340B
The Drug
Discount Program
What it is. Why it matters.
DHSc, PhD, MBA, BSRT(ARRT)R
A program with a public purpose
340B helps eligible healthcare organizations stretch limited resources.
Created in 1992 and administered by the Health Resources and Services Administration, the program requires participating drug manufacturers to offer discounted covered outpatient medicines to eligible, enrolled providers.
Qualifying hospitals, federally qualified health centers, Ryan White providers, and specified clinics can participate. Eligibility differs across categories, and neither every hospital nor every prescription qualifies. The intended benefit is organizational capacity: reaching more eligible patients and providing more comprehensive services. Understanding that design is the first step toward evaluating whether the program delivers on its mission.
How the savings work
The purchasing discount and the insurer’s payment are separate parts of the transaction.
The amount available varies with drug mix, payer mix, pharmacy arrangements, and administrative expenses. A large purchase total is not the same as a savings total, and retained revenue should not automatically be labeled profit. Because 340B operates through manufacturer discounts, it differs from a direct federal grant for each prescription.
Why safety-net care matters
For some providers, lower drug costs can help sustain care that ordinary reimbursement does not fully finance.
Potential uses include medication assistance, pharmacy services, and other underfunded care. These resources matter where patients face financial barriers or where essential services are difficult to support.
Research provides qualified support for this role. Studies link participation to expanded care at community health centers and to additional service availability at some public hospitals. Benefits differ across organizations, and these findings do not establish that every institution achieves the same results. Local evidence should show which services were supported and who gained access.
What patients actually pay
A lower provider purchase price does not automatically reduce a patient’s copay.
Eligible patients may have insurance. Their out-of-pocket costs depend on coverage and the provider’s assistance arrangements, including where a prescription is dispensed. The program is therefore not a universal patient discount card.
Patients can ask their care team or pharmacy about financial assistance and how to apply. Leaders should make those pathways understandable and explain how program resources support affordability, access, and services. GAO documented differing contract-pharmacy discount practices among selected providers, but its sample does not establish a current national rate.
Two policy developments
A revised rebate pilot and a Medicare payment proposal affect different parts of 340B financing.
HRSA’s revised pilot would replace upfront discounts with later rebates for selected Medicare-negotiated drugs. Approved manufacturer plans would begin January 1, 2027; approvals, if any, are scheduled by September 24, 2026. The pilot is not yet operating.
Separately, CMS has proposed lower Medicare payments for certain 340B-acquired hospital outpatient drugs. That proposal concerns reimbursement and includes exceptions. Neither development means that 340B is ending. Their implications for patient costs, provider cash flow, and resources available for care require careful evaluation.
Protect the program. Show its value.
Preserving safety-net capacity and demanding evidence of patient benefit belong together.
Leadership reporting should connect program resources with medication assistance, care for underserved patients, and support for services that are difficult to finance. Report relevant program costs as well as benefits, and explain where the evidence remains incomplete.
Compliance also matters. Confirm covered-entity and patient eligibility, prevent diversion, and avoid prohibited duplicate Medicaid discounts. GAO has identified weaknesses in oversight that merit attention. Financial scale alone does not establish performance. Evaluate proposed reforms through patient affordability, provider capacity, and public spending, while keeping those distinct outcomes visible.
Keep patients at the center
340B can be an essential resource for providers serving communities with limited access to care.
Its promise is broader access to medicines and services. Protecting that promise requires a clear account of the people reached, the barriers reduced, and the services sustained.
What the research supports
The evidence supports a meaningful safety-net role, with important differences between settings.
Watts and colleagues studied 1,468 federally qualified health centers during 2004–2022. More registered 340B locations were associated with greater subsequent service to underserved groups and more preventive services. Locations were a proxy for financial benefit; the study cannot establish causation.
Owsley and colleagues used a difference-in-differences design to examine hospital participation. Public hospitals entering 340B showed increased availability of service lines classified as unprofitable. Nonprofit hospitals did not show a meaningful corresponding change. Selection and sample limitations remain.
Five questions for leaders
Translate the safety-net mission into questions your organization can answer.
Use these questions for board discussion, operating reviews, and patient-access planning. They are proposed management practices, not a validated research instrument.
- Who benefits? Describe the patients reached, assistance provided, and access barriers addressed.
- What does the program support? Identify services and their associated costs.
- Can patients find help? Make assistance criteria and application routes understandable.
- Are controls reliable? Review eligibility, duplicate-discount prevention, and corrective actions.
- How could policy changes affect care? Examine cash flow, affordability, service capacity, and public spending.
Research and oversight
Peer-reviewed studies and government analyses underpin the evidence discussion.
- Watts, E., McGlave, C., Quinones, N., Bruno, J. P., & Nikpay, S. (2024). 340B participation and safety net engagement among federally qualified health centers. JAMA Health Forum, 5(10), e243360.
- Owsley, K. M., Hasnain-Wynia, R., Rooks, R. N., Tung, G. J., Mays, G. P., & Lindrooth, R. C. (2024). US hospital service availability and new 340B program participation. JAMA Health Forum, 5(5), e240833.
- U.S. Government Accountability Office. (2018). Drug discount program: Federal oversight of compliance at 340B contract pharmacies needs improvement (GAO-18-480).
- U.S. Government Accountability Office. (2025). 340B drug discount program: Agency oversight has improved, but actions needed to address weaknesses (GAO-26-108784).
- Congressional Budget Office. (2025). Growth in the 340B Drug Pricing Program.
Policy sources and notes
Policy status was checked September 6, 2026. Proposals may change.
- Health Resources and Services Administration. (2026). 340B Drug Pricing Program.
- HRSA. (2026). 340B eligibility.
- HRSA. (2026, August 3). Notice regarding 340B Rebate Model Pilot Program. Federal Register, 91, 48883.
- Centers for Medicare & Medicaid Services. (2026, July 2). Calendar year 2027 OPPS and ASC proposed rule.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
Adapted from the 340B video and executive evidence brief. AI narration and illustrative generated scenes do not document actual patients or institutional outcomes.
THE LEADERSHIP FIELDBOOK 2026 EDITION
People. Judgment. Performance.
THE LEADERSHIP FIELDBOOK
Radiology
Leadership
Building teams. Improving care.
Leading the imaging enterprise.
DHSc, PhD, MBA, BSRT(ARRT)R
2026 EDITION
The central commitment
Leadership is a responsibility to care.
Every leadership decision eventually reaches a patient.
A staffing decision affects whether an examination happens on time. A scheduling rule can ease a patient’s uncertainty or prolong it. A technology investment can strengthen professional judgment or create another interruption. Radiology leadership connects these choices to the people who experience their consequences.
The work requires clinical partnership, operational discipline, financial understanding, and the courage to address difficult problems. These capabilities become useful when they lead to reliable care and a working environment in which people can do their best work.
Eight priorities. One connected service.
Inside the fieldbook
Select a chapter to turn directly to it. Research references are linked throughout the book.
A guide for the working leader
Read. Discuss. Put it to work.
This fieldbook is written for radiology executives, physician leaders, imaging directors, managers, and emerging leaders. Its purpose is to connect leadership judgment with the everyday work of delivering diagnostic care. Read it in sequence to examine the service as a whole, or choose a chapter that addresses the decision in front of you.
Each chapter pairs a leadership principle with a practical application. Use the closing question in a leadership meeting, a one-to-one conversation, or a discussion with the people closest to the work. The companion tools help turn those conversations into defined responsibilities and measurable action.
- 1Identify a real problem.Describe the patient or staff experience you want to improve.
- 2Agree on the next decision.Name an owner, define useful evidence, and set a review date.
- 3Return to the result.Ask whether the change worked and what the team learned.
Turn pages with the arrows or swipe. Contents includes topic search. Reading view displays the complete text; Print / PDF opens a print-ready edition.
Purpose and responsibility
Make the mission operational
Radiology leadership begins with a practical promise: help patients obtain the right examination, receive a reliable interpretation, and move to an appropriate next step. That promise depends on coordinated decisions across scheduling, technology, clinical practice, staffing, finance, and referring relationships. An imaging service can produce more examinations while patients continue to experience fragmented care. The leader's responsibility is to connect the entire pathway and make its performance visible to the people who can improve it.
This book proposes an operating model built around eight leadership responsibilities. Each responsibility should have an accountable owner, a measurable outcome, a review rhythm, and a clear response when performance falls short. Radiologist and administrator partnership literature emphasizes intentional collaboration and shared institutional values, although it does not establish a universally superior organizational structure (France et al., 2019). Translate that principle into daily practice: give clinical and operational leaders a common problem to solve, an agreed decision process, and the resources to act.
If a patient followed one referral through your service today, where would responsibility become unclear? Begin your leadership agenda at that handoff.
A working leadership charter
Build a physician partnership
A physician and an administrator bring different knowledge to the same responsibility. The physician partner can lead clinical standards, interpretation quality, and professional practice. The administrative partner can lead resource planning, staffing systems, patient access, and financial execution. Decisions that affect both domains require explicit joint review. Partnership should extend to technologists, nurses, schedulers, and informatics staff whose work determines whether a policy can function at the point of care.
The following charter is a proposed management tool informed by the partnership principles described by France et al. (2019). Define the shared mission, each leader's decision rights, spending authority, clinical escalation route, and method for resolving disagreement. Reserve a weekly meeting for decisions and unresolved barriers. Maintain a short decision log stating the issue, evidence, owner, deadline, and outcome. When a dispute persists, identify the governing question, such as patient safety, service capacity, or financial sustainability, and bring it to the designated executive or clinical authority.
Write one joint goal for the next quarter. Name the physician owner, operational owner, frontline contributors, decision date, and evidence that will demonstrate success.
Psychological safety in practice
Make speaking up worthwhile
Psychological safety concerns whether people believe they can raise questions, admit uncertainty, or discuss mistakes without interpersonal punishment. Edmondson's foundational field study linked team psychological safety with learning behavior in organizational work teams; the setting was not radiology, so application here requires judgment (Edmondson, 1999). For an imaging leader, the practical question is whether the newest employee can interrupt a familiar workflow when something appears unsafe, incomplete, or inconsistent with the patient’s needs.
Create a dependable response to concerns. Thank the person, clarify the immediate risk, protect the patient when needed, and report back on the action taken. Pair that invitation with clear standards and fair follow-through. When expectations are missed, examine training, workload, equipment, process design, and individual choices before deciding on a response. Avoid public humiliation and vague reassurance. A confidential concern that disappears into an inbox teaches staff little about whether leaders will act. A visible correction, with appropriate confidentiality, makes the leadership commitment concrete.
Ask: What problem do experienced staff quietly work around? Choose one recurring workaround, identify its cause, and tell the team when it will be reviewed.
Workforce and professional fulfillment
Design work people can sustain
A 2025 Canadian physician survey found that more favorable ratings of supervisors' leadership qualities were associated with lower burnout and greater organizational satisfaction. Its cross-sectional design cannot establish that improving leadership alone causes those outcomes, and its pandemic context matters (Spilg et al., 2025). The implication for radiology leaders is to examine the organization of work alongside individual support, while avoiding promises that a single wellness initiative will solve structural pressures.
Build a workforce plan that reflects modality skills, examination complexity, coverage hours, leave, orientation, and noninterpretive duties. Count the time spent answering calls, resolving protocols, supervising procedures, training colleagues, and managing exceptions. Review vacancies, overtime, turnover, and confidential staff feedback together. Where demand exceeds available capacity, make the tradeoff explicit: adjust access commitments, obtain additional coverage, redesign tasks, or change the service scope. Document how decisions affect patient safety and the remaining team. Treat leadership development as a continuing responsibility, including coaching, feedback, succession planning, and protected opportunities to learn.
Select one source of avoidable daily friction. Assign an owner to remove it, then ask affected staff whether the change actually reduced their workload.
Access across the referral pathway
Follow the patient before the slot
Access begins when an imaging need is identified. Map the pathway from referral receipt through order clarification, authorization, scheduling, preparation, examination, and report delivery. Record elapsed time and unresolved work at each step. A short scheduling call does not establish good access if the order waited several days before that call occurred. Review the patient's experience of uncertainty, transportation, language, preparation instructions, and out-of-pocket expectations alongside operational measures.
Do not assume reminders will solve every missed appointment. In one MRI study, an automated reminder intervention produced no statistically significant overall reduction in missed appointments, although a Medicaid subgroup improved (Kenniff & Ginat, 2023). Use that finding as a reason to diagnose local barriers. Offer a simple route to confirm, cancel, or reschedule; provide assistance matched to the barrier; and evaluate completion rather than message delivery alone. Examine differences by modality, location, appointment lead time, and relevant patient groups. Ensure that access interventions support patients who face additional obstacles to completing care.
Trace ten recently delayed referrals. Record the first unresolved barrier, who could resolve it, and whether the patient knew what would happen next.
An illustrative planning exercise
Capacity is a usable resource
Separate theoretical capacity, staffed capacity, scheduled capacity, and completed examinations. A scanner's opening hours do not automatically translate into patient appointments. Protocol mix, preparation, room turnover, maintenance, interruptions, and available personnel determine what the service can reliably complete. Build schedules from observed local process times and clinical requirements, then reserve appropriate flexibility for variation. A plan that assumes every minute will proceed exactly as scheduled can leave little room for a patient who needs additional support.
Illustrative example: an eight-hour day provides 480 minutes. If planned nonappointment time totals 48 minutes, 432 minutes remain. At an average appointment duration of 40 minutes, the arithmetic is 10.8 appointment equivalents. For a uniform 40-minute template, ten complete slots fit; the remaining time is not an eleventh full appointment. Actual templates require examination-specific durations and staffing validation. Track why offered capacity becomes unused, including cancellations, preparation failures, equipment downtime, and staffing gaps. Test one constraint at a time and monitor whether increased throughput changes delays, overtime, or repeat imaging.
Define every capacity denominator. Report booked slots, completed examinations, and staffed appointment minutes separately so that a change in one cannot conceal deterioration in another.
From examination to action
Own the whole diagnostic pathway
Diagnostic safety requires attention to how images are acquired, interpreted, communicated, and used. Itri et al. (2018) describe perceptual, cognitive, and environmental contributors to imaging error. Their review supports examining both professional judgment and the conditions in which judgment occurs. A leadership response should therefore include the quality of clinical information, access to prior studies, image adequacy, communication practices, and the operational environment surrounding interpretation.
Use a locally approved pathway for findings that require urgent communication or recommended follow-up. Define who receives the result, what constitutes acknowledgment, who owns the next clinical action, and how unresolved communication escalates. Responsibilities and timelines should reflect clinical urgency and organizational policy. Audit a sample from finding identification to documented handoff or follow-up status. A report marked final is an important milestone, but leaders still need evidence that the communication process functions as intended. Include modality safety, contrast processes, emergency readiness, and equipment quality assurance in the broader quality program, with appropriate clinical and technical oversight.
For the last significant communication failure, can you identify the exact handoff that failed and the system change that would make recurrence less likely?
A practical peer-learning cycle
Turn experience into learning
Peer learning gives radiologists a structured opportunity to examine discrepancies, strong interpretations, and opportunities for improvement. In a 2023 survey of ACR members, participating users commonly perceived benefits for safety culture and continuous improvement. Those responses describe professional perceptions, not proof of reduced patient harm; the low survey response rate also limits confidence in broad generalization (Sharpe et al., 2023). Use the evidence to inform program design while measuring local results.
Establish an agreed route for submitting learning opportunities and a respectful process for discussion. Ask what made the case difficult, what information was available at the time, and whether an underlying workflow can be improved. Assign selected changes to named owners and revisit them after implementation. Include excellent calls so that the service learns from effective practice as well as mistakes. Keep educational discussion distinct from the organization's formal response to serious competence or conduct concerns. Evaluate participation and completed improvement actions, then connect selected actions to relevant process or patient outcomes when meaningful measurement is feasible.
End each learning discussion with one sentence: Because of this case, we will change this practice, with this owner, and review it on this date.
Economics of the care process
Know what care actually costs
Sound financial leadership starts with clear definitions. Charges are prices presented for billing; recognized revenue and cash collections answer different questions. Cost per examination also depends on the expenses included and how shared resources are assigned. A service can report higher volume while its collection yield, staffing requirements, or service mix changes unfavorably. Leaders need a view of clinical activity that reconciles with the financial statements and cash position.
Time-driven activity-based costing offers a process-level approach to understanding resource use in radiology (Shankar et al., 2020). Map the activities needed to deliver care, estimate practical resource capacity, calculate a cost per available minute, and apply observed time requirements. For management decisions, supplement that analysis with transparent treatment of overhead, financing, and capital replacement. Use a consistent definition when comparing modalities or sites. Examine variation as a starting point for investigation: longer time may reflect necessary clinical complexity, an inefficient process, or incomplete measurement. Financial stewardship requires distinguishing those explanations before changing staffing or appointment duration.
Ask finance and operations to reconcile one modality's examinations, recognized revenue, cash collections, direct expenses, allocated expenses, and capacity assumptions on a single worksheet.
Illustrative break-even thinking
Make the business case explicit
A proposed service deserves an explicit economic model and a clinical rationale. Define expected demand, realistic collections, variable expense, fixed commitments, implementation time, and the resources needed to sustain quality. Separate the incremental decision from the full service-line result. Filling an otherwise available slot and adding a new scanner involve different cost commitments. Document which assumptions come from actual organizational data and which remain uncertain.
Illustrative example: assume monthly fixed operating costs of $90,000, revenue of $600 per completed examination, and variable cost of $300 per examination. Contribution is $300 per examination. Operating break-even is $90,000 divided by $300, or 300 examinations, producing $180,000 in revenue. This simplified model assumes a stable service mix and collection yield, and excludes taxes, financing, and capital costs unless included explicitly. It is not a cash-flow forecast. Test lower reimbursement, slower volume growth, and higher staffing expense before committing resources. Where a clinically necessary service requires subsidy, present that decision transparently with its patient and strategic value.
Before approval, ask which assumption would most change the decision and what evidence could reduce that uncertainty. Set a date to compare actual results with the business case.
Technology selection and governance
Start with a clinical purpose
Begin technology decisions with a precise description of the problem. Is the service trying to reduce repetitive work, improve communication, support detection, shorten reporting delay, or improve image quality? Define the affected users and the outcome that matters to patients or operations. Establish a baseline before reviewing product demonstrations. A technically impressive system may still fit poorly with the workflow or produce benefits that the organization cannot realize.
The 2024 multisociety AI statement recommends evaluating practical, ethical, technical, and monitoring considerations throughout adoption. It is professional consensus, not a trial showing that any particular tool improves care (Brady et al., 2024). For a proposed purchase, request evidence relevant to the intended population, imaging protocols, and use case. Designate clinical and operational owners, involve information security and informatics, and clarify the clinician's responsibility for reviewing outputs. Include integration, training, support, and ongoing oversight in the cost assessment. Make approval contingent on an agreed evaluation plan and a workable response when the system is unavailable or produces an unexpected result.
What specific patient or workflow outcome will change, how will we measure it, and who will remain responsible after the vendor implementation team leaves?
Monitoring the human and system
Treat deployment as a beginning
An AI system's performance can change when equipment, protocols, patient populations, or software versions change. The multisociety statement emphasizes monitoring and local evaluation across the technology's lifecycle (Brady et al., 2024). Define performance measures relevant to the intended task, including failure modes and subgroup performance where feasible. Assess how clinicians use the output and whether the combined workflow delivers the intended benefit. Accuracy in a development dataset does not establish effectiveness in every clinical setting.
The proposed leadership tool is a technology register that gives each application a purpose, owner, version, review date, and escalation route. At each review, ask whether the original problem remains, whether staff can use the tool reliably, and whether its operating burden remains justified. Record incidents and operational interruptions alongside benefits. When material changes occur, determine whether renewed evaluation is needed before wider use. Protect the ability to pause the application and continue essential services through a defined alternative process. Treat staff feedback as evidence for investigation, then compare it with objective performance data.
For every deployed application, identify the person authorized to pause its use and the evidence that would trigger that decision. Rehearse the alternative workflow.
A coherent service portfolio
Design the imaging enterprise
Enterprise strategy asks where imaging capabilities should sit, how sites should work together, and which patient needs the organization can reliably meet. Evaluate access, clinical capability, workforce availability, capital requirements, referring relationships, and continuity during disruption. A new location may expand geographic access while adding staffing and coordination demands. A centralized function may improve consistency while creating a dependency that needs an alternative operating plan.
Workforce projections illustrate why strategy should examine several futures. Christensen et al. (2025) modeled future US radiologist supply under different assumptions about residency growth and attrition; these estimates are conditional forecasts, not guarantees of local availability. Build a service portfolio that identifies essential services, opportunities for development, and activities requiring redesign. Standardize patient safety expectations, core definitions, and data exchange while allowing justified differences in clinical capability and local demand. Give each investment a named sponsor, expected patient benefit, resource plan, and measurable checkpoint. Discuss how the enterprise will function if recruitment is slower or equipment replacement takes longer than planned.
Compare three futures: expected demand, constrained staffing, and accelerated growth. Identify the decisions that remain useful across all three and the commitments that depend on one forecast.
Learning through disciplined execution
Test change before scaling it
A change initiative should begin with a defined problem, an explanation of why the proposed intervention might help, and a prediction that can be examined. Taylor et al. (2014) found that published applications of plan-do-study-act often omitted essential features such as iterative cycles and data over time. Naming the method is insufficient. Leaders need to ensure that the work produces usable learning rather than a series of activities labeled improvement.
For a practical test, select one site, team, or workflow where the problem is clear. State the baseline, proposed change, expected result, responsible owner, and observation period. Identify a balancing measure that could reveal a new burden, such as overtime, repeat examinations, or delays elsewhere. Review results with the people doing the work and decide whether to adapt, expand, or stop. Record what differed from the original prediction. Before spreading the change, confirm that necessary staffing, training, and technical conditions exist in the receiving setting. Include patients or referring clinicians when the change alters their experience.
Complete this sentence before starting: We predict that this change will improve this measure because of this mechanism, while this balancing measure remains acceptable.
A proposed execution sequence
Move from assessment to action
Days 1 to 30: establish the leadership charter and learn how the service actually operates. Observe representative workflows, speak with staff and patients, review urgent safety concerns, and reconcile the essential activity and financial measures. Build a short baseline covering access, workforce, quality, and financial sustainability. Confirm who owns each result. Resolve immediate hazards through the appropriate clinical and operational channels while the broader assessment proceeds.
Days 31 to 60: choose a manageable improvement portfolio. Define one access problem, one workforce or workflow problem, and one quality priority, then select projects the team can realistically support. For each, document the intervention, owner, baseline, intended outcome, balancing measure, and review date. Days 61 to 90: examine results, adapt unsuccessful tests, and standardize changes that have demonstrated useful local improvement. Present decisions and unresolved risks to the leadership partners. This sequence is an original management framework, not a validated implementation protocol. The timing should reflect service size, urgency, and readiness rather than forcing work to meet an arbitrary calendar.
Bring a concise record of what changed, what the evidence shows, what remains unresolved, and which decisions require additional resources or executive action.
A proposed leadership review
Use a scorecard that prompts action
Use a compact scorecard with a few measures that connect to decisions. Suggested domains are patient access, completed care, workforce stability, diagnostic communication, technical quality, financial sustainability, and improvement execution. These are proposed management domains, not a validated scale. Choose measures that the service can define consistently and influence directly. For access, examine referral-to-examination time and its distribution; for completed care, examine completion among eligible scheduled appointments; for quality, examine documented closure of selected communication tasks.
Give every measure a definition, source, owner, baseline, target rationale, and review frequency. Include the numerator and denominator for rates, relevant exclusions, and case-mix context. Compare trends with experience from patients and frontline staff. Do not interpret a single favorable month as proof of sustained improvement. In the leadership review, ask what changed, what explains the change, who is affected, and what decision follows. Close with an action log that carries forward unresolved commitments. The scorecard earns its place when it helps leaders decide where to investigate, where to support staff, and where to direct resources.
Choose one measure that needs action now. Name the accountable leader, the next decision, the evidence required, and the date the team will hear what happened.
Companion tool 01
Make decision rights visible
Use this starting point to write a local charter. Confirm authority with medical staff leadership and organizational policy, then assign named people to each responsibility.
| Decision | Accountable lead | Required partners |
|---|---|---|
| Clinical protocols and interpretation standards | Physician lead | Technologists, nursing, quality |
| Staffing and patient access | Operational lead | Physician lead, frontline managers |
| Capital and service expansion | Named executive sponsor | Clinical, operations, finance |
| Diagnostic communication | Designated clinical owner | Referrers, operations, informatics |
| AI approval and monitoring | Designated governance lead | Clinical, IT, security, operations |
Proposed management tool. Titles vary by organization; accountability must be assigned to a person with the necessary authority.
For each decision, add spending or scope limits, the escalation route, the meeting at which it is reviewed, and the person who communicates the outcome.
Companion tool 02
A scorecard with a purpose
Select a manageable set from these examples. Set local targets after validating the baseline, case mix, and operational capacity.
| Domain | Example definition |
|---|---|
| Access | Median and 90th-percentile calendar days from accepted referral to completed examination. |
| Completion | Completed eligible appointments ÷ all eligible scheduled appointments × 100. |
| Workforce | Voluntary departures ÷ average headcount × 100 for a stated period. |
| Communication | Sampled actionable results with documented required handoff ÷ sampled results requiring that handoff × 100. |
| Technical quality | Exams repeated for technical failure ÷ completed exams × 100, using agreed exclusions. |
| Financial performance | Recognized revenue less explicitly defined operating expenses, reconciled with finance. |
| Execution | Improvement actions completed by due date ÷ actions due in the period × 100. |
Specify the source, owner, exclusions, observation window, and denominator. Review modality and site differences before interpreting an enterprise average.
Companion tool 03
Connect capacity and contribution
Two simple calculations connect the schedule with the operating model. These assumed values are teaching examples and do not describe KCAI or another organization.
480 minutes − 48 planned minutes = 432 minutes
432 ÷ 40 minutes = 10.8 appointment equivalents
Ten complete uniform 40-minute slots
$600 revenue − $300 variable cost = $300 contribution per examination
$90,000 fixed costs ÷ $300 = 300 examinations
300 × $600 = $180,000 revenue
At 20 operating days and ten complete slots each day, the illustrative template offers 200 slots. It cannot deliver 300 examinations without a change in operating days, hours, resources, or the validated examination mix. Demand and staffing must also support the revised plan.
Assumes uniform durations and a stable service mix. This operating example excludes financing, taxes, and capital costs unless explicitly included in the stated expenses. Cash flow requires a separate model.
Reconcile the volume required by the business case with the capacity the team can safely and reliably provide.
Companion tool 04
Your 90-day leadership worksheet
Complete this worksheet with the physician and operational leadership partners. Keep the initial project portfolio small enough to receive meaningful attention.
| Period | Working deliverable |
|---|---|
| Days 1 to 30 Understand | Shared charter; patient pathway; validated baseline; urgent risks with accountable owners. |
| Days 31 to 60 Test | Defined access, workforce, and quality priorities; practical tests; balancing measures. |
| Days 61 to 90 Decide | Results reviewed; changes adapted or adopted; resource decisions and next-quarter priorities. |
- 1Priority and patient benefitWhat problem are we addressing, and who should experience an improvement?
- 2Owner, measure, and dateWho can act, what evidence will we review, and when?
- 3Decision after the testWill we adapt, expand, or stop the change, and why?
Explain the results, acknowledge uncertainty, identify unfinished work, and make the next decision clear. Calendar milestones support judgment; they do not replace it.
Research and professional guidance · 1 of 4
References
REFERENCE 01
France, T. J., Menaker, R., & Thielen, K. R. (2019). The importance of a radiologist-administrator partnership to future health care. Journal of the American College of Radiology, 16(8), 1114–1118.
doi:10.1016/j.jacr.2018.10.023REFERENCE 02
Spilg, E. G., McNeill, K., Dodd-Moher, M., Dobransky, J. S., Sabri, E., Maniate, J. M., & Gartke, K. A. (2025). Physician leadership and its effect on physician burnout and satisfaction during the COVID-19 pandemic. Journal of Healthcare Leadership, 17, 49–61.
doi:10.2147/JHL.S487849REFERENCE 03
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383.
doi:10.2307/2666999Research and professional guidance · 2 of 4
References
REFERENCE 04
Christensen, E. W., Parikh, J. R., Drake, A. R., Rubin, E. M., & Rula, E. Y. (2025). Projected US radiologist supply, 2025 to 2055. Journal of the American College of Radiology, 22(2), 161–169.
doi:10.1016/j.jacr.2024.10.019REFERENCE 05
Kenniff, J., & Ginat, D. (2023). Evaluation of an automated reminder system for reducing missed MRI appointments. Journal of Patient Experience, 10, 23743735231151548.
doi:10.1177/23743735231151548REFERENCE 06
Itri, J. N., Tappouni, R. R., McEachern, R. O., Pesch, A. J., & Patel, S. H. (2018). Fundamentals of diagnostic error in imaging. RadioGraphics, 38(6), 1845–1865.
doi:10.1148/rg.2018180021Research and professional guidance · 3 of 4
References
REFERENCE 07
Sharpe, R. E., Jr., Tarrant, M. J., Brook, O. R., Chatfield, M., Chaudhry, H., City, R. B., Donnelly, L. F., Goldberg-Stein, S., Hernandez, D., Hwang, G. L., Kunst, M. M., Lee, R., Moriarity, A. K., Pahade, J. K., Patel, S., & Broder, J. C. (2023). Current state of peer learning in radiology: A survey of ACR members. Journal of the American College of Radiology, 20(7), 699–711.
doi:10.1016/j.jacr.2023.03.018REFERENCE 08
Shankar, P. R., Hayatghaibi, S. E., & Anzai, Y. (2020). Time-driven activity-based costing in radiology: An overview. Journal of the American College of Radiology, 17(1, Part B), 125–130.
doi:10.1016/j.jacr.2019.07.010Research and professional guidance · 4 of 4
References
REFERENCE 09
Brady, A. P., Allen, B., Chong, J., Kotter, E., Kottler, N., Mongan, J., Oakden-Rayner, L., Pinto dos Santos, D., Tang, A., Wald, C., & Slavotinek, J. (2024). Developing, purchasing, implementing and monitoring AI tools in radiology: Practical considerations. A multi-society statement from the ACR, CAR, ESR, RANZCR & RSNA. Insights into Imaging, 15, Article 16.
doi:10.1186/s13244-023-01541-3REFERENCE 10
Taylor, M. J., McNicholas, C., Nicolay, C., Darzi, A., Bell, D., & Reed, J. E. (2014). Systematic review of the application of the plan-do-study-act method to improve quality in healthcare. BMJ Quality & Safety, 23(4), 290–298.
doi:10.1136/bmjqs-2013-001862Scope and interpretation
About the evidence
This book combines selected research and professional guidance with an original practical framework for radiology leadership. Targeted searches of PubMed, journal websites, and academic repositories were completed on September 4, 2026. The ten selected sources span 1999 to 2025 and include observational studies, surveys, a workforce projection, reviews, and a multisociety AI statement. This is a focused narrative synthesis, not a systematic review. Associations do not establish causation; forecasts depend on assumptions; local improvement findings may not generalize. The AI statement underwent formal review by its participating societies rather than the publishing journals. The book's operating model, worksheets, scorecard, and 90-day sequence are proposed management tools, not independently validated instruments. All financial and capacity examples are illustrative, with no organizational performance data supplied.
The leadership charter, scorecard, worksheets, and integrated capacity example extend the chapter discussion into practical applications. They should be adapted with the people accountable for the work. Establish local definitions and confirm that changes improve the intended process without creating unacceptable burdens elsewhere.
Prepared as an original leadership fieldbook for Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R. September 2026. Cover image is an AI-generated editorial illustration depicting fictional professionals; it is not a portrait of the author or a photograph of a named facility.
Use research to sharpen the question, local data to understand the setting, and professional judgment to choose the next action.
Radiology Leadership
Lead the work. Care for the people.
A reliable imaging service is built through many connected decisions. The leader’s task is to make their purpose clear, equip people to act, and return to the results with honesty.
Start with one patient pathway, one avoidable barrier, or one commitment the team needs you to keep. Give it an owner. Make the next decision explicit. Then follow through.
National Healthcare Operations Study | 2026
Beyond the Emergency Waiting Room
The national emergency department waiting and boarding crisis: a system-level evidence synthesis, policy analysis, and enterprise operating model for safer hospital flow. Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
NATIONAL HEALTHCARE OPERATIONS STUDY | 2026
Beyond the Emergency
Waiting Room
The National Emergency Department Waiting and Boarding Crisis
A system-level evidence synthesis, policy analysis, and enterprise operating model for safer hospital flow
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
An independent scholarly synthesis prepared for healthcare leaders, clinicians, policymakers, and researchers.
National Healthcare Operations Study | 2026
Beyond the Emergency Waiting Room
An integrative evidence synthesis, policy analysis, and implementation framework for the national emergency department waiting and boarding crisis.
Evidence available through August 30, 2026. No patient-level data were collected or analyzed; institutional review board review was not required. No external funding was received. The author reports no competing interests.
This work is a scholarly synthesis and a proposal for implementation. It must not be represented as a completed trial, a validated predictive instrument, or a substitute for local clinical governance.
Reader map
Contents
- 1Executive judgment5
- 2The crisis in a familiar room15
- 3Methods21
- 4National epidemiology25
- 5The harm signal32
- 6Operational physics40
- 7Inequity in the queue46
- 8Why remedies disappoint51
- 9Intervention evidence56
- 10Enterprise Flow Reliability System62
- 11Implementation68
- 12Measurement and governance75
- 13Policy and conclusion80
- 14Appendices88
Part one
Executive judgment
The decision, the evidence strength, and the action logic
“
Boarding is not a location problem. It is accumulated clinical, operational, and social work that a hospital shifts into its least protected space.
Treating it as emergency department congestion preserves the bottlenecks that create it.
The governance error
The decisive governance error is assigning accountability to the emergency department, since that is where the queue becomes visible. Emergency clinicians can redesign triage, initiate orders, split streams, and use vertical care. Those steps may improve waiting-room performance and should be pursued when locally indicated.
Once a patient requires a staffed inpatient bed, however, the dominant constraint usually lies in the hospital’s output system: unavailable staffed beds, delayed specialty decisions, asynchronous discharge work, uneven bed turnover, post-acute scarcity, and authority dispersed across services.
The emergency department can absorb risk, but it cannot manufacture inpatient capacity.
The recommended decision
Classify boarding beyond locally validated risk thresholds as an enterprise safety event. The chief executive, chief medical officer, chief nursing officer, and governing board should own a balanced set of measures.
- The median and the tails of boarding time
- Time to inpatient-team management
- Deterioration and delirium
- Left-without-being-seen rates
- Transfer acceptance
- Patient experience and staff harm
- Readmission and equity gaps
Capital expansion should follow, not precede, a disciplined assessment of staffed capacity, temporal mismatch, and avoidable process variation.
Figure 1
National signals show both scale and post-decision delay.
Evidence type: observed national or multicenter estimates. Measures are not directly additive. The CMS value is a visit-duration measure, not a boarding-only measure.
Table 3
Executive decision matrix
| Question | Judgment | Confidence |
|---|---|---|
| Is this mainly an emergency department demand problem? | No. Demand matters, but admitted-patient boarding is principally an output and staffed-capacity problem with hospital-wide causes. | High for system mechanism; moderate for any one local cause. |
| Does prolonged boarding cause harm? | The total evidence is consistent, biologically plausible, temporally ordered, and supported by dose-response signals. Residual confounding remains. | Moderate overall; higher for association than for exact causal effect size. |
| Will a command center solve it? | Not by itself. It improves shared situational awareness and execution only when paired with decision rights, standard work, and accountable services. | Low to moderate; outcome evidence is largely observational. |
Confidence ratings are the author's synthesis of design quality, consistency, directness, and implementation dependence. They are not formal GRADE ratings.
Table 3, continued
Executive decision matrix
| Question | Judgment | Confidence |
|---|---|---|
| Should hospitals add beds? | Only when analysis shows a structural deficit in staffed capacity after timing, progression, discharge, and post-acute constraints are addressed. | Moderate; highly context-dependent. |
| What should happen first? | Standardize care for boarded patients, define clocks, establish executive control, measure tails and equity, and remove the top few verified constraints. | High as a safety and implementation sequence. |
Local context determines whether any single intervention is appropriate.
Importance and objective
Abstract
Importance. Emergency department boarding is the continued care of a patient in an emergency department or temporary location after a decision to admit or transfer. It is now a routine manifestation of hospital-wide capacity failure rather than an episodic inconvenience. Contemporary evidence associates prolonged boarding with mortality, delirium, medication delay, longer stays, worse patient experience, clinician harm, and diminished regional transfer access.
Objective. To synthesize current national and peer-reviewed evidence, distinguish causal findings from associations and modeled inferences, identify why common countermeasures underperform, and propose an enterprise operating system for reducing boarding while protecting safety and equity.
Findings
Abstract
CDC estimated 155.4 million U.S. emergency department visits in 2022. National data from 46.2 million hospitalizations showed that at the January 2022 peak, 40.1% boarded for more than 4 hours and 6.3% for more than 24 hours.
In a 2026 multicenter study across 56 emergency departments, 263,666 of 492,135 general-medicine admissions, or 53.6%, waited at least four hours for inpatient management.
A prospective cohort of adults aged 75 years or older found mortality of 15.7% among those remaining in the emergency department overnight versus 11.1% among those reaching a ward before midnight, adjusted risk ratio 1.39.
Conclusions and relevance
Abstract
Boarding should be governed as an enterprise patient-safety failure with executive and board accountability. Front-end emergency department interventions may shorten selected intervals but cannot reliably remove an output constraint.
Hospitals should combine immediate boarded-patient protections with demand sensing, timely transition to inpatient management, staffed-capacity control, inpatient progression, predictable discharge, post-acute coordination, and equity-stratified measurement.
The proposed Enterprise Flow Reliability System offers a testable operating hypothesis and measurement architecture, not a validated clinical tool.
Keywords. Emergency department; boarding; crowding; hospital occupancy; patient flow; patient safety; health equity; interhospital transfer; operations management; implementation science.
Part two
The crisis in a familiar room
Problem framing and the definitions that determine accountability
At 3:14 in the morning
Every staffed inpatient bed is assigned. A frail older adult who was admitted six hours earlier remains under bright emergency department lights. Her home medications are being reconciled between teams. A second patient needs telemetry. A third has completed an emergency evaluation but cannot reach the regional specialty center that declined the transfer while managing its own internal queue. New ambulances continue to arrive.
This is a composite operational scene, not an account of a named patient. Its ordinariness is the point.
Emergency departments were designed to identify immediate threats, stabilize patients, and determine disposition. Boarding converts them into unscheduled inpatient units without the physical environment, staffing model, pharmacy routines, therapy access, sleep conditions, or accountability structures of wards.
A national question of access
The mismatch is not simply uncomfortable. It creates a state in which care responsibility, surveillance, medication administration, diagnostic follow-up, mobility, nutrition, privacy, and communication can all become less reliable.
The national scale makes this an issue of access, quality, workforce, and regional equity at once. CDC estimated 155.4 million U.S. emergency department visits in 2022, with 17.8 million resulting in hospital admission.
CMS publicly reports emergency care timeliness, and current data show wide variation across hospitals and states. Those measures describe total visit intervals and should not be mistaken for a pure boarding clock. The more specific national boarding evidence is nevertheless stark: hours after the admission decision are common rather than exceptional.
CDC, 2024; CMS, 2026; Janke et al., 2025, 2026
Section 1.1
Definitions determine accountability
Crowding, waiting, length of stay, access block, and boarding are related but distinct. Crowding is the condition in which demand exceeds the ability to provide timely care. Waiting-room delay occurs before placement or clinician evaluation. Length of stay begins at arrival and ends at physical departure. Boarding begins only after a disposition decision to admit or transfer.
The Joint Commission defines boarding as holding patients in the emergency department or another temporary location after the decision to admit or transfer has been made.
A hospital can improve its door-to-provider metric even as boarding worsens. It can report a median that improves while the 90th percentile and the number of 24-hour boarders deteriorate. For governance, no single clock is sufficient.
Table 4
Operational definitions and their limits
| Measure | Clock | Primary risk of misuse |
|---|---|---|
| Door to clinician | Arrival to first qualified clinician | Can improve while downstream boarding worsens |
| ED length of stay | Arrival to physical departure | Combines diagnostic work with boarding |
| Boarding time | Admission or transfer decision to physical departure | Timestamp definitions vary |
| Time to inpatient management | Admission decision to inpatient team assumption of care | Does not guarantee a ward bed or ward environment |
| Occupancy | Occupied staffed beds divided by available staffed beds | Licensed-bed denominators conceal staffing constraints |
| Left without being seen | Departure before medical screening milestone | May be sensitive to documentation and workflow |
Recommended practice: report a metric dictionary with numerator, denominator, clock start, clock stop, exclusions, stratifiers, data latency, and accountable owner.
Section 1.2
The moral and clinical category error
The phrase waiting for a bed can imply administrative inconvenience. In reality, a boarded person is already a hospital patient with active inpatient needs. Clinical work does not pause when the bed is unavailable.
The question is whether that work occurs with clear ownership, appropriate staffing, reliable medication and nutrition routines, mobility and toileting support, escalation pathways, privacy, and the capacity to notice deterioration.
Treating boarding as a transportation problem obscures the transfer of clinical risk.
Part three
A transparent integrative review
Evidence and analytic approach
How the evidence was assembled
This manuscript used a rapid integrative review to support a time-sensitive operational and policy decision. Searches targeted boarding and crowding, hospital occupancy, mortality, delirium, older adults, behavioral health, patient experience, discrimination, transfer acceptance, discharge timing, command centers, and system-level interventions.
Sources included PubMed-indexed literature, JAMA Network publications, Health Affairs, CDC and NCHS, CMS Provider Data, AHRQ Patient Safety Network, and Joint Commission standards. The evidence horizon ended August 30, 2026.
Evidence was prioritized in order: systematic reviews and meta-analyses; nationally representative or multicenter cohorts; prospective cohorts; controlled or quasi-experimental evaluations; mixed-methods research; and single-center quality-improvement reports.
Table 5
Evidence taxonomy used in the synthesis
| Class | Typical design | Interpretive rule |
|---|---|---|
| A: strongest available | Systematic review with credible appraisal; national or large multicenter study with robust adjustment; prospective multicenter cohort | Supports the direction of the core judgment; effect magnitude still interpreted cautiously when exposure is nonrandom. |
| B: informative | Single-center cohort; interrupted time series; quasi-experimental or mixed-methods study | Supports mechanism or implementation choice, but transportability must be demonstrated locally. |
| C: hypothesis-generating | Descriptive report; uncontrolled quality-improvement project; expert consensus | Useful for design options, not proof of effectiveness. |
| O: original contribution | Framework, threshold proposal, roadmap, or scorecard created for this report | Must be piloted, measured, and revised; never described as observed evidence. |
The synthesis separates observed evidence, causal evidence, and original modeled proposals.
What the rapid design cannot claim
Limitations follow from the rapid design. The search was not protocol-registered, dual screening was not performed, and a formal risk-of-bias instrument was not applied to every included study.
Publication bias is likely because unsuccessful operational interventions are less likely to be published. Boarding definitions, patient populations, occupancy denominators, and health systems differ.
The report therefore emphasizes convergent findings, mechanisms, and safe implementation rather than pooled effect sizes that the literature cannot support.
Because screening was purposive and not protocol-registered, this manuscript does not claim PRISMA-compliant completeness.
Part four
Two queues, one hospital
Scale and distribution of the national burden
The front-door queue and the post-decision queue
The United States has both a front-door queue and a post-decision queue. The first is visible in waiting rooms and door-to-clinician measures. The second is evident among admitted patients who remain in emergency areas or await transfer.
These queues compete for the same nurses, rooms, monitors, diagnostic services, and attention. When boarders occupy treatment spaces, waiting-room delays increase even if incoming volume remains unchanged. When new emergencies consume emergency resources, inpatient-level care for boarders becomes harder to sustain.
Figure 2
More than half of general-medicine admissions crossed four hours before inpatient management.
Observed evidence. Study period: June 1, 2024, to May 31, 2025; 56 emergency departments in 17 health systems. Time to inpatient management ended at the earlier of a health-system-defined inpatient-management milestone or ED departure. Source: Janke et al. (2026).
Two transitions, not one
The 2026 RESQUE-NET analysis is important because it moves beyond the bed-placement clock. Among 492,135 adults admitted to general medical services, 263,666 waited at least four hours for inpatient management. Delay varied substantially by hospital.
A patient can therefore experience two transitions: the clinical decision to admit, and the later transfer of responsibility to a team structured for inpatient care.
Reducing the second delay is a safety intervention, even when a physical bed cannot be created immediately.
Janke et al., 2026
Figure 3
At the national peak, four-hour boarding affected two in five hospitalizations.
Observed evidence from 46.2 million hospitalizations in a large national data source, 2017-2024. Values show the January 2022 peak, not the average across all months. Source: Janke et al. (2025).
A rising burden, not a pandemic anomaly
The Health Affairs analysis of 46.2 million hospitalizations documented a rising burden rather than an isolated pandemic anomaly. The January 2022 peak is analytically useful because it shows the system’s vulnerability under stress. It should not be used as a current national average.
The deeper lesson is that hospitals operating close to staffed capacity have limited resilience. Once a queue develops, each delayed discharge, bed closure, imaging bottleneck, or post-acute denial propagates backward toward the emergency department.
Leaders should use CMS measures for external comparison while maintaining granular internal boarding clocks, preferably with percentile and threshold distributions.
Section 3.1
The tail matters more than the average
A median is resistant to extreme values, which is useful for stability but dangerous for governance. If most patients move promptly while a clinically vulnerable minority remains for 18 or 30 hours, the median can appear acceptable.
What makes the tail visible:
- The 90th percentile
- The percentage exceeding 4, 8, 12, and 24 hours
- The number of patient-hours spent boarding
Patient-hours are particularly useful because they measure the volume of risk being carried, not only the number of people who cross a threshold.
Part five
The harm signal
Mortality, delirium, delay, and loss of dignity
Causal judgment without a trial
No randomized trial will assign patients to prolonged boarding. Causal judgment must therefore integrate observational design, consistency, temporal ordering, dose-response patterns, plausible mechanisms, and the possibility of residual confounding.
Sicker patients may board longer because they need specialized beds. But longer exposure can also worsen outcomes through delayed surveillance, interrupted medication routines, immobility, sleep deprivation, noise, light, handoff ambiguity, and constrained nursing attention.
The evidence supports a moderate-confidence conclusion that prolonged boarding contributes to harm, while exact causal effect sizes remain uncertain.
Boudi et al., 2020; Lauque et al., 2022; Oskvarek et al., 2026
Figure 4
Older adults who remained in the emergency department overnight had higher in-hospital mortality.
Observed prospective association, not a randomized causal estimate. The adjusted risk ratio was 1.39 (95% CI, 1.07-1.81). Source: Roussel et al. (2023).
Section 4.1
Mortality and deterioration
A systematic review by Boudi and colleagues found an association between boarding and in-hospital mortality, though heterogeneity and observational designs limited certainty. Lauque and colleagues similarly reported an association between longer length of stay and mortality. Earlier cohorts found increasing mortality across longer boarding strata.
The Roussel study adds prospective, multicenter evidence in a population for whom environmental and functional hazards are especially salient. Among 1,598 patients aged 75 or older across 97 French emergency departments, overnight stay was associated with higher mortality, more adverse events, and a longer stay. The association was stronger among patients with limited autonomy.
Generalizability across all U.S. settings is limited, but the study supports prioritizing age, frailty, mobility, cognition, and autonomy in escalation rules rather than relying solely on elapsed time.
Figure 5
Longer mean boarding time accompanied delirium or severe agitation in a large cohort.
Observed association in seven hospitals from 2018-2022. Each additional hour was associated with a small increase in adjusted odds, with a larger absolute risk among older adults and patients with dementia. Source: Joseph et al. (2024).
Section 4.2
Delirium, agitation, and the environment
Joseph and colleagues analyzed 236,169 adults across seven hospitals. Delirium or severe agitation occurred in 27,194 patients, or 11.5%. Age and preexisting dementia amplified risk.
Hallway exposure introduces potential hazards including light, noise, sleep disruption, sensory deprivation or overload, immobility, and reduced access to family and orientation cues.
A hospital that cannot immediately eliminate boarding can still reduce exposure-related harm.
Joseph et al., 2024; van Loveren et al., 2021
Section 4.3
Treatment reliability and spillover
Boarding affects more than admitted patients. As boarders occupy rooms and nurse assignments, new patients wait longer, and more may leave before evaluation. Diagnostic and treatment delays can affect time-sensitive conditions.
Medication delays and adverse events have been reported among boarded patients, and crowding degrades the working conditions in which clinicians notice changes, communicate, and recover from interruptions.
The spillover is also economic. A room used for a boarded inpatient cannot perform its emergency function. Ambulances queue or divert. Elective surgery competes for beds. Staff work under higher cognitive load and greater exposure to verbal or physical aggression.
The cost is not captured by a single emergency department budget, because risk and lost capacity are distributed across departments, payers, patients, and the regional network.
“
The correct counterfactual is not an empty emergency department. It is the same patient receiving the same needed care in a setting with clear ownership, appropriate staffing, reliable routines, privacy, mobility, and escalation capacity.
Part six
Operational physics
Why queues accelerate near saturation
Input, throughput, output
The input-throughput-output model remains useful because it prevents the front door from becoming the sole explanatory frame.
Input is demand for emergency care, including volume, acuity, primary care access, behavioral health needs, and regional transfers. Throughput includes triage, clinician evaluation, diagnostics, consultation, decision-making, and handoff. Output includes the availability of the next appropriate setting: a staffed inpatient bed, an intensive care unit, a behavioral health facility, a home service, a skilled nursing facility, or an accepting regional hospital.
Asplin et al., 2003
Figure 6
The visible queue is created by the interaction of input, throughput, and output constraints.
Original synthesis. The conceptual structure draws on Asplin et al. (2003), queueing theory, and contemporary evidence. It is explanatory, not a validated prediction model.
Little's Law and the shape of a queue
Little’s Law provides a simple operational identity: work in process equals throughput multiplied by cycle time. The number of boarded patients rises when admissions continue and the cycle time to an appropriate destination lengthens.
Adding another waiting area changes the location of work in process. It does not reduce cycle time unless care, decisions, or destinations change.
Queueing systems behave nonlinearly near saturation. When demand and service time are variable, waiting rises sharply as utilization approaches full capacity. Hospitals are not single-server queues, and a universal occupancy threshold is too crude. Still, empirical work has repeatedly found worse boarding at high occupancy, often above 85% to 90%.
The operational denominator must be staffed and usable beds, not licensed beds that cannot accept a patient.
Section 5.1
Temporal mismatch
Hospitals often admit continuously but discharge in a narrower afternoon window. Rounds, consultations, transportation, pharmacy, equipment, and family readiness occur sequentially. The bed is unavailable during the morning demand peak even when the same patient leaves later that day.
The remedy is to move uncertainty upstream. The expected discharge date should be estimated early, updated daily, and linked to named barriers. Tests required for discharge should be identified before the day of departure. Transport, medications, durable equipment, caregiver training, and post-acute placement should be treated as part of the clinical workflow, not as last-minute logistics.
The goal is not discharge by noon at any cost. It is a predictable match between available staffed capacity and the next wave of demand, without increasing readmissions, mortality, or unsafe transitions.
Section 5.2
Responsibility mismatch
A patient may be admitted in one information system state, managed by an emergency clinician, accepted by a hospitalist, assigned to a specialty service, and awaiting a bed controlled by another office.
When responsibility changes without a standardized handoff, the patient can be everyone’s concern and no one team’s full responsibility.
The 2026 multicenter study shows that delay to inpatient management is a separate target from physical departure. Hospitals should define who writes orders, reviews results, responds to deterioration, communicates with the family, and performs medication reconciliation at every stage.
Part seven
Inequity is built into the queue
Equity and regional access
Queues are not socially neutral
Patients with behavioral health needs, limited English proficiency, disability, homelessness, complex placement needs, or weak social support may require resources that are scarce or administratively fragmented.
Rural patients often depend on transfer for specialty care. When referral hospitals are crowded, internal boarding becomes a barrier to regional access.
A queue at one tertiary center can delay definitive care across hundreds of miles.
Figure 7
Prolonged boarding is associated with both a worse experience and reduced acceptance of regional transfers.
Observed associations from distinct studies. Odds ratios above one indicate higher reported discrimination or dissatisfaction. Odds ratios below one indicate lower transfer acceptance. Do not compare magnitudes across outcomes as if they share a common scale. Sources: Olson et al. (2024); Greenwood-Ericksen et al. (2025).
Experience, dignity, and transfer access
In a cross-sectional study of 525 adults boarding during general-medicine admission, those boarding for at least 24 hours had higher adjusted odds of reporting discrimination (1.84) and dissatisfaction (1.77) than those boarding less than four hours. Patients from marginalized racial and ethnic groups reported more concerns about courtesy, respect, intelligence, and involvement in decisions.
Greenwood-Ericksen and colleagues analyzed 26,020 interhospital transfer requests. Higher boarding and inpatient census at the receiving hospital were associated with lower transfer acceptance. Requests from rural settings had an adjusted odds ratio of acceptance of 0.66 compared with urban requests.
This does not prove boarding alone caused each denial. It does show that internal hospital strain and regional access are coupled.
Sections 6.1 and 6.2
Behavioral health and equity measurement
Behavioral health boarding reveals the limits of a bed-only explanation. Patients may wait because inpatient psychiatric beds, crisis stabilization, intensive outpatient programs, community clinics, substance-use treatment, and supportive housing are insufficient. Emergency environments can worsen distress, isolation, and agitation.
Every boarding metric should be stratified by age, race and ethnicity using locally defined categories, language, disability, behavioral health status, payer, housing instability, rurality, transfer origin, and service line.
The purpose is not to rank patients or infer biology from social categories. It is to determine whether the operating system distributes delay, dignity, surveillance, or access to destinations unequally.
Part eight
Why familiar remedies disappoint
Failure analysis of the standard responses
The central implementation question
Hospitals often respond to crowding with visible, local changes: add hallway spaces, accelerate triage, build another waiting area, create a command center, set a discharge-by-noon target, or purchase predictive software.
Each can be useful in a defined constraint. None is a universal cure.
The central implementation question is whether the intervention changes a verified bottleneck, transfers risk, or merely changes where delay is counted.
Table 6
Intervention logic and failure modes
| Common response | Why it fails when used alone | Required safeguard |
|---|---|---|
| Provider in triage or split flow | Does not create staffed inpatient capacity; can draw clinicians from higher-acuity care | Track admitted boarding, adverse events, and staffing balance |
| More ED treatment spaces | Boarders may rapidly occupy the added space | Demonstrate structural space deficit and fund staffing |
| Observation unit | Becomes another holding area without selection criteria and exit discipline | Diagnosis-specific protocols, length limits, and escalation |
| Command center | Dashboards without decision rights digitize delay | Named authority, standard work, service-level commitments |
The table distinguishes the intended mechanism from common implementation failure.
Table 6, continued
Intervention logic and failure modes
| Common response | Why it fails when used alone | Required safeguard |
|---|---|---|
| Discharge by noon | Can produce gaming or unsafe haste; not every patient can or should leave early | Readmission, mortality, experience, and destination balancing measures |
| Full-capacity protocol | May relocate risk to inpatient hallways and raise dignity or staffing concerns | Strict criteria, executive activation, safety evaluation, and time limits |
| Predictive analytics | Predictions do not change capacity or accountability | Action thresholds, calibration, bias review, and downtime plans |
Local context determines whether any intervention is appropriate.
What the newest systematic review found
The newest systematic review located 226 studies published from 2018 through January 2025.
- 83% were retrospective
- 62% were rated low quality
- 35% were rated acceptable quality
Provider-in-triage, nurse-initiated orders, and split flow showed promise for selected front-end outcomes. Output interventions required system-level coordination.
The evidence base supports action, but not false precision about which package will produce a specific effect in every hospital.
Oskvarek et al., 2026
Part nine
Act, but know what is known
Critical appraisal of the intervention evidence
Constraint-driven, multifaceted work
The strongest intervention strategy is not a single tactic. It is constraint-driven, multifaceted work with balancing measures.
Mixed-methods research has found that high-performing hospitals use executive involvement, data transparency, operational accountability, and multiple coordinated strategies.
Systematic reviews of hospital length-of-stay interventions show inconsistent effects across heterogeneous populations, though selected clinical pathways and case management programs can reduce length of stay. Single-center programs report large improvements, but their estimates are vulnerable to secular trends, concurrent changes, and local context.
Chang et al., 2018; Siddique et al., 2021; Al Harbi et al., 2024
Table 7
Evidence-to-decision summary
| Intervention domain | Confidence | Decision implication |
|---|---|---|
| Front-end clinical flow | Moderate for selected process measures | Use when front-end delay is verified; do not claim it solves admitted boarding |
| Active bed management | Low to moderate | Pair visibility with authority and service-level standard work |
| Capacity command centers | Low | Treat as infrastructure, not the intervention itself |
| Early discharge and progression | Low to moderate | Redesign the full discharge pathway; avoid target gaming |
Confidence is intentionally conservative because many operational studies are observational, single-center, and bundled.
Table 7, continued
Evidence-to-decision summary
| Intervention domain | Confidence | Decision implication |
|---|---|---|
| Case management | Moderate for some high-risk populations; low for generalizability | Target defined barriers and evaluate equity and readmission |
| Full-capacity protocols | Low | Use only with patient-safety design and comparative evaluation |
| ED-based critical care units | Low to moderate | Consider where ICU boarding is recurrent and staffing can match acuity |
| Regional transfer coordination | Low to moderate | Create transparent regional rules and load-balancing mechanisms |
Mechanistic strength is not the same as comparative outcome evidence.
Section 8.1
Immediate safety before throughput
Hospitals should not wait for a perfect redesign of flow before protecting patients already boarding. The safety bundle begins when the disposition decision is recorded and scales with time, acuity, frailty, environment, and handoff status.
- Assign an accountable attending and nurse at every transition, with closed-loop handoff and coverage visibility
- Complete medication reconciliation and administer time-critical medications without waiting for physical bed placement
- Use scheduled reassessment and deterioration criteria, including rapid response or critical care consultation when indicated
- Screen for delirium, falls, pressure injury, suicide risk, pain, nutrition, hydration, toileting, mobility, and sensory needs
- Protect sleep and dignity: lights, noise, privacy, family communication, interpreter access, and avoidance of hallway exposure whenever possible
- Escalate older adults, patients with dementia or limited autonomy, critically ill patients, children, and behavioral health patients using stricter thresholds
Section 8.2
Structural improvement after stabilization
Once immediate harm controls are in place, leaders should map the queue by service, shift, day, location, destination, and delay reason.
The purpose is not a long list of causes. It is to identify the few constraints producing most patient-hours.
One hospital may have delayed environmental services. Another may close staffed medical beds because of nurse vacancies. Another may have a post-acute placement crisis. Another may schedule elective demand without reserving resilience for unscheduled care.
The same dashboard cannot substitute for local diagnosis.
Part ten
The Enterprise Flow Reliability System
An original operating model of five coupled control loops
Figure 8
Five coupled control loops organize an enterprise response.
Original contribution. The EFRS is a proposed implementation model and has not been prospectively validated.
Loops 1 and 2
Sense and Stabilize
Sense combines near-term demand forecasting with current staffed capacity and known downstream constraints. The unit of analysis is not an empty bed. It is an accept-ready care position with the required nurse, monitoring, isolation, equipment, and service capability.
Accuracy matters, but operational usefulness matters more: every forecast state must map to a preauthorized action.
Stabilize activates the boarded-patient safety bundle, clarifies clinical ownership, and prioritizes patients whose risk increases steeply over time. The target is not only bed movement. It is reliable care while movement is delayed.
This loop should be audited through missed medications, delayed reassessment, deterioration, delirium, falls, pressure injuries, elopement, restraint, privacy events, and patient communication.
Loops 3 and 4
Synchronize and Release
Synchronize eliminates the gaps between decision, order, handoff, bed assignment, environmental services, transport, and physical arrival. Diagnostic and consult service-level agreements should specify response times, escalation, and exception pathways.
The transition to inpatient management should occur at a defined clinical milestone and not depend entirely on physical relocation.
Release starts at admission, not discharge morning. Each patient receives an expected date and destination, a daily readiness assessment, and named barriers. Weekend and evening capability should match the hospital’s seven-day demand profile.
Hospitals should measure the interval from medically ready to actual departure, stratified by destination and payer, without pressuring clinicians to declare readiness prematurely.
Loop 5
Learn and govern
Learn and govern assigns executive decision rights, reviews variation, tests countermeasures, and reports performance to the governing board.
Every severe boarding episode should generate a brief learning review focused on system conditions rather than individual blame. Recurrent 24-hour boarding, preventable deterioration, or unsafe hallway care should be treated as a serious organizational signal.
Equity review is not a separate annual project. It is part of every measure-and-improvement cycle.
Table 8
EFRS control-loop specification
| Loop | Executive question | Primary measure |
|---|---|---|
| Sense | What demand-capacity mismatch will occur in the next 4, 12, and 24 hours? | Forecast error; staffed-bed deficit; transfer pressure |
| Stabilize | Are boarded patients receiving reliable inpatient-level protections now? | Bundle reliability; deterioration; delirium; missed medication |
| Synchronize | Where is responsibility or movement waiting for another function? | Time to inpatient management; diagnostic and consult delay |
| Release | Which patients are ready next, and what prevents safe departure? | Medically-ready-to-departure time; discharge timing |
| Learn + govern | Who can remove the constraint today and prevent recurrence? | Tail boarding; patient-hours; variation; equity gaps |
A hospital should be able to specify, for each loop, the action, the responsible role, the trigger, the expected mechanism, the measure, and the balancing risk.
Part eleven
365 days to a different operating system
Roadmap and escalation architecture
Figure 9
The roadmap moves from visibility to institutional reliability.
Original contribution. Sequence should be adapted to local readiness, labor agreements, regulatory requirements, and patient populations.
Section 10.1
Days 0 to 30: make risk visible
- Appoint one executive owner with written authority across emergency, inpatient, diagnostic, environmental, transport, and case-management functions
- Publish the metric dictionary, including physical-location and clinical-ownership clocks
- Baseline median, 90th percentile, percentages beyond 4, 8, 12, and 24 hours, and boarding patient-hours by service and shift
- Activate the boarded-patient safety bundle and create a daily list of high-risk boarders
- Establish two daily enterprise flow huddles timed before major demand and discharge decisions
- Review all current queues and labels to identify metric displacement or hidden boarding locations
Sections 10.2 and 10.3
Days 31 to 90, and beyond
- Implement a standardized transition to inpatient management, with named responsibility for orders, results, reassessment, and family communication
- Create a small reason taxonomy for delays and validate it through observation, not only electronic selections
- Use forecasting to trigger preauthorized staffing, bed, surge, elective, and transfer actions
- Select the top three constraints by boarding patient-hours and assign service-level improvement compacts
- Add balancing measures for readmission, mortality, adverse events, elective cancellations, patient experience, staff injury, and equity
The middle phase redesigns discharge readiness, diagnostic and consult response, bed turnaround, nurse staffing, behavioral health pathways, and post-acute coordination. The later phase embeds flow in the quality plan, capital planning, medical staff accountability, labor strategy, and board oversight.
Figure 10
Risk-based escalation prevents normalization of severe boarding.
Original contribution. Thresholds are illustrative and require local validation. Use stricter criteria when clinical acuity, age, cognitive status, behavioral health risk, or the environment increases the risk of harm.
Table 9
Illustrative escalation standard
| Level | Illustrative trigger | Required response | Accountable role |
|---|---|---|---|
| 0: Stable | Performance within local control limits and no forecasted deficit | Routine forecasting, huddles, and safety-bundle readiness | Flow leader |
| 1: Strained | Rising >4-hour rate, forecasted bed or staffing deficit | Service owners remove near-term barriers; protect high-risk patients | Administrator on call |
| 2: Severe | >8-hour boarders, high staffed occupancy, delayed care, or transfer pressure | Executive huddle; open authorized capacity; inpatient teams assume care promptly | CMO/CNO or designee |
| 3: Critical | >12-hour tail, any >24-hour boarder, serious safety event, or failure of regional access | Incident command; consider elective modification; systemwide staffing and transfer actions; board notification | CEO or incident commander |
Triggers should be tested against local outcomes and false-alarm burden. Clinical judgment overrides a lower operational level.
Section 10.4
Elapsed time is necessary but insufficient
Escalation should also consider staffed occupancy, forecasted deficits, the number of critical care boarders, time to inpatient management, waiting-room acuity, ambulance offload, transfer requests, care delays, and any safety event.
Level 3 should activate a hospital incident structure or an equivalent executive control, with authority to adjust elective schedules, open staffed surge capacity, redistribute clinical teams, expedite transfers, and communicate transparently with patients and regional partners.
Part twelve
A board scorecard that cannot be gamed easily
Measures and governance
What a useful scorecard does
A useful scorecard links outcomes, processes, and balancing measures. It reports the distribution, not only the median. It shows clinical ownership and physical location separately. It stratifies equity. It displays trends with statistical process control when volume permits.
Most important, every measure maps to a decision. A dashboard that cannot change staffing, timing, or service behavior is an archive.
Table 10
Enterprise flow scorecard
| Domain | Core metric | Stratification |
|---|---|---|
| Access | Door to qualified clinician; LWBS | Shift, acuity, age, race/ethnicity, language |
| Boarding | Median, P90, >4/>8/>12/>24 h, patient-hours | Service, destination, unit, day, shift |
| Clinical ownership | Time to inpatient management | Hospital, service, shift, location |
| Safety | Deterioration, delirium, falls, missed medication, pressure injury | Age, cognition, acuity, environment |
| Capacity | Staffed occupancy and closed-bed hours | Service, skill mix, isolation capability |
P90 means 90th percentile. LWBS means left without being seen.
Table 10, continued
Enterprise flow scorecard
| Domain | Core metric | Stratification |
|---|---|---|
| Progression | Expected-date reliability; medically-ready-to-departure | Service, destination, payer, barrier |
| Regional access | Transfer acceptance and time to definitive site | Rurality, diagnosis, payer, distance |
| Experience and equity | Communication, dignity, discrimination, delay gaps | Patient-defined demographic groups |
| Workforce | Injury, violence, missed breaks, turnover intention | Role, shift, department |
| Balancing | Readmission, mortality, elective cancellations, return visits | Service and intervention exposure |
Metric definitions should be version-controlled and audited against source records.
Sections 11.1 and 11.2
Board oversight and analytic design
The governing board should receive a monthly flow and access report within the quality agenda, not solely the operations or finance agenda. The report should include severe-event narratives, equity gaps, corrective actions, and whether leaders met escalation obligations.
The Joint Commission’s 2026 National Performance Goal requires hospitals to measure and set goals for patient flow components, manage boarding, review results, and take action when goals are not met. It explicitly names the governing body, chief executive, senior managers, nurse executive, medical staff, and clinical leaders in accountability.
Improvement evaluation should use an interrupted time series or stepped-implementation design when feasible. Report absolute times and rates, not only relative improvement. Track implementation fidelity so a null result can be distinguished from failure to deliver the intervention.
Part thirteen
From hospital to region
Policy implications, research agenda, and conclusion
Hospital leaders control only part of the chain
Post-acute workforce shortages, behavioral health capacity, ambulance availability, payer authorization, home services, and regional transfer rules determine whether patients can move safely. A credible national strategy must align measurement and incentives across settings rather than penalize the emergency department for downstream scarcity.
Public reporting should distinguish total length of stay, boarding, and time to inpatient management. Measures should include tail performance and stratification, not only a median. Payment systems should avoid creating incentives to relabel locations or accelerate unsafe discharge.
Workforce policy is capacity policy. A licensed bed without nurses, respiratory therapists, environmental services, pharmacy support, transport, and medical coverage is not available capacity.
Table 11
Leadership and policy agenda
| Actor | Priority action | Accountability signal |
|---|---|---|
| Governing board | Classify severe boarding as a quality and safety risk | Regular review of tail, harm, equity, and corrective action |
| Chief executive | Create enterprise decision rights and remove service silos | Documented response to escalation and recurrent constraints |
| Clinical leaders | Standardize ownership, reassessment, and service-level commitments | Time to inpatient management and care-bundle reliability |
| Payers | Reduce avoidable authorization and placement delay | Medically-ready-to-departure time by payer and destination |
| Regional authorities | Coordinate transfers and surge load balancing | Acceptance, time to definitive care, and rural equity |
| Accreditors and regulators | Align definitions and require action on severe tails | Comparable measures plus evidence of governance response |
Researchers should evaluate multicomponent system interventions with causal designs, reporting protocolized outcomes, fidelity, equity, and balancing measures.
Section 13.1
What must be tested next
The evidence is sufficient to reject complacency but insufficient to identify one universally superior intervention.
- Does earlier assumption of care by inpatient teams reduce adverse events when physical bed placement remains delayed?
- Which components of a boarded-patient safety bundle reduce delirium, missed medications, deterioration, restraint, falls, and patient distress?
- What staffed-occupancy thresholds predict nonlinear boarding risk by service line, hospital type, and demand variability?
- Do command centers improve outcomes beyond the decision rights, staffing, and standard work that accompany them?
- Can regional load balancing improve time to definitive care without destabilizing receiving hospitals?
- Which public measures resist relabeling and accurately reflect both typical performance and extreme tails?
Section 13.2
Limitations of this manuscript
This is not a de novo systematic review, meta-analysis, or original patient-level study. Search and selection may have missed relevant literature.
Current national reporting uses nonidentical populations and clocks, preventing direct combination. International outcome evidence may not fully transport to the United States. Operational interventions are sensitive to local staffing, labor agreements, facility design, information systems, and referral networks.
The EFRS, thresholds, roadmap, and scorecard are original proposals and have not been prospectively validated.
“
The queue is a clinical condition of the hospital. Its treatment is reliable enterprise flow, and its first duty is to protect the people already inside it.
Section 13.3
Conclusion
The national emergency department boarding crisis is not solved by asking emergency clinicians to work faster inside a saturated system. The best available evidence shows a large, persistent post-decision queue associated with mortality, delirium, medication delay, poor experience, workforce strain, and reduced regional access.
The most defensible response is both immediate and structural: protect every boarded patient now, transfer clinical ownership reliably, govern staffed capacity as a system, advance inpatient progression and safe discharge, coordinate post-acute and regional resources, and hold executives and boards accountable for the tail of delay and its unequal distribution.
Act with epistemic discipline
- Associations are not randomized effects.
- Single-center success is not universal proof.
- Predictive software is not capacity.
- A bed is not usable without staff.
- A median is not the patient at 24 hours.
These distinctions do not weaken the case for action. They make the action safer, more honest, and more likely to endure.
Appendices
Definitions, diagnostics, and reproducibility
Appendices A, B, and C
Table A1
Minimum data dictionary
| Term | Recommended local definition |
|---|---|
| Boarding start | Timestamp of documented decision to admit or transfer. Preserve source event and version history. |
| Physical boarding end | Timestamp the patient physically leaves the emergency department or temporary boarding location for the definitive care setting. |
| Inpatient-management start | Timestamp when an identified inpatient clinician or team explicitly assumes primary management responsibility under a standardized handoff. |
| Boarding patient-hour | One admitted or transferring patient remaining in a boarding location for 1 hour after boarding starts. |
| Staffed usable bed | A bed that can accept a patient at that time with required nursing, medical coverage, equipment, monitoring, isolation, and service capability. |
| Severe boarding event | Locally defined combination of tail duration, high-risk patient factors, unsafe environment, care delay, or deterioration requiring executive review. |
Appendix A. Operational definitions.
Table B1
Executive diagnostic checklist
| Diagnostic question | Red flag |
|---|---|
| Where does the clock start and stop? | Different departments use different clocks |
| Who owns the patient clinically? | No explicit inpatient assumption of care |
| Which constraints generate most patient-hours? | Long unprioritized list without magnitude |
| Are beds physically or operationally unavailable? | Capacity plan counts unstaffed beds |
| Is delay temporal? | Afternoon discharge peak follows morning demand |
| Who is harmed disproportionately? | No equity or high-risk stratification |
| What happens at escalation? | Alerts repeat without a change in authority or resources |
| Are interventions moving the metric? | Improvement in reported time without improved patient environment |
Appendix B. Diagnostic questions before capital or technology investment.
Appendix C
Reproducibility and reporting notes
All quantitative figures in this manuscript were recreated from values reported in the cited primary sources or federal datasets. No patient-level records were downloaded or analyzed.
Figures 6, 8, 9, and 10, as well as the implementation tables, are original conceptual outputs and are explicitly labeled as such. The report avoids combining estimates with different denominators into a synthetic national rate.
A future reproducible analysis should preserve source extracts, variable dictionaries, code, software versions, date of data access, and a machine-readable table linking each figure value to its source location. Hospital implementations should pre-register evaluation outcomes where feasible, define exclusions before analysis, and retain negative or null results to reduce publication bias.
Part I
Selected references
- Asplin, B. R., et al. (2003). A conceptual model of emergency department crowding. Annals of Emergency Medicine, 42(2), 173-180.
- Boudi, Z., et al. (2020). Association between boarding in the emergency department and in-hospital mortality: A systematic review. PLOS ONE, 15(4), e0231253.
- Centers for Disease Control and Prevention, NCHS. (2024). National Hospital Ambulatory Medical Care Survey: 2022 emergency department summary tables.
- Centers for Medicare & Medicaid Services. (2026). Hospitals: Timely and effective care. Provider Data Catalog.
- Chang, A. M., et al. (2018). Hospital strategies for reducing emergency department crowding: A mixed-methods study. Annals of Emergency Medicine, 71(4), 497-505.e4.
- Greenwood-Ericksen, M., et al. (2025). Emergency department boarding, inpatient census, and interhospital transfer acceptances. JAMA Network Open, 8(5), e2512299.
- Janke, A. T., et al. (2026). Time to inpatient management for boarding emergency department patients. JAMA. Advance online publication.
- Janke, A. T., et al. (2025). Hospital boarding of patients in the emergency department as a delayed form of admission. Health Affairs, 44.
Part II
Selected references
- Joseph, J. W., et al. (2024). Boarding duration in the emergency department and inpatient delirium and severe agitation. JAMA Network Open, 7(6), e2416343.
- Lauque, D., et al. (2022). Length-of-stay in the emergency department and in-hospital mortality: A systematic review and meta-analysis. Journal of Clinical Medicine, 12(1), 32.
- Olson, R. M., et al. (2024). Prolonged boarding and racial discrimination and dissatisfaction among emergency department patients. JAMA Network Open, 7(9), e2433429.
- Oskvarek, J. J., et al. (2026). Emergency department crowding in the modern era: A systematic review, 2018-2025. Clinical and Experimental Emergency Medicine, 13(2), 119-139.
- Roussel, M., et al. (2023). Overnight stay in the emergency department and mortality in older patients. JAMA Internal Medicine, 183(12), 1378-1385.
- Siddique, S. M., et al. (2021). Interventions to reduce hospital length of stay in high-risk populations: A systematic review. JAMA Network Open, 4(9), e2125846.
- The Joint Commission. (2025). National Performance Goals effective January 2026 for the Hospital Program.
- van Loveren, K., et al. (2021). Increased emergency department hallway length of stay is associated with development of delirium. Western Journal of Emergency Medicine, 22(3), 726-735.
The full reference list appears in the published manuscript.
About the author
Kelly Emrick
DHSc, PhD, MBA, BSRT(ARRT)R
Healthcare executive, author, and registered radiologic technologist. Her research and practice span imaging service line transformation, population health, fiscal management, predictive analytics, value-based care, and adaptive leadership.
This report is an independent scholarly synthesis prepared for healthcare leaders, clinicians, policymakers, and researchers. Correspondence information can be inserted before submission to a selected journal.
Beyond the Emergency Waiting Room
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
2026
National Healthcare Operations Study | 2026
Beyond the Emergency Waiting Room
The National Emergency Department Waiting and Boarding Crisis
Boarding is not a location problem. It is accumulated clinical, operational, and social work that a hospital shifts into its least protected space.
An integrative evidence synthesis, policy analysis, and enterprise operating model for safer hospital flow.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Prepared for journal submission · Evidence through August 30, 2026
Research report | Policy status: proposed
Who Owns Outpatient Healthcare?
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
August 2026
Who Owns Outpatient
Healthcare?
Site-neutral payment, hospital cross-subsidies, and the struggle toredesign the economic architecture of American ambulatory care
Kelly Emrick
Independent evidence synthesis and scenario analysis
POLICY STATUS: PROPOSED
Who Owns Outpatient Healthcare?
Site-neutral payment, hospital cross-subsidies, and the struggle to redesign the economic architecture of American ambulatory care.
What is inside
- 1The Policy Has Become a Fight Over the Outpatient Enterprise
- 2Why the First Site-Neutral Reform Barely Moved the Market
- 3Cross-Subsidy Is Real, but It Is Not a Blank Check
- 4Radiology Is the Clearest Test Case
- 5The Ownership Premium and the Consolidation Flywheel
- 6What Happens When the Subsidy Is Removed?
- 7Rural and Safety-Net Hospitals Require Precision, Not Rhetoric
- 8A Better Policy Architecture
- 9Strategic Conclusions
Followed by methods, limitations, an integrity statement, and references.
Purpose
This report examines site-neutral payment as a contest over ownership, price, and the financing of hospital public goods.
It asks four linked questions.
- Why does the same ambulatory service command different payment when a hospital owns the site?
- Which portions of those differentials reflect legitimate hospital obligations?
- Which portions primarily reward market power and consolidation?
- What payment architecture could reduce ownership-based distortions without hollowing out emergency, rural, teaching, and safety-net capacity?
Design
The analysis integrates current federal policy, peer-reviewed studies of site-based payment, hospital-physician integration, commercial prices, rural hospital margins, 340B economics, and hospital cross-subsidization.
It includes an independent scenario framework but no patient-level claims analysis. Estimates reported from CMS, MedPAC, and peer-reviewed studies remain attributable to those sources; the illustrative margin bridge is not a forecast or formal budget score.
Primary finding
The strongest evidence supports a dual conclusion.
First, payment differentials tied only to hospital ownership raise spending, increase beneficiary liability, and help finance consolidation. Hospital acquisition of physician practices has repeatedly been associated with higher prices, while quality gains are inconsistent.
Second, many hospitals do rely on high-margin outpatient lines, commercial contracts, drug spreads, and facility payments to support services that are structurally underfunded.
Eliminating these margins without an explicit replacement mechanism does not abolish the social obligation. It merely forces hospitals to ration, close, relocate, or reprice it elsewhere.
Decision implication
A defensible reform should be service-specific, risk-adjusted, phased, transparent, and paired with direct support for public goods.
The policy should neutralize payment for low-complexity, substitutable services while preserving explicit supplements for rural access, standby capacity, trauma, teaching, uncompensated care, and clinically justified hospital complexity.
Cross-subsidy should be converted from an opaque ownership premium into an accountable public financing mechanism.
THE GOVERNING PRINCIPLE
Pay the same base rate for the same clinically comparable service, then pay separately and transparently for measurable hospital obligations.
Ownership should not be the basis of the subsidy formula.
The Policy Has Become a Fight Over the Outpatient Enterprise
Why a payment rule became a contest over the ambulatory front door.
The hospital is no longer only a place of beds
For decades, the hospital was understood primarily as a place of beds, operating rooms, and emergency care. That description is now economically incomplete.
Much of modern hospital strategy is organized around ambulatory surgery, oncology infusion, diagnostic imaging, physician practices, cardiovascular testing, hospital-based clinics, and off-campus departments.
These services do more than deliver care. They generate referrals, protect downstream volume, strengthen payer leverage, and, in many systems, produce margins used to sustain less profitable obligations.
Two payments for one service
Site-neutral payment challenges this architecture at its foundation. Under Medicare, a service delivered in a hospital outpatient department can receive both a professional payment and a facility payment under the Outpatient Prospective Payment System, whereas a clinically similar service in an independent physician office is generally paid through the Physician Fee Schedule.
Commercial payment differentials can be larger.
The policy dispute therefore asks whether Medicare and private purchasers should pay more because the service is hospital-owned, or only when the patient, resources, and readiness requirements justify the difference.
Reimbursement is capitalized into acquisition value
The ownership question matters because reimbursement is capitalized into acquisition value. If hospital ownership converts the same visit, scan, or procedure into a higher-paying encounter, the payment system creates a premium for integration.
Capps, Dranove, and Ody (2018) found that prices for services provided by acquired physicians increased an average of 14.1% after acquisition. Neprash et al. (2015), Baker et al. (2014), and Lin et al. (2021) likewise associated integration with higher spending or prices.
A 2025 systematic review of 43 hospital-physician integration studies concluded that integration was associated with higher costs while effects on quality and utilization remained unclear (Harris et al., 2025).
WHAT IS ACTUALLY BEING CONTESTED
Site neutrality is a struggle over who controls the ambulatory front door, who captures the facility margin, who owns referral pathways, and whether profitable services remain attached to the hospital enterprise.
The 2027 policy frontier
CMS has proposed extending its volume-control method to imaging without contrast delivered in excepted off-campus provider-based departments. The proposed policy would apply Physician Fee Schedule-equivalent rates to affected services and exempt rural Sole Community Hospitals.
CMS estimates approximately $260 million in first-year Part B expenditure reductions, including about $190 million in Medicare savings and $70 million in lower beneficiary premiums. Beneficiary cost-sharing obligations would decline by an estimated $70 million (CMS, 2026a).
A second front: the 340B drug payment change
The same proposed rule attacks a second source of outpatient margin. CMS proposes paying for 340B-acquired drugs at average sales price minus 33.4%, estimating $4.55 billion less in Original Medicare drug payments and $1.15 billion less in beneficiary drug payments in the first year.
Because the change is budget neutral, the payment system would redistribute an equivalent amount toward non-drug outpatient services.
This is not a uniform hospital cut. It is a reallocation across service lines and institutions, with different winners and losers depending on drug intensity, outpatient mix, eligibility, and organizational structure (CMS, 2026a).
Figure 1
Policy scope determines the financial shock. Estimated annual payment effects depend primarily on policy scope.
Published figure reproduced from the report. Sources: CMS (2026a); Lou et al. (2025).
Reading Figure 1
The two panels answer different questions, which is why they use different scales.
- The narrow panelSeparates two figures that sit close together. The gap between the $212 million lower bound and the $260 million current proposal is $48 million, which disappears entirely on a scale wide enough to show billions.
- The full-range panelPlaces the current proposal against the broadest options under discussion. The upper estimate is more than 28 times the imaging proposal.
- The governing pointScope, not the existence of site neutrality, determines the size of the financial effect. Arguments about whether the policy is large or small are usually arguments about which option is being priced.
Why the First Site-Neutral Reform Barely Moved the Market
A policy that touched 1.5% of spending was never going to restructure an industry.
The 2015 reform and what it actually reached
The Bipartisan Budget Act of 2015 applied site-neutral rates to certain newly established off-campus departments, while grandfathering existing sites and creating other exceptions.
Post et al. (2025) examined Medicare claims from 2013 through 2020 and found that only 1.5% of outpatient department spending during 2017-2020 occurred at facilities subject to site-neutral payments. Counties affected by the policy did not experience a statistically significant reduction in hospital-physician integration.
The mechanism was too narrow to materially change acquisition incentives or total spending.
What that result does and does not prove
This result is politically important. Opponents may cite the limited savings as evidence that site neutrality is ineffective.
A more accurate interpretation is that exceptions shielded nearly the entire payment base. The study does not demonstrate that equalizing rates cannot affect behavior. It demonstrates that a policy touching 1.5% of spending is unlikely to restructure an industry.
The study does not demonstrate that equalizing rates cannot affect behavior. It demonstrates that a policy touching 1.5% of spending is unlikely to restructure an industry.
Scope changes the answer
Lou et al. (2025) show how scope changes the answer. Across three policy options, estimated annual Medicare reductions ranged from $212 million to $7.36 billion.
Small and rural hospitals absorbed the smallest shares because their affected outpatient volume was lower, and effects varied little across beneficiary groups defined by age or dual eligibility.
The analysis weakens claims that every version of site neutrality necessarily concentrates harm among vulnerable hospitals or beneficiaries, while leaving open the question of hospital-specific exposure within each category.
Four ways to design the policy
| Policy design | Economic mechanism | Likely result | Principal risk |
|---|---|---|---|
| Narrow, prospective-only | Applies lower rates only to new or limited off-campus sites | Small savings; weak change in acquisition incentives | Creates grandfathered winners and strategic avoidance |
| Service-specific neutrality | Targets substitutable visits, tests, imaging, and procedures | Moderate savings; stronger competition by site | May remove margin from departments supporting other services |
| Broad HOPD neutrality | Aligns a wide range of HOPD services with office or ASC rates | Large payment redistribution and lower cost sharing | Potential restructuring, closures, or service migration |
| Neutral base plus explicit supplements | Equal base rate plus separate public-good payments | Reduces ownership premium while protecting named obligations | Requires credible cost accounting and political discipline |
The fourth design is the architecture this report recommends.
Cross-Subsidy Is Real, but It Is Not a Blank Check
Hospitals may be right that margins finance public goods and still be wrong about how to finance them.
The financing claim, and the evidence for it
Hospitals often argue that higher payments for profitable outpatient services support emergency readiness, trauma care, teaching, behavioral health, obstetrics, uncompensated care, and 24-hour infrastructure. At an organizational level, this claim is plausible and frequently true.
David, Helmchen, and Henderson (2014) found evidence consistent with cross-subsidization among hospital service lines.
The 2026 analysis of 1,384 Critical Access Hospitals by Jia and colleagues provides contemporary evidence of payer-level asymmetry: average commercial margins were 15.6% in 2023, Medicare fee-for-service margins were near break-even at 0.5%, and Medicaid margins remained deeply negative at -36.3%. Overall operating margins declined from 8.0% in 2011 to 3.3% in 2023.
Figure 2
Commercial margins mask the public-payer deficit. Critical Access Hospital payer margins demonstrate the financial logic behind cross-subsidy claims.
Published figure reproduced from the report. Source: Jia et al. (2026), 15,819 hospital-year observations.
Three cautions
- Cross-subsidy is not cost shifting.Frakt (2011) concluded that careful empirical studies generally found cost shifting to be limited and contingent on market power, not an automatic dollar-for-dollar response to public payment shortfalls. Hospitals able to command high commercial prices may already be exercising bargaining leverage independent of Medicare rates.
- An organization-wide need does not size an individual fee.A demonstrated need at the enterprise level does not prove that every facility fee or ownership premium is properly sized.
- Opacity can conceal inefficiency.Purchasers cannot determine which margin finances a public good, which finances expansion, and which reflects administrative cost or market power.
The 340B program illustrates the ambiguity
The program unquestionably generates revenue that can support services, and a 2023 scoping review found evidence of expanded service capacity in some settings (Knox et al., 2023).
However, Desai and McWilliams (2018) associated hospital 340B eligibility with oncology consolidation and found limited evidence that participation in the program improved care for low-income populations.
This does not establish that 340B margins are socially valueless. It shows that a subsidy tied to drug acquisition or organizational status does not automatically yield transparent, proportional community benefits.
CRITICAL INTERPRETATION
Hospitals may be right that margins finance public goods and still be wrong that ownership-based payment is the best way to finance them.
Both propositions can be true at the same time.
Radiology Is the Clearest Test Case
Standardized, code-defined, and deliverable in both settings. Imaging is where the argument is hardest to avoid.
Why imaging concentrates the dispute
Radiology concentrates the site-neutral dispute because many diagnostic services are standardized, code-defined, technologically comparable, and deliverable in both hospital and freestanding settings.
A noncontrast CT, MRI, radiograph, or ultrasound can often be performed safely outside a hospital. Yet the hospital technical payment may exceed the office or independent-center rate, and beneficiary coinsurance rises with the allowed amount.
The price difference can persist even when the scanner, technologist qualifications, radiologist, protocol, and clinical indication are similar.
The price difference can persist even when the scanner, technologist qualifications, radiologist, protocol, and clinical indication are similar.
The hospital argument is not frivolous
Hospital imaging departments frequently support emergency coverage, inpatient throughput, sedation, complex patients, infection controls, after-hours staffing, subspecialty coverage, and redundant infrastructure.
However, those costs are not evenly present in every off-campus scan. Paying an across-the-board ownership premium treats a routine scheduled outpatient examination as if it carries the full readiness burden of the hospital enterprise.
It also gives hospitals more acquisition currency when purchasing independent imaging centers.
Narrow in dollars, large in signal
The proposed 2027 policy is narrow but strategically significant. It focuses on imaging without contrast in excepted off-campus departments rather than all imaging everywhere. Rural Sole Community Hospitals are exempt.
The reform does not eliminate hospital imaging revenue; it reduces the premium for a set of services CMS regards as substitutable.
If finalized, the immediate dollar effect is modest relative to total hospital spending, but the ownership signal is large: historical grandfathering may no longer be permanent.
The immediate dollar effect is modest. The ownership signal is large. Historical grandfathering may no longer be permanent.
Who is affected, and how
Radiology stakeholders under the proposed policy
| Stakeholder | Near-term effect | Strategic response | Unresolved question |
|---|---|---|---|
| Hospital systems | Lower technical revenue on affected off-campus studies | Model code-level exposure; redesign site portfolio; protect ED and inpatient capacity explicitly | How much affected margin currently supports non-imaging obligations? |
| Independent imaging centers | Reduced Medicare ownership disadvantage | Compete on access, cost, service, and contracting | Will commercial payers follow Medicare? |
| Radiologists | Potential pressure on stipends, coverage, and professional arrangements | Separate professional value from technical cross-subsidy | Who finances 24/7 subspecialty coverage? |
Who is affected, and how
Patients, employers, and payers
| Stakeholder | Near-term effect | Strategic response | Unresolved question |
|---|---|---|---|
| Patients | Lower coinsurance and potentially more site choice | Use transparent scheduling and benefit navigation | Will lower rates improve access or trigger local closures? |
| Employers and payers | Greater leverage to steer routine imaging | Contract by site, quality, and total episode cost | Can network design preserve complex-care capacity? |
Beneficiary coinsurance falls with the allowed amount, which is why patient liability is a policy outcome and not a side effect.
The Ownership Premium and the Consolidation Flywheel
Payment policy intended to support hospitals can reduce the number of independent alternatives.
A repeatable acquisition cycle
Site-based payment creates a repeatable acquisition cycle. Higher hospital reimbursement increases the value of owning the outpatient site. Acquisition then converts independently billed services into hospital-based encounters.
Higher revenue and greater referral control strengthen the system’s balance sheet and its bargaining position with payers. That stronger position finances additional acquisitions.
Over time, payment policy intended to support hospitals can reduce the number of independent alternatives and make the market more dependent on the institutions receiving the premium.
The five stages of the flywheel
- 1DifferentialHospital-owned site receives a higher allowed amountOwnership becomes a revenue strategy
- 2AcquisitionSystem purchases physician practice, imaging center, infusion site, or ASC interestIndependent supply contracts
- 3ConversionServices move to hospital billing or referral channelsPatients and payers face higher prices
- 4LeverageSystem controls more physicians, sites, and downstream volumeNegotiating power rises
- 5ReinvestmentAdditional margin supports capacity, acquisitions, or enterprise overheadFlywheel repeats; local alternatives decline
What the evidence supports
Hospital ownership of physician practices has been associated with higher prices and spending (Baker et al., 2014; Capps et al., 2018; Neprash et al., 2015).
Lin et al. (2021) estimated that hospital prices increased by 3% to 5% following physician integration, without commensurate evidence of higher quality. McCarthy and Sheehan-Connor (2022) found that hospital costs increased by 1% to 3% after integration, driven in part by outpatient visits.
Hu et al. (2024) further documented that Medicare site-based payment incentives were associated with hospital-physician integration. The cumulative literature makes a narrow coordination-only explanation increasingly difficult to sustain.
Integration can produce real benefits
Shared records, standardized protocols, unified call coverage, capital access, and population-health infrastructure can improve coordination.
The problem is that the payment premium is not conditioned on demonstrating those gains.
A system can receive higher facility payment because it owns the site, even if quality, access, and patient complexity are unchanged.
The payment premium is not conditioned on demonstrating the gains that would justify it.
What Happens When the Subsidy Is Removed?
The accounting effect is straightforward. The organizational response is not.
One accounting effect, many possible responses
The immediate accounting effect is straightforward: affected outpatient revenue declines. The organizational response is not.
Hospitals can lower costs, renegotiate contracts, reduce administrative overhead, consolidate locations, change hours, shift services to exempt sites, cut cross-subsidized programs, or seek higher commercial rates.
The distribution of these responses determines whether site neutrality produces efficiency or merely relocates financial pressure.
Figure 3
Illustrative outpatient margin transmission. Values are indexed and do not estimate any hospital-specific or national-budget effect.
Published figure reproduced from the report. Illustrative scenario constructed by the author.
Reading the scenario
The scenario begins with an outpatient margin index of 100. An illustrative rate shock removes 18 points, volume migration removes 7 points, and operational redesign recovers 10 points, leaving 85 points.
The arithmetic is intentionally simple. Its purpose is to force leadership teams to distinguish gross reimbursement exposure from residual operating exposure.
A hospital that models only the rate change may overstate the final loss; a hospital that assumes every lost dollar can be absorbed through efficiency may understate the threat to access.
Likely organizational responses
- Portfolio sortingMove routine services into lower-cost facilities while retaining complex care on the hospital campus.
- Cost unbundlingIdentify the actual incremental cost of emergency readiness, teaching, standby capacity, and complex-patient support.
- Contract redesignSeek commercial arrangements that pay explicitly for access, quality, network adequacy, and after-hours coverage.
- Service-line reprioritizationReduce or close programs whose losses can no longer be supported by outpatient margins.
- Ownership restructuringPursue joint ventures, management agreements, or clinically integrated networks rather than full acquisition.
- Revenue substitutionExpand higher-acuity procedures, specialty pharmacy, ambulatory surgery, or other remaining margin sources.
The danger is cumulative policy exposure
Site-neutral imaging, 340B redistribution, Medicaid changes, rising uninsured volumes, workforce costs, and commercial resistance do not arrive in isolation.
A policy that is manageable on its own can become destabilizing when stacked against simultaneous losses.
This is why system-level simulation should incorporate payer mix, service-line margins, fixed-cost absorption, debt covenants, capital obligations, and local substitution capacity rather than relying on a single national percentage.
A policy that is manageable on its own can become destabilizing when stacked against simultaneous losses.
Rural and Safety-Net Hospitals Require Precision, Not Rhetoric
The evidence supports caution, but not categorical exemption from analysis.
What the evidence actually shows
Rural hospitals are frequently invoked as the decisive objection to site-neutral reform.
Lou et al. (2025) found that small and rural OPPS hospitals would absorb the smallest aggregate shares of reductions under the options studied because their affected volumes were lower. CMS has proposed exempting rural Sole Community Hospitals from its 2027 noncontrast imaging policy.
These safeguards reduce, but do not eliminate, concern for individual hospitals with unusually concentrated outpatient exposure.
A uniform narrative is analytically weak
Jia et al. (2026) add needed nuance. Critical Access Hospitals remained dependent on profitable commercial care, yet higher Medicare and Medicaid shares were modestly associated with higher overall margins, while uncompensated care was associated with lower margins.
Because CAHs receive cost-based Medicare reimbursement, public payer mix does not operate identically across hospital classes. A uniform narrative that Medicare volume is inherently loss-producing is therefore analytically weak.
The more direct threat may be uncompensated care, local commercial weakness, insufficient scale, or service-line concentration.
Exemption should follow access, not labels
A credible rural policy should be hospital-specific and access-specific.
Exemption should depend on substitutability, travel time, local capacity, financial vulnerability, and the service’s role in emergency readiness, not simply on a rural label.
Some rural hospital outpatient services are the only local option and deserve explicit support for access. Others may be delivered efficiently through freestanding or shared regional models without undermining the hospital.
Exemption should depend on substitutability, travel time, local capacity, and financial vulnerability, not simply on a rural label.
A Better Policy Architecture
Separate the clinical service from the social obligation, then finance each one openly.
A false dichotomy
The choice between unrestricted hospital differentials and universal flat payment is a false dichotomy.
The better model separates the clinical service from the social obligation.
The base payment should follow the resources needed for a clinically comparable service. Supplemental payment should follow measurable obligations that society wants maintained.
Base payment follows the resources a comparable service requires. Supplemental payment follows the obligations society wants maintained.
Six design principles
| Design principle | Operational rule | Reason |
|---|---|---|
| Clinical comparability | Neutralize rates only when services, patient acuity, resources, and safety requirements are genuinely comparable | Avoids treating unlike care as identical |
| Explicit public-good supplements | Pay separately for trauma, emergency readiness, teaching, rural access, uncompensated care, and standby capacity | Replaces opaque ownership premium with accountable support |
| Hospital-specific transition | Phase changes using code-level exposure, financial vulnerability, and local substitution capacity | Reduces avoidable access shocks |
Six design principles (continued)
| Design principle | Operational rule | Reason |
|---|---|---|
| Commercial market guardrails | Pair neutrality with antitrust enforcement, price transparency, and network adequacy | Prevents savings from being captured only by dominant payers or providers |
| Quality and access monitoring | Track wait times, closures, travel, service migration, outcomes, and patient liability | Detects unintended harm early |
| Budget-neutrality transparency | Show which hospitals and service lines gain from redistribution as well as which lose | Makes policy incidence visible |
Recommended federal sequence
- Start narrow and substitutableBegin with high-volume, low-complexity services that demonstrate strong clinical substitutability across settings.
- Publish the impact files firstRelease code-level and hospital-level impact files before finalization, including beneficiary cost-sharing effects.
- Create a readiness payment layerDefine a public-good payment layer that is independent of where routine outpatient services are billed.
- Use transition corridorsGive unusually exposed hospitals time to redesign rather than abruptly withdraw services.
- Require public reportingReport how supplemental payments support named services and community access.
- Evaluate it as market structureAssess the reform as a market-structure intervention, not only as a Medicare budget intervention.
This is a portfolio problem, not a revenue-cycle problem
Health-system boards should not treat site-neutrality as a revenue-cycle problem. It is an enterprise portfolio problem.
Leadership should map every affected code to location, payer, cost, margin, patient complexity, downstream contribution, and cross-subsidized obligation.
The analysis should show which services remain economically viable at a neutral rate, which require redesign, and which public goods would lose support. The final product should be a board-level access and capital plan, not merely a reimbursement variance report.
The final product should be a board-level access and capital plan, not merely a reimbursement variance report.
Five questions a board should ask
- Where is the exposure?Code, payer, location, volume, allowed amount, variable cost, fixed-cost absorptionPrioritize sites and service lines
- What does the margin support?Traceable transfers to ED, trauma, teaching, behavioral health, call coverage, and charity careDefend or redesign the subsidy
- Can care move safely?Acuity, travel time, network capacity, clinical exclusions, scheduling accessChoose site-of-care strategy
- What can be redesigned?Labor model, hours, equipment utilization, lease, overhead, referral flowBuild mitigation plan
- What must be protected?Community need, unique capability, regulatory obligation, quality, equitySeek explicit supplemental funding
Strategic Conclusions
Reform redistributes both market power and dollars.
Not a minor billing detail
Site-neutral payment is becoming a battle over who owns outpatient care because the payment differential is not a minor billing detail.
It is embedded in acquisition valuations, referral control, service-line profitability, beneficiary cost sharing, and the financing of hospital obligations.
Reform therefore redistributes both market power and dollars.
Where each case is strongest
The case for reform is strongest where the service is routine, clinically substitutable, and paid more solely because a hospital owns the location.
The case for caution is strongest where a facility maintains unique access, treats patients who are meaningfully more complex, or incurs demonstrable standby and public-service costs.
Policy fails when it assumes all hospital care is identical to office care. It also fails when it assumes hospital ownership alone proves added value.
Policy fails when it assumes all hospital care is identical to office care. It also fails when it assumes hospital ownership alone proves added value.
The central error is concealment
The central policy error of the current system is not the existence of cross-subsidy. It is that cross-subsidy is largely hidden.
Society has allowed higher prices for commercially attractive and outpatient services to function as an unofficial financing mechanism for emergency readiness, teaching, rural access, uncompensated care, and organizational expansion.
That mechanism is politically convenient because it avoids naming the tax, defining the benefit, or auditing the transfer. It is also economically distortive because it rewards ownership, encourages consolidation, and raises patient liability.
FINAL JUDGMENT
Site-neutral payment should proceed, but it should be redesigned as a transition from implicit subsidy to explicit public financing.
The same service should receive the same base payment when clinical and resource requirements are comparable. Hospitals should then be paid directly, transparently, and adequately for the public goods only hospitals are expected to provide.
Methods and research design
This report is an independent synthesis of evidence and a scenario analysis.
Sources were identified through targeted searches of CMS, the Federal Register, MedPAC, PubMed, PubMed Central, Health Affairs, Health Affairs Scholar, and peer-reviewed health economics literature.
Priority was given to 2024 to 2026 studies, systematic reviews, national claims analyses, and primary federal documents. Older foundational studies were retained when they defined cost shifting, cross-subsidy, or vertical-integration mechanisms that remain central to the current debate.
Limitations
- The report does not analyze patient-level claims, hospital cost-report microdata, or proprietary commercial contracts.
- National estimates should not be applied mechanically to an individual hospital. Exposure depends on service mix, payer mix, geography, cost structure, grandfathered status, and local alternatives.
- Observational integration studies can be affected by selection, market structure, and unmeasured differences between acquiring and nonacquiring systems.
- Cross-subsidy is difficult to verify because hospital cost accounting and internal transfers are not consistently public or standardized.
- The CY 2027 OPPS and ASC provisions discussed here were proposed, not final, as of August 29, 2026.
- The illustrative margin index is a conceptual stress test, not a budget score, forecast, or estimate of any named organization.
INTEGRITY STATEMENT
No conclusion was conditioned on preserving a hospital revenue stream or maximizing payer savings.
The analysis applies the same test to both sides: claims require evidence, and social obligations should be financed in proportion to demonstrable need rather than institutional status alone.
References
- Baker, L. C., Bundorf, M. K., & Kessler, D. P. (2014). Vertical integration: Hospital ownership of physician practices is associated with higher prices and spending. Health Affairs, 33(5), 756-763.
- Capps, C., Dranove, D., & Ody, C. (2018). The effect of hospital acquisitions of physician practices on prices and spending. Journal of Health Economics, 59, 139-152.
- Centers for Medicare & Medicaid Services. (2026a, July 2). Calendar year 2027 Hospital Outpatient Prospective Payment System and Ambulatory Surgical Center proposed rule fact sheet.
- Centers for Medicare & Medicaid Services. (2026b). Medicare program: Hospital Outpatient Prospective Payment and Ambulatory Surgical Center payment systems; proposed rule, CMS-1850-P. Federal Register.
- David, G., Helmchen, L. A., & Henderson, R. A. (2014). Do hospitals cross-subsidize? Journal of Health Economics, 37, 198-218.
- Desai, S., & McWilliams, J. M. (2018). Consequences of the 340B Drug Pricing Program. New England Journal of Medicine, 378(6), 539-548.
- Frakt, A. B. (2011). How much do hospitals cost shift? A review of the evidence. Milbank Quarterly, 89(1), 90-130.
- Harris, A., Philbin, S., Post, B., et al. (2025). Cost, quality, and utilization after hospital-physician and hospital-post acute care vertical integration: A systematic review. Medical Care Research and Review, 82(1), 3-42.
- Hu, X., et al. (2024). Association between Medicare site-based payment policy and hospital-physician integration. JAMA Health Forum.
- Jia, Y., et al. (2026). Payer mix shifts and profitability at Critical Access Hospitals. Health Affairs Scholar, 4(4), qxag083.
- Knox, R. P., et al. (2023). Outcomes of the 340B Drug Pricing Program: A scoping review. Exploratory Research in Clinical and Social Pharmacy, 12, 100326.
References (continued)
- Lin, H., McCarthy, I. M., & Richards, M. (2021). Hospital pricing following integration with physician practices. Journal of Health Economics, 77, 102444.
- Lou, K. K., Linehan, K. E., da Fonte, L. N., Lai, P., & Buntin, M. B. (2025). Medicare site-neutral payment policies: Effects of proposals on hospitals and beneficiary groups. Health Affairs, 44(6), 668-676.
- McCarthy, S., & Sheehan-Connor, D. (2022). The effect of hospital-physician integration on hospital costs. Health Economics, 31(11), 2333-2368.
- Medicare Payment Advisory Commission. (2023). Aligning fee-for-service payment rates across ambulatory settings. In Report to the Congress: Medicare and the health care delivery system.
- Murray, R. C., et al. (2025). Site-neutral payment for routine services could save commercial purchasers and patients billions. Health Affairs Scholar.
- Neprash, H. T., Chernew, M. E., Hicks, A. L., Gibson, T., & McWilliams, J. M. (2015). Association of financial integration between physicians and hospitals with commercial health care prices. JAMA Internal Medicine, 175(12), 1932-1939.
- Parente, S. T., et al. (2024). Impact of site-neutral payments for commercial and Medicare purchasers. Health Affairs Scholar.
- Post, B., Thai, N., Noor-E-Alam, M., & Young, G. J. (2025). Site-neutral payment reform: Little impact on outpatient Medicare spending or hospital-physician integration. Health Affairs, 44(6), 659-667.
- Sen, A. P., et al. (2022). Site-based payment differentials for ambulatory services. Health Affairs, 41(2).
- Song, Z., Rose, S., Safran, D. G., Landon, B. E., Day, M. P., & Chernew, M. E. (2014). Changes in health care spending and quality four years into global payment. New England Journal of Medicine, 371(18), 1704-1714.
About this edition
This flipbook edition reproduces the research report Who Owns Outpatient Healthcare? by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
Figures 1, 2 and 3 are the published figures, reproduced as printed in the source report. All estimates remain attributable to the sources cited beneath them.
The spoken narration was generated from the report’s own text and is provided as a reading aid. Where narration and page differ, the page governs.
This report is an independent evidence synthesis. It is not legal, financial, or reimbursement advice. The CY 2027 OPPS and ASC provisions it analyzes were proposed, not final, as of August 29, 2026.
Kelly Emrick
DHSc PhD MBA BSRT(ARRT)R
Who Owns Outpatient Healthcare?
Pay the same base rate for the same clinically comparable service, then pay separately and transparently for measurable hospital obligations.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)RIndependent evidence synthesis and scenario analysis | August 2026Investigative report
Where Radiology Lawsuits Begin
The top 10 primary liability factors, claim outcomes, and a defensible risk-control system
Investigative report
Where Radiology Lawsuits Begin
The top 10 primary liability factors, claim outcomes, and a defensible risk-control system
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
U.S.-focused evidence review | Evidence current through August 28, 2026
A ranked evidence synthesis of national paid claims, large closed-claim cohorts, legal-database studies, insurer analyses, and contemporary safety guidance. It is a scholarly risk-management analysis, not legal advice.
How to use
Reading, listening, and moving around
Contents
What is inside
The premise
The principal liability problem is not the isolated miss. It is the uncontained miss.
Radiology malpractice risk is overwhelmingly diagnostic, but the lawsuits with the greatest clinical and financial force usually require more than an imperfect image interpretation. They are strengthened when an abnormality is missed or minimized, urgency is not communicated, follow-up is not assigned, or a complication is not recognized and rescued.
Kundu et al. (2026), National Practitioner Data Bank paid claims.
Bottom line
Five links, not one signature
Bottom line
The most defensible risk strategy is a five-link reliability system: acquire the right study, interpret it with appropriate context, communicate actionable findings, close every follow-up loop, and recognize and rescue complications. A report signed on time is not the endpoint. Verified clinical action is.
Figure 1.Where the four handoff risks sit.
Author framework, redrawn for this edition in the report palette.
Table 1
The ranking, factors 1 to 5
| # | Primary factor | Signal | Principal risk owner |
|---|---|---|---|
| 1 | Diagnostic misinterpretation and failure to diagnose | Very high | Radiologist and department |
| 2 | Delay in diagnosis and escalation failure | Very high | Shared clinical system |
| 3 | Failure to communicate critical, unexpected, or revised findings | High | Radiologist and department |
| 4 | Failure to close recommended follow-up and incidental findings | High | Department and health system |
| 5 | Teleradiology, after-hours context, and handoff failure | High | Group and contracting entity |
Table 1 (part 1). Author synthesis of the sources reviewed in the report.
Table 1
The ranking, factors 6 to 10
| # | Primary factor | Signal | Principal risk owner |
|---|---|---|---|
| 6 | Interventional radiology technical injury | Moderate to high | Operator and facility |
| 7 | Failure to recognize, monitor, or rescue complications | Moderate to high | Team and facility |
| 8 | Inadequate informed consent and deficient documentation | Moderate | Physician and facility |
| 9 | Equipment, protocol, image-quality, and radiation-dose failure | Moderate | Department and facility |
| 10 | Wrong patient, site, or procedure and MR environmental safety failure | Lower frequency, extreme severity | Department and facility |
Table 1 (part 2). Ranking is an evidence synthesis, not an actuarial model.
Claim outcomes
Most claims do not end in a verdict
A 2025 U.S. radiology malpractice guide reports that about 63% of claims are dropped or dismissed, 28% settle, and the remainder proceed to trial, where defense verdicts predominate. A nonpayment claim is not a costless claim.
Figure 2.Average indemnity and allocated legal expense by case disposition.
Wojdyla and Chen (2025). Descriptive averages, not predictions for a particular jurisdiction or case.
Table 2
What each disposition means operationally
| Case outcome | Indemnity | Legal expense | Operational meaning |
|---|---|---|---|
| Dropped, denied, or dismissed | $0 | $28,000 | No NPDB payment report; substantial time and stress remain |
| Settlement | $428,021 | $67,524 | Payment generally reportable to the NPDB |
| Defense verdict | $0 | $125,661 | No indemnity, but high defense cost and prolonged exposure |
| Plaintiff verdict | $980,000 | $258,421 | Payment reportable; excess exposure possible |
Table 2. Source: Wojdyla and Chen (2025).
The short list
Six controls that carry the most weight
- Closed-loop communication with named recipients, acknowledgment, escalation timers, and durable audit trails.
- A centralized follow-up registry for actionable and incidental findings, with patient notification and documented clinical closure.
- Risk-tiered worklists for cancer, acute vascular disease, trauma, and preliminary-final discrepancies.
- Teleradiology contracted as a clinical service: full priors, complete clinical data, discrepancy deadlines, aligned credentialing.
- Standardized time-outs, consent, technical checklists, post-procedure surveillance, and rescue pathways.
- Equipment, protocols, dose, MR access, identity verification, staffing, and competency governed as board-visible risks.
Section 1
Scope, definitions, and method
What are the ten primary factors that most often make U.S. radiology care the subject of a physician-based or department-based malpractice claim, what legal and clinical outcomes follow, and which controls can materially reduce the risk?
Scope
What the report covers
The report addresses diagnostic radiology, breast imaging, interventional radiology, contrast administration, teleradiology, imaging operations, and MR safety.
It includes direct physician negligence and organizational theories such as vicarious liability, apparent agency, negligent credentialing or retention, inadequate policy and staffing, equipment failure, and failure to operate a reliable results-management system.
The focus is the United States because malpractice doctrine, claim reporting, insurance structures, damages, and procedural rules vary sharply across jurisdictions.
Important legal limitation
This is a scholarly risk-management analysis, not legal advice. The applicable standard of care, statutes of limitations, damages rules, reporting duties, privilege protections, and theories of institutional liability vary by state and by case. A practice should validate policies with local counsel, its insurer, and medical staff leadership.
Method
Four layers of evidence, four limitations
| Evidence layer | Strength | Primary limitation |
|---|---|---|
| National paid claims (NPDB) | Best for allegation mix, severity, payment association, and long-term trends | Only claims with a payment; public-use coding limits detail |
| Closed claims (insurer cohorts) | Adds nonpayment outcomes, contributing factors, and defense perspective | Proprietary participation; not a national sample |
| Legal database studies | Captures allegation narrative and litigated detail | Overrepresents appealed or published matters; cannot estimate incidence |
| Safety surveillance and guidance | Identifies foreseeable injury pathways and current standards | Not evidence of negligence, causation, or lawsuit frequency |
Evidence searched and synthesized through August 28, 2026.
Method
How the ten were ranked
Disease entities such as breast cancer and fractures are treated as high-risk manifestations of diagnostic failure, not as independent root causes. Rare sentinel events appear lower in the frequency ranking even when their preventability and reputational consequences are extreme.
Section 2
How radiology error becomes legal liability
An error is not automatically malpractice. A plaintiff generally must establish duty, breach of the applicable standard of care, injury, and proximate causation.
The legal test
Breach, causation, and the weight of severity
In radiology, breach often becomes a battle over whether the interpretation, procedure, communication, or follow-up response was reasonable under the circumstances.
Causation can be decisive. A missed lesion may be real, yet damages require evidence that earlier diagnosis or different action would probably have changed the clinical outcome.
adjusted odds ratio for payment when the patient outcome was severe
Kundu et al. (2026), NPDB radiology analysis.
Figure 3
The radiology liability chain
Figure 3. The radiology liability chain and its four high-risk handoffs. Source: Author framework, as published in the report.
Why it matters
The same miss, two very different cases
Contained
- Error detected on review or comparison
- Addendum issued and actively communicated
- Receipt acknowledged by a named clinician
- Follow-up completed or clinically closed
- Often no compensable injury
Uncontained
- Error persists through sign-off
- Addendum sits unread in an inbox
- No acknowledgment, no escalation
- Recommendation never completed
- A delay claim with a preventability narrative
Defensibility therefore depends on both diagnostic performance and system reliability.
Table 4
Liability architecture: who can be named
| Potential defendant | Typical duty or control | Common theory |
|---|---|---|
| Radiologist | Interpretation, comparison with priors, reporting, communication, procedure, supervision, consent | Direct professional negligence |
| Radiology group | Staffing, peer review, contracting, schedule design, credentialing, teleradiology integration | Vicarious liability and direct organizational negligence |
| Hospital or imaging center | Policies, technologists, equipment, identity systems, emergency response, follow-up infrastructure | Employee liability, apparent agency, corporate negligence |
| Ordering or treating clinician | Clinical context, response to report, referral, follow-up, discharge communication | Shared or comparative fault; direct negligence |
| Vendor or manufacturer | Software, device, injector, contrast agent, worklist, decision support | Product, contract, or contribution theories; fact-specific |
Table 4. General U.S. framework; state law varies.
Section 3
National claim landscape
399,568 malpractice payment records involving 229,276 physicians, September 1990 through December 2024. Only 1,978 were radiology claims, but their profile was distinctive.
The profile
Small in number, distinctive in shape
mean payment, 2010 to 2025, against $184,000 from 1990 to 2010
Kundu et al. (2026). These figures describe paid claims, not every filed case.
Figure 4
Diagnostic failure dominates paid claims
Figure 4.Distribution of paid radiology claims by allegation category.
Kundu et al. (2026). Rebuilt for this edition. Other diagnostic error is derived by subtraction and is not a separately coded NPDB allegation.
Failure to diagnose is usually an interpretation allegation. Delay in diagnosis is a systems narrative that can involve the radiologist, the ordering clinician, the outpatient practice, and the facility.
Section 4
The top 10 primary factors
Each factor is presented the same way: why it ranks where it does, what the outcomes look like, the controls that work, and the measure that proves the control is working.
Diagnostic misinterpretation and failure to diagnose
Why it ranks first
Failure to diagnose is the single largest coded allegation in the newest national paid-claim evidence. In the earlier, larger radiologist cohort, diagnostic error occurred at 14.83 claims per 1,000 person-years, far exceeding procedural complications, communication failures, and failures to recommend additional testing.
The legally important error may be perceptual, cognitive, technical, or contextual: the abnormality was not seen, was seen but discounted, was assigned the wrong diagnosis, or was interpreted without the priors or history that would have changed the conclusion.
Whang et al. (2013).
Factor 1 | Figure 5
Where diagnostic claims concentrate
Figure 5.High-risk diagnosis-related claim subjects.
Whang et al. (2013), 4,793 claims involving 8,401 radiologists. Rebuilt for this edition.
In nonspinal musculoskeletal claims, 91.8% were attributed to failure to diagnose. In a breast cancer legal-database series, delay was alleged in 82% of cases and radiologists were named in 43%.
Factor 1
Controls that work
- Prioritize comparison imaging and surface missing priors before sign-off, especially for cancer, trauma, and postoperative studies.
- Use structured search patterns and indication-specific report templates for high-harm examinations, while preserving room for unexpected findings.
- Apply targeted second review to selected high-risk worklists, major discrepancies, new staff onboarding, and domains flagged by peer learning.
- Manage cognitive load: subspecialty routing, protected reading conditions, interruption controls, reasonable shift length, and turnaround targets that do not reward unsafe speed.
- Create nonpunitive peer learning linked to pathology, surgery, return visits, and later imaging so error classes drive education and protocol redesign.
Measure
Major discrepancy rate by modality and diagnosis; rate of prior-image availability; amended reports with clinically significant change; pathology or surgery discordance closed within 30 days; and repeat imaging prompted by initial interpretive error.
Outcomes
Diagnostic claims disproportionately involve severe injury or death. Severe outcomes appeared in 64.2% of paid radiology claims. In one insurer review, 44% of diagnostic cases resulted in payment, with a mean indemnity of $300,000 and a median of $100,000.
Delay in diagnosis and escalation failure
Why it ranks second
Delay is a distinct legal theory because it connects an earlier opportunity to a later, worse outcome. It includes equivocal findings that were not resolved, undercalled abnormalities, failure to recommend the next step, recommendations with vague timing, and abnormal studies that reached no accountable clinician.
It is especially potent in oncology, where serial imaging can visually demonstrate progression to a jury.
average alleged delay in a 253-case breast cancer litigation series
Lee et al. (2020). Delays of 12 months or longer were more likely to be paid, with an odds ratio of 2.13.
Factor 2
Controls that work
- Write recommendations that specify the finding, modality, time interval, and responsible clinical action. Avoid unsupported hedging and indefinite phrases.
- Use urgency categories tied to maximum communication and acknowledgment times.
- Escalate unresolved discordance between imaging and the clinical picture, including a direct clinician conversation when a negative study does not safely end the diagnostic pathway.
- Track interval recommendations to completion or documented clinical closure, not merely to message delivery.
- Notify patients in plain language when local law and policy permit, while preserving clinician coordination and documenting the notification.
Measure
Median days from abnormal image to definitive action; percentage of recommendations with explicit modality and interval; overdue high-risk recommendations; and time from amended report to treatment-plan reassessment.
Outcomes
The mean payment across payment cases was approximately $979,000, though dispersion was very wide. Lack of referral to a surgeon was also associated with payment, whereas documented patient failure to follow up was associated with a lower likelihood of payment.
Failure to communicate critical, unexpected, or revised findings
Why it ranks third
A correct report does not protect the patient if its urgency is not understood. Communication allegations include failure to make direct contact, sending a message to the wrong clinician, relying on an unmonitored inbox, omitting a change from preliminary to final interpretation, and failing to document receipt.
Communication failure reduces a medically complex defense to a simpler operational story: the defendant knew or should have known, but the information did not reach someone who could act.
communication-failure claims were less often dropped, denied, or dismissed
Humphrey et al. (2022), malpractice claims across specialties.
Factor 3
Controls that work
- Define three or four urgency tiers with named recipient roles and acknowledgment deadlines.
- Use closed-loop technology that records sender, recipient, timestamp, acknowledgment, escalation, and final disposition.
- Prohibit voicemail-only closure for life-threatening findings and define the chain for an unreachable ordering clinician.
- Treat clinically important addenda and preliminary-final discrepancies as new results requiring active communication.
- Audit semantic clarity: the impression should state the abnormality, urgency, and recommended action without contradiction.
Measure
Critical-result acknowledgment within policy; unacknowledged alerts at each escalation interval; revised-finding communication compliance; unreachable-clinician events; and harm reviews linked to result communication.
Outcomes
In teleradiology claims, provider communication problems occurred in 25.9%, compared with 12.6% in other radiology claims. Historical radiology liability work also found materially higher indemnity when communication, rather than interpretation, was the primary failure.
Failure to close recommended follow-up and incidental findings
Why it ranks fourth
A report can be delivered and acknowledged, yet the recommended CT, MRI, biopsy, or specialty referral never occurs. Ownership becomes diffuse across the radiologist, ordering clinician, primary care clinician, patient, and facility.
Outpatient prevalence intensifies the risk because there may be no inpatient team to reconcile the plan.
Figure 6.A multistage recommendation-tracking program, before and after.
Wandtke and Gallagher (2017). Completion figures are as published. The 53.5% unknown-status baseline is derived from the published 13.9% endpoint and the reported 74% relative reduction; it is not a published value.
Factor 4
Controls that work
- Create a single enterprise registry for actionable recommendations, independent of ordering location and EHR inbox.
- Assign a named service-line owner and a patient-level accountable clinician before the timer starts.
- Use standardized recommendation language that can be parsed reliably, with evidence-based intervals and indication.
- Escalate overdue items to a coordinator, the patient, the responsible service chief, and risk management according to harm tier.
- Permit documented clinical closure when follow-up is not appropriate because of goals of care, outside completion, specialist management, or informed refusal.
Measure
Completed or clinically closed by due date; unknown status; patient awareness; finding inclusion in discharge summaries; time to first escalation; and cancer diagnoses originating from tracked findings.
Outcomes
A 2024 inpatient study still found incidental findings in only 51% of discharge summaries after an electronic notification intervention. Direct nurse-coordinator contact found patients and primary care clinicians aware of 79% when follow-up was due. Technology without accountable human closure remains incomplete.
Teleradiology, after-hours context, and handoff failure
Why it ranks fifth
Teleradiology is not inherently unsafe, but its claims reveal predictable vulnerabilities: incomplete clinical records, absent priors, overnight workload, weekend coverage, separation from clinicians, preliminary-final discrepancies, and unclear responsibility for revised findings.
The largest recent closed-claim comparison included 135 teleradiology and 3,474 other radiology cases. Teleradiology claims more often involved weekends or holidays, night shifts, failure to read the medical record, and failure or delay in reporting revised findings.
Schaffer et al. (2024).
Factor 5 | Figure 7
More likely to pay, and larger when paid
Figure 7.Teleradiology claim payment and indemnity compared with other radiology claims.
Schaffer et al. (2024), 135 teleradiology and 3,474 other closed claims, 2010 to March 2022. Redrawn for this edition because the published figure printed its two panel titles over the chart title; the values are unchanged.
These findings should inform contract and workflow design, not lead to a blanket conclusion that remote interpretation causes harm.
Factor 5
Controls that work
- Give remote radiologists the same clinical history, priors, protocols, and communication tools available on site.
- Define whether reports are preliminary or final, who performs overreads, maximum overread latency, and how clinically important discrepancies are acknowledged.
- Align credentials, privileges, quality review, peer learning, and continuing education across on-site and remote groups.
- Monitor workload, queue aging, nocturnal shift length, interruptions, and subspecialty mismatch. Do not use turnaround time as the only performance signal.
- Require tested downtime procedures and auditable escalation to the emergency department or responsible clinician.
Measure
Overread discrepancy rate and notification time; missing-prior and missing-history rate; critical-result acknowledgment by shift; studies per work hour by complexity; and teleradiology claim or event trend.
Outcomes
Contract for teleradiology as a clinical service, not a commodity. The claim profile is driven by information access, coverage design, and discrepancy handling, all of which sit inside the contract.
Interventional radiology technical injury
Why it ranks sixth
Interventional radiology adds direct procedural duty to the diagnostic role. Claims include vascular or organ injury, hemorrhage, perforation, pneumothorax, infection, embolization injury, wrong-target treatment, retained devices, and complications of biopsy, drainage, ablation, and access.
In one insurer cohort, improper performance was alleged in 59% of interventional cases, but expert reviewers judged the procedure technically correct in many cases despite the adverse outcome. A complication can be nonnegligent, but the record must show sound selection, consent, technique, monitoring, and response.
mean indemnity in paid interventional claims, against a 30% payment rate
The Doctors Company (2020). Median indemnity was $308,800.
Factor 6
Controls that work
- Use indication and contraindication checks, case-specific planning, pre-procedure laboratory and medication reconciliation, and documented alternative analysis.
- Perform a standardized time-out that includes patient, procedure, site, device, imaging, antibiotics, blood availability, and rescue plan.
- Define competency and proctoring by procedure, not only by broad privilege category. Monitor low-volume high-risk work.
- Use real-time device and specimen counts, implant or device traceability, and postprocedure imaging when indicated.
- Standardize postprocedure observation, discharge instructions, contact access, and thresholds for immediate reassessment.
Measure
Major complication by procedure and risk class; unplanned admission or escalation; return to emergency care within 7 days; retained item or device event; procedure-specific volume and competency status; and time to rescue.
Outcomes
Low frequency and high severity make interventional radiology a classic tail risk. Claims can also name the hospital, nursing staff, technologists, anesthesia professionals, and device manufacturers.
Failure to recognize, monitor, or rescue complications
Why it ranks seventh
Many defensible initial complications become difficult claims when deterioration is missed. This factor spans contrast reactions, extravasation, hemorrhage, sedation-related respiratory depression, sepsis, pneumothorax, compartment syndrome, and allergic-like reactions.
It is separate from technical injury because the alleged negligence occurs after the initiating event: monitoring was inadequate, symptoms were dismissed, emergency drugs or equipment were unavailable, or escalation was delayed.
Khan et al. (2023); ACR Manual on Contrast Media.
Factor 7 | Figure 8
Rescue failure, not the decision to give contrast
Figure 8.Major allegation categories in contrast-media litigation.
Khan et al. (2023), 151 published U.S. legal case summaries through October 2022. Rebuilt for this edition. Selection bias limits population inference.
All eleven alleged kidney-injury cases that went to trial resulted in defense verdicts. The pattern suggests that visible rescue failure, not the mere administration of contrast, drives the strongest allegations.
Factor 7
Controls that work
- Use current contrast screening and agent-selection protocols, but avoid overstating renal risk or creating unsupported barriers to clinically necessary imaging.
- Keep adult and pediatric reaction cards, appropriately stocked emergency kits, oxygen, suction, and monitoring immediately available. Audit expiration and seal integrity.
- Require recurring simulation for anaphylaxis, airway compromise, extravasation with neurovascular injury, oversedation, hemorrhage, and MR-zone emergency response.
- Document the reaction, treatment, response, disposition, patient instructions, and future contrast recommendations in a visible longitudinal record.
- Set escalation criteria that favor early evaluation when pain, swelling, neurologic findings, hypotension, hypoxia, or delayed deterioration occurs.
Measure
Reaction-to-epinephrine time; emergency-kit readiness; percentage of staff completing simulation; unplanned transfer after contrast or sedation; extravasation neurovascular documentation; and complication-to-escalation interval.
Inadequate informed consent and deficient documentation
Why it ranks eighth
Consent and documentation rarely substitute for safe care, but they strongly influence what can be proven. Claims allege failure to explain material risks, alternatives, expected benefits, who will perform the procedure, or the possibility of conversion and rescue.
Documentation failures also include contradictory reports, unsigned addenda, copy-forward errors, missing procedure details, and records altered after an event.
The Doctors Company (2020).
Factor 8
Controls that work
- Use a procedure-specific consent conversation covering material risks, alternatives, likely benefit, no-treatment consequences, and patient questions.
- Document who participated, interpreter use, capacity, anticoagulant plan, allergies, and any material change in risk after consent.
- Use structured procedure notes that capture indication, technique, devices, contrast and dose, specimens, complications, estimated blood loss, and disposition.
- Correct reports through transparent, time-stamped addenda with direct communication when the change is clinically important.
- After an adverse event, preserve the record, notify the insurer or risk office, and use the established disclosure process. Do not alter prior documentation.
Measure
Procedure-specific consent completeness; interpreter documentation; unsigned reports and notes; clinically significant addenda with communication; template copy-forward defects; and adverse-event disclosure timeliness.
Outcomes
Consent claims can proceed even when technical performance is reasonable, especially if the undisclosed risk is the exact injury that occurs. Post-event alteration or defensive rewriting may create credibility and discovery problems beyond the original clinical issue.
Equipment, protocol, image-quality, and radiation-dose failure
Why it ranks ninth
A radiologist cannot interpret anatomy that was not adequately imaged, and a department can be directly liable for unsafe equipment, poorly governed protocols, inadequate training, or failure to inspect.
One insurer cohort identified equipment inspection and maintenance allegations in 13% of diagnostic closed claims. A cluster involved updated CT scanners and unexpectedly high radiation settings, with allegations including hair loss and fear of future injury. Operational contributors included training, policy, protocol, and inspection failures.
The Doctors Company (2020).
Factor 9
Controls that work
- Use formal change control for software, scanner, injector, reconstruction, protocol, and dose updates, including acceptance testing before clinical release.
- Set modality-specific image-quality and dose alerts with named review responsibility and documented resolution.
- Require radiologists to identify nondiagnostic limitations and obtain repeat or alternative imaging when the clinical question cannot be answered safely.
- Maintain preventive maintenance, calibration, service records, and medical physics review. Reconcile device alerts and recalls.
- Train and validate technologist competency after upgrades, new protocols, and cross-modality assignments.
Measure
Repeat examination rate by cause; diagnostic-quality rejection rate; dose-alert investigation time; post-upgrade competency completion; overdue preventive maintenance; and report limitations that do not trigger a corrective plan.
Outcomes
Equipment claims can expand beyond a single physician to the facility, physicist, technologist, manufacturer, service contractor, and supervising radiologist. Multiple exposed patients create notification, regulatory, reputational, and class-action risk.
Wrong patient, site, or procedure and MR environmental safety failure
Why it ranks tenth
These events are less common in malpractice datasets but are usually highly preventable, emotionally salient, and difficult to defend. They include wrong-patient imaging or contrast, wrong-side or wrong-target intervention, mislabeled specimens, retained objects, patient falls, projectile events, RF burns, implant-related injury, and failures of MR access control.
Delfino et al. (2019), ten-year review of FDA MRI adverse-event reports. Mechanical events accounted for 11% and acoustic events 6%.
Factor 10
Controls that work
- Use two identifiers at acquisition, contrast administration, procedure, specimen handling, and report finalization. Barcode where feasible.
- Conduct a final active pause immediately before invasive procedures, confirming patient, site, procedure, imaging, and laterality with the full team.
- Adopt MR zones, controlled access, Level 1 and Level 2 training, ferromagnetic detection, implant verification, a full-stop final check, and emergency drills.
- Prevent RF burns through patient positioning, insulation, removal of unnecessary conductors, cable management, communication, and immediate response to warmth or pain.
- Use fall-risk screening, safe transfers, adequate staffing, and equipment suited to patient size, mobility, and sedation state.
Measure
Near misses involving identity or laterality; time-out reliability; unscreened Zone III entry; MR burn or projectile events; implant-condition deviations; falls; and corrective-action closure after every serious safety event.
Outcomes
The ACR 2026 MR Safety update emphasizes controlled access to Zone III, Level 2 personnel oversight, equipment inspection, removal of unnecessary conductive materials, visual integrity checks, adherence to device conditions, and improved hearing protection.
Section 5
What the evidence says about outcomes
Claim frequency is not the same as payment probability, and payment probability is not the same as expected indemnity.
Figure 9
What is associated with payment
Figure 9.Adjusted odds ratios associated with claim payment.
Assembled from Kundu et al. (2026), Lee et al. (2020), and Schaffer et al. (2024). Different datasets and periods; these values should not be pooled.
Association is not causation. Severe injury, venue, policy limits, and evidentiary quality all interact.
Table 5
Factors associated with disposition or cost
| Outcome driver | Observed association | Evidence type |
|---|---|---|
| Severe patient outcome | Higher payment likelihood; adjusted OR 8.64 | National paid claims |
| Recent claim period | Higher payment; adjusted OR 2.23 after 2010 | National paid claims |
| Teleradiology | 58.5% paid; adjusted OR 1.74; median indemnity $339,230 | Closed claims |
| Breast cancer delay of 12+ months | Associated with payment; OR 2.13 | Legal database |
| Communication failure | Less often dropped, denied, or dismissed: 54% versus 67% | All-specialty closed claims |
| IR claim | 30% paid, but mean indemnity $587,800 in one insurer cohort | Insurer closed claims |
Table 5. Descriptive synthesis; data sources and periods differ and should not be pooled.
The hidden outcome
Organizational burden
Indemnity captures only part of the cost. Claims consume physician and leadership time, generate legal expense, may trigger NPDB reporting, affect credentialing and insurance, and can produce moral injury, burnout, and defensive practice.
Section 6
A defensible risk-control system
Education alone is weak because it depends on memory under variable workload. Strong controls make the safe action easier, detect failure quickly, assign ownership, and preserve evidence of completion.
Table 6
Control architecture, layers 1 to 4
| Layer | Required control | Accountable owner | Core evidence |
|---|---|---|---|
| Worklist | Risk-tier by harm and time sensitivity; surface priors and clinical context | Reading-workflow owner | Queue aging by harm tier |
| Interpretation | Structured search and report patterns; targeted second review | Section chief | Major discrepancy rate |
| Communication | Closed-loop acknowledgment and escalation | Radiology operations | Acknowledged within policy |
| Follow-up | Enterprise registry to completion or clinical closure | Named program director | Overdue and unknown status |
Table 6 (part 1). Author synthesis.
Table 6
Control architecture, layers 5 to 8
| Layer | Required control | Accountable owner | Core evidence |
|---|---|---|---|
| Procedure | Time-out, competency, surveillance, and rescue standards | IR medical director | Complication and time to rescue |
| Technology | Change control, validation, cybersecurity, downtime, and audit logs | Imaging informatics | Failed interfaces and downtime |
| Safety | MR zones, identity, contrast, dose, falls, and drills | Modality safety leads | Near miss and event closure |
| Governance | Board-visible dashboard, event learning, contract oversight | Chief medical officer | Action closure and trend |
Table 6 (part 2). Defense-in-depth control architecture.
Table 7
30, 90, 180, and 365 days
| Horizon | Objective | Actions |
|---|---|---|
| First 30 days | Find uncontrolled risk | Inventory critical-result policies, overdue recommendations, teleradiology terms, major complications, dose alerts, MR access, emergency kits, and unresolved serious events. Name executive and clinical owners. |
| By 90 days | Install high-leverage controls | Launch one urgency taxonomy, acknowledgment and escalation rules, a follow-up registry, revised-finding workflow, procedure rescue drills, and a unified event-review taxonomy. |
| By 6 months | Measure reliability | Report control performance by site, modality, shift, and vendor. Audit high-risk reports and follow-up closures. Correct teleradiology data access, contract, and overread gaps. |
| By 12 months | Prove sustained control | Demonstrate stable improvement, complete competency cycles, conduct external or cross-site audits, test downtime and disaster response, and brief the board on residual risk and resource needs. |
Table 7. Author recommendation.
Section 7
Board and department dashboard
A liability dashboard should measure failed controls before it measures claims. Every metric needs a numerator, a denominator, an owner, a review cadence, and an escalation threshold.
Table 8
Minimum dashboard, metrics 1 to 5
| Metric | Operational definition | Cadence | Owner |
|---|---|---|---|
| Critical results acknowledged in policy time | Acknowledged alerts / all critical alerts | Weekly | Radiology operations |
| Actionable follow-up closed by due date | Completed or clinically closed / due recommendations | Monthly | Follow-up program |
| Unknown follow-up status | Unknown / all due recommendations | Monthly | Follow-up program |
| Clinically important overread discrepancy notified | Notified in policy time / important discrepancies | Monthly | Teleradiology lead |
| Major diagnostic discrepancy | Major discrepancies / reviewed exams | Monthly, stratified | Section chiefs |
Table 8 (part 1). Author recommendation. Targets should be locally validated.
Table 8
Minimum dashboard, metrics 6 to 10
| Metric | Operational definition | Cadence | Owner |
|---|---|---|---|
| Complication-to-rescue interval | Median and 90th percentile minutes | Monthly | IR and modality leads |
| Emergency readiness | Passed kit and drill checks / expected | Monthly | Nursing or technologist lead |
| MR access or screening near miss | Events per 1,000 MR encounters | Monthly | MR medical director |
| Dose-alert closure | Alerts investigated and closed / alerts | Monthly | Medical physicist |
| Serious corrective actions overdue | Overdue actions / all open actions | Monthly to board | Quality and safety |
Table 8 (part 2).
Section 8
Post-event response
When harm or a major diagnostic correction is discovered, the first obligation is clinical, the second is integrity of the record, and the third is learning and communication.
The sequence
Six steps, in order
Disclosure
Communication-and-resolution programs
These programs use transparent disclosure, investigation, explanation, apology when appropriate, and proactive compensation when warranted.
Multi-hospital studies have found positive or neutral liability effects, with no evidence of worsening liability trends. A 2026 AHRQ rapid review similarly found decreases or no significant changes in measured liability and cost outcomes, while noting limited evidence for patient-centered and safety outcomes.
Design note
These programs require coordination among legal, insurer, clinical, and leadership teams and should not be improvised by an individual clinician.
Kachalia et al. (2018); Sokol-Hessner et al. (2026).
Section 9
Limitations and unanswered questions
Claims data are an imperfect lens. Every conclusion in this report inherits the limits of the datasets behind it.
Limitations
Seven honest limits
- NPDB public-use data identify paid claims, not all allegations, dismissed cases, insurer reserves, or the full clinical record.
- Closed-claim cohorts reflect participating insurers and organizations. Case mix, defense strategy, policy limits, jurisdiction, and coding vary.
- Legal databases overrepresent appealed, published, or otherwise notable matters and cannot estimate incidence.
- Safety-event reports identify harm pathways but are not evidence of negligence, causation, or lawsuit frequency.
- Dollar values span different years and may use different inflation adjustments. They are not directly poolable.
- Associations between a factor and payment do not establish that the factor caused payment.
- The ranking is an author-developed synthesis intended for prioritization. It has not been validated as an actuarial risk score.
Research priorities
What the field still needs
Conclusion
Radiology lawsuits begin where diagnostic uncertainty meets a failed system.
The evidence does not support a risk program built only around telling radiologists to be more careful. It supports a reliability program that makes context available, identifies high-harm work, communicates urgency, verifies follow-up, standardizes procedures, rehearses rescue, governs remote coverage, and measures whether each safeguard actually worked.
If you do one thing
Fund the closed loop
Leadership decision
Fund the closed loop. If one investment must be made first, build enterprise ownership and auditable closure for critical results, revised findings, and recommended follow-up. It addresses multiple top-ranked factors at once and creates evidence that the organization acted reliably.
- It answers factor 3, communication of critical and revised findings.
- It answers factor 4, closure of recommended follow-up and incidental findings.
- It shortens factor 2, delay in diagnosis and escalation failure.
- It produces the evidence a defense needs: who was told, when, and what happened next.
Appendix A
Evidence map, factors 1 to 5
| Factor | Anchor sources | Primary signal | Confidence |
|---|---|---|---|
| 1. Diagnostic failure | Kundu 2026; Whang 2013; Baker 2014 | Highest frequency; severe outcomes | High |
| 2. Delay and escalation | Kundu 2026; Lee 2020 | 14.7% national; oncology payment association | High |
| 3. Communication | Humphrey 2022; Gale 2011; Schaffer 2024 | Harder to dismiss; prominent in teleradiology | Moderate to high |
| 4. Follow-up closure | Wandtke 2017; Mattay 2024 | Controllable failure; large closure gap | Moderate |
| 5. Teleradiology | Schaffer 2024 | Higher payment, indemnity, and death share | Moderate to high |
Table A1 (part 1). Confidence reflects the liability-ranking evidence, not the clinical importance of the hazard.
Appendix A
Evidence map, factors 6 to 10
| Factor | Anchor sources | Primary signal | Confidence |
|---|---|---|---|
| 6. IR technical injury | The Doctors Company 2020; Magnavita 2013 | Lower payment frequency, higher severity | Moderate |
| 7. Recognition and rescue | Khan 2023; ACR Contrast Manual | Anaphylaxis and extravasation dominate | Moderate |
| 8. Consent and documentation | The Doctors Company 2020; Wojdyla 2025 | Common contributing allegation | Low to moderate |
| 9. Equipment, protocol, dose | The Doctors Company 2020 | 13% of diagnostic claims in one cohort | Low to moderate |
| 10. Identity and MR safety | Delfino 2019; ACR 2026; Joint Commission | Rare but foreseeable, highly preventable harm | Low for frequency; high for hazard |
Table A1 (part 2).
Appendix B
Risk ownership, domains 1 to 4
| Risk domain | Accountable | Responsible | Oversight |
|---|---|---|---|
| Diagnostic accuracy | Section chief | Radiologists | Chief medical officer |
| Critical results | Radiology operations | Radiologists and coordinators | Quality committee |
| Follow-up registry | Program medical director | Coordinators | CMO and board |
| Teleradiology | Radiology chair | Vendor and local leads | Medical executive committee |
Table B1 (part 1). Illustrative RACI-style ownership model.
Appendix B
Risk ownership, domains 5 to 8
| Risk domain | Accountable | Responsible | Oversight |
|---|---|---|---|
| IR safety and rescue | IR medical director | IR team | Procedural safety committee |
| Contrast and sedation | Modality medical director | Technologists and nurses | Radiology quality |
| Equipment and dose | Imaging administrator | Physics and modality leads | Environment of care |
| MR safety | MR medical director | MR safety officer and technologists | Radiology chair |
Table B1 (part 2). Local bylaws, scope, and governance determine final assignments.
References
Sources, A to H
- American College of Radiology. (2025). ACR Manual on Contrast Media.
- American College of Radiology. (2026). ACR Manual on MR Safety, updated March 2026.
- American College of Radiology. (n.d.). ACR Practice Parameter for Communication of Diagnostic Imaging Findings.
- Baker, S. R., Patel, R. H., Yang, L., Lelkes, V. M., Castro, A., and Sanzone, C. F. (2014). Non-spinal musculoskeletal malpractice suits against radiologists in the United States. Emergency Radiology, 21(1), 29 to 34.
- Brenner, R. J., Bartholomew, L., and Berlin, L. (2005). Communication errors in radiology: A liability cost analysis. Journal of the American College of Radiology, 2(5), 428 to 431.
- Delfino, J. G., et al. (2019). MRI-related FDA adverse event reports: A 10-year review. Medical Physics.
- Gale, B. D., Bissett-Siegel, D. P., Davidson, S. J., Juran, D. C., and Abujudeh, H. H. (2011). Failure to notify reportable test results. Journal of the American College of Radiology, 8(11), 776 to 779.
- Geijer, H., and Geijer, M. (2018). Added value of double reading in diagnostic radiology: A systematic review. Insights into Imaging, 9, 287 to 301.
- Humphrey, K. E., Sundberg, M., Milliren, C. E., Graham, D. A., and Landrigan, C. P. (2022). Frequency and nature of communication and handoff failures in medical malpractice claims. Journal of Patient Safety, 18(2), 130 to 137.
References
Sources, K to W
- Kachalia, A., et al. (2018). Effects of a communication-and-resolution program on hospitals' malpractice claims and costs. Health Affairs, 37(11), 1836 to 1844.
- Khan, J. M., et al. (2023). Contrast agent administration as a source of liability: A legal database analysis. Radiology, 308(3), e230802.
- Kundu, A., Govind, S., Funaki, E., and Kumari, D. (2026). Insights into radiology malpractice cases in the United States from the National Practitioner Data Bank. Radiology, 320(2), e253352.
- Lee, C. S., Nagy, P. G., Weaver, S. J., Newman-Toker, D. E., and others. (2020). Breast cancer malpractice litigation: A 10-year analysis and update in trends. Clinical Imaging, 60(1), 26 to 32.
- Magnavita, N., et al. (2013). Medical liability in interventional radiology: An Italian national claim study. La Radiologia Medica, 118, 504 to 517.
- Mattay, G., Mallikarjun, K., Grow, P., et al. (2024). Communication of incidental imaging findings on inpatient discharge summaries. Journal of Patient Safety, 20(5), 370 to 374.
- Schaffer, A. C., et al. (2024). Assessment of claimant, clinical, and financial characteristics of teleradiology medical malpractice cases. Radiology, 311(1), e232806.
- Sokol-Hessner, L., et al. (2026). Programs for responding to harms experienced by patients during clinical care. AHRQ rapid review.
- The Doctors Company. (2020). Diagnostic and interventional radiology closed claims study.
- Voreis, S., Mattay, G. S., and Cook, T. S. (2022). Informatics solutions to mitigate legal risk associated with communication failures. Journal of the American College of Radiology, 19(7), 823 to 828.
- Wandtke, B., and Gallagher, S. (2017). Reducing diagnostic delay: Multistage recommendation tracking. American Journal of Roentgenology, 209(5), 970 to 975.
- Whang, J. S., Baker, S. R., Patel, R., Luk, L., and Castro, A. (2013). The causes of medical malpractice suits against radiologists in the United States. Radiology, 266(2), 548 to 554.
- Wojdyla, L. M., and Chen, J. Y. (2025). Navigating malpractice: Guide for U.S. radiologists. RadioGraphics, 45(5), e240092.
- The Joint Commission. (2026). National Performance Goals for hospitals.
Colophon
About this edition
This flipbook reproduces the investigative report Where Radiology Lawsuits Begin by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, evidence current through August 28, 2026.
- Figures 4, 5, 6, 7, 8, and 9 were rebuilt as inline vector graphics from the values published in the report.
- Figure 7 was redrawn because the published version printed its two panel titles over the chart title. The values are unchanged.
- Figures 1 and 3 present the report's own liability chain framework.
- Every dollar value, percentage, and odds ratio is reproduced exactly as published in the source report.
Where Radiology Lawsuits Begin
The top 10 primary liability factors, claim outcomes, and a defensible risk-control system
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Narration. Select Listen to hear the pages in front of you read aloud, or Play through to have the book narrate itself and turn its own pages. Every page is a recorded track, about 29 minutes of narration in all.
Executive research report · 2026
When Every Strategic Problem Becomes an Operations Problem
The expanding healthcare chief operating officer role: an integrative evidence synthesis of role expansion, governance overload, and the operating-system redesign that health systems now require.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
EXECUTIVE RESEARCH REPORT
The Expanding Healthcare
Chief Operating Officer Role
WHEN EVERY STRATEGIC PROBLEM
BECOMES AN OPERATIONS PROBLEM
An integrative evidence synthesis of role expansion,
governance overload, and the operating-system redesign
that health systems now require
Executive research report
When Every Strategic Problem Becomes an Operations Problem
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R. An integrative evidence synthesis of role expansion, governance overload, and the operating-system redesign that health systems now require. Evidence current through August 28, 2026.
THESIS AND PRIMARY FINDING
The coherence problem
The modern healthcare COO is becoming accountable not simply for more departments, but for organizational coherence across domains that no longer work independently.
PRIMARY FINDING
Broader COO scope can create value when it closes a real coordination gap and is matched by decision rights, leadership capacity, operating cadence, and independent assurance. Without those conditions, the COO becomes the health system’s organizational shock absorber: accountable for the combined operational consequences of strategies owned elsewhere.
The language of an organizational shock absorber is an original interpretive construct. It describes a role that absorbs ambiguity, interdependence, and unresolved tradeoffs so the rest of the executive system can continue operating. It makes hidden coordination work visible. It is not yet a validated measure or a causal model.
ABSTRACT
What this synthesis examines
Health systems increasingly assign chief operating officers responsibility for enterprise integration across clinical operations, digital transformation, workforce, finance, ambulatory growth, mergers, supply resilience, access, and patient experience. The empirical literature rarely measures the COO role itself, leaving practice ahead of evidence.
This critical integrative review and theory-building case synthesis examines how the mandate is changing, which structural conditions may explain that change, and when broader scope is more likely to produce coordination than overload. Evidence was drawn from peer-reviewed leadership and management research, national administrative and survey data, public organizational records, and current health-system cases through August 28, 2026.
The synthesis finds no strong causal evidence that adding or enlarging a healthcare COO role improves performance. Management quality, role clarity, information sharing, leadership-team integration, usable authority, and disciplined operating routines are the more defensible mechanisms.
The evidence boundary
No strong peer-reviewed healthcare study isolates the causal effect of adding a COO, expanding a portfolio, or assigning a particular number of functions. This study therefore rejects title inflation as an intervention and treats organizational design as the unit of analysis.
FIGURE 1
Evidence discipline for COO design. The hierarchy distinguishes inferential strength from practical usefulness. Original synthesis generates testable propositions but does not constitute empirical confirmation.
Author synthesis. Redrawn from the published figure to correct a label overlap in the bottom tier; all wording, colour, and proportion are unchanged.
FOR BOARDS AND EXECUTIVE TEAMS
Eight conclusions
Numbers 1 through 4
FOR BOARDS AND EXECUTIVE TEAMS
Eight conclusions
Numbers 5 through 8
The COO’s job is no longer to run operations. It is to make the enterprise operable.Author synthesis
PART I
Problem and context
The expanding COO remit, contemporary operating pressures, and the Northwell case
- 1. When strategy reaches the care-delivery floor
- 2. Method and evidence discipline
- 3. The operating environment that is rewriting the role
- 4. From hospital operator to enterprise integrator
- 5. Northwell Health as a revelatory case
1. THE CONTEMPORARY OPERATING PROBLEM
When strategy reaches the care-delivery floor
A health system can approve a strategy in a boardroom in one afternoon. The strategy becomes real only after thousands of people reconcile it with medication schedules, clinic templates, union rules, payer edits, information-system constraints, construction phasing, credentialing, transportation, language access, and the ordinary unpredictability of illness.
Every elegant strategic sentence eventually becomes a queue, a handoff, a staffing decision, a screen, a supply request, a clinical judgment, or a patient waiting for an answer. That conversion has always been the province of operations. What has changed is the density of the conversion.
The contemporary health system is a network of acute-care facilities, physician groups, ambulatory sites, digital channels, laboratories, pharmacies, post-acute partners, payer contracts, joint ventures, and acquired organizations. Its components share patients, clinicians, capital, data, risk, and reputation, but they do not automatically share priorities or operating routines.
The chief operating officer increasingly occupies the space between those dependencies. Each discipline can optimize its own portfolio while the enterprise still fails at the points where their work meets. The COO is asked to make the parts cohere.
SECTION 1
The organizational shock absorber
Strategy owners generate commitments. Local leaders encounter conflicts. Committees defer decisions. Functional executives protect legitimate domain priorities. The COO bridges the gap through personal relationships, late-night follow-ups, repeated escalations, and informal exception management.
The arrangement can look highly effective because work continues. Its hidden cost is that the operating system depends on one person’s stamina, memory, and influence.
A meta-analysis of 169 samples found that work-demand stressors were negatively related to performance, with role ambiguity and situational constraints among the strongest relationships; overload was more negatively associated with performance among managers (Gilboa et al., 2008). The relevant distinction is not between an easy COO job and a hard one. It is between productive stretch and structurally unexecutable accountability.
GOVERNING DISTINCTION
A high-performing COO solves consequential problems. A high-reliability operating system prevents all consequential problems from requiring the COO’s personal attention.
SECTION 1 AND 2
The thesis in one relationship
COO scope has a contingent and probably nonlinear relationship with enterprise performance. Narrow scope can preserve functional silos and force the CEO to broker every conflict. Moderate breadth can reduce coordination costs by creating a single accountable integrator. Excessive breadth can centralize too many dependencies, slow decision-making, dilute strategic attention, and turn the COO into a bottleneck.
This is a theory of fit, not a recommendation for a universally powerful COO. Health systems may achieve integration through regional presidents, clinical-administrative dyads, service-line structures, chief transformation officers, or operating councils. The presence of a COO does not prove coherence. The absence of one does not prove fragmentation.
2. Method and evidence discipline
The study used a critical integrative review combined with a theory-building case synthesis. It was explanatory rather than effect-estimating. PRISMA 2020 principles informed transparency, but the work should not be described as a fully PRISMA-compliant effectiveness review. Three statement types were kept distinct: observed, inferred, and proposed.
TABLE 1
Evidence-grading framework
| Tier | Evidence type | Appropriate claim |
|---|---|---|
| 1 | Randomized or rigorously identified causal evidence | Causal effects of a defined operating intervention |
| 2 | Multisite longitudinal, comparative, or well-controlled observational | Associations, temporal patterns, moderators, plausible mechanisms |
| 3 | Validated cross-sectional research, audited administrative data | Prevalence, organizational characteristics, financial facts |
| 4 | Documented case, official record, executive interview, survey | Formal scope, stated intent, implementation context |
| 5 | Theoretical interpretation or original synthesis | Explanatory propositions and testable model design |
3. THE OPERATING ENVIRONMENT
Financial recovery can coexist with operating fragility
The COO remit is not expanding in a vacuum. It is expanding within a delivery environment that combines financial contradictions, migration beyond hospital walls, digital concentration without complete interoperability, cyber risk, acquisition pressure, and leadership capacity constraints.
MedPAC’s 2026 hospital analysis illustrates two financial realities. Aggregate all-payer operating margin excluding federal relief improved to 6.4% in 2024, while the fee-for-service Medicare margin remained negative at 12.1%. One quarter of hospitals had all-payer margins below 2%, while another quarter exceeded 12%.
Approximately half of the 2024 year-over-year improvement was tied to one-time 340B remedy payments, limiting inference about recurring operating strength.
For a COO, the implication is that enterprise operating decisions occur under different local economic conditions, while patients and clinicians experience a single system brand. Aggregate performance can conceal operating inequality.
FIGURE 2
Two financial realities occupy the same hospital. Aggregate cost-report margins. The 2024 all-payer increase includes a material one-time 340B effect.
Source: MedPAC (2026), Chapter 3. Published figure.
SECTION 3
Care disperses while accountability remains connected
Between 2019 and 2024, the number of Medicare-participating ambulatory surgical centers rose from 5,760 to 6,436, while the number of hospital providers declined from 4,620 to 4,530. In 2024, 248 ASCs opened and 108 closed. Knee and hip procedure volume in ASCs grew 27.6% and 28.7% that year.
These counts do not describe all outpatient migration and should not be compared as equivalent facilities. They do document an operating landscape in which care, revenue, workforce, equipment, and risk are increasingly distributed across sites.
Distributed care multiplies interfaces. Scheduling rules, preauthorization, clinical pathways, sterilization, anesthesia coverage, referral management, revenue capture, post-procedure escalation, and patient transportation must work across ownership forms and geographies.
A system can expand access through outpatient growth while increasing coordination debt. The COO mandate grows when the enterprise lacks another mechanism to manage that debt.
FIGURE 3
Care disperses beyond hospital walls. Indexed counts use different provider definitions and are presented for directional operating context.
Sources: MedPAC (2026), Chapters 3 and 11. Published figure.
SECTION 3
Digital adoption is not operational integration
Certified EHR adoption among nonfederal acute-care hospitals reached 99.4% in 2024, and 91% reported the same developer across inpatient and outpatient settings. Those achievements do not mean that information flows reliably through care.
In 2023, 70% of hospitals reported engaging in all four interoperability domains at least sometimes, but only 43% did so routinely. Seventy-one percent reported routine access to needed outside information, while only 42% said clinicians routinely used that information at the point of care. Routine exchange to most or all long-term and post-acute partners was 16%; the corresponding figure for behavioral-health partners was 17%.
The gap between adoption and use is operational. It involves workflow design, identity matching, governance, data quality, training, cognitive burden, local variation, partner connectivity, and clinical trust. This is why a digital initiative so often arrives at the COO’s door.
FIGURE 4
Digital adoption is not operational integration. Indicators use mixed years and distinct survey questions and are not a literal cohort funnel.
Sources: Barker and Chang (2026); Gabriel et al. (2024). Published figure.
SECTION 3
Cyber risk has become an operating-continuity problem
From 2010 to 2023, the number of annual reported health-care breaches affecting at least 500 records rose from 216 to 745. Hacking or information-technology incidents rose from 4% to 81% of breaches. The number of affected records grew from about 6.1 million to 166.5 million, and hacking accounted for 95% of affected records in 2023.
These figures do not represent distinct patients, and the reporting year may differ from the occurrence year. They nevertheless show a change in category.
A cyber event is not contained within information security. It can interrupt medication administration, imaging, registration, laboratory routing, claims, pharmacy, payroll, communications, transfer, and emergency diversion.
Continuity requires clinical downtime procedures, command structure, alternative communications, supply and staffing contingencies, data restoration priorities, and recovery assurance. Once again, a technical risk becomes an enterprise operating problem.
FIGURE 5
Cyber exposure changed category. Records are not distinct patients, and reporting year may differ from occurrence year.
Source: Jiang et al. (2025). Published figure.
SECTION 3
Integration activity is not realized value
Consolidation creates a second kind of operating pressure: the promise that scale will translate into value. A 2025 systematic review of 37 U.S. studies found that 13 of 14 price studies reported increases, 13 of 16 cost or spending studies reported an increase or no change, and 20 of 26 quality studies reported worse outcomes or no change.
The review combined neutral and adverse results in two categories, and the underlying studies were heterogeneous and observational. The conclusion is not that every merger fails. It is that legal and financial integration does not reliably create clinical or operational value on its own.
After a transaction closes, someone must integrate credentialing, clinical pathways, supply contracts, service distribution, leadership roles, data, culture, capital planning, and community access. Acquisition strategy becomes execution debt. Broad COO mandates often represent the attempt to place that debt somewhere.
FIGURE 6
Integration does not reliably create measurable value. Bars preserve the systematic review’s reported categories and denominators.
Source: Satiani et al. (2025). Published figure.
SECTION 3
The agenda converges faster than capacity expands
In ACHE’s 2025 Top Issues survey, financial challenges ranked as the most pressing issue, followed by workforce. Government mandates, access, and technology also remained prominent.
Among workforce concerns, 59% of responding CEOs selected hospital-leader burnout, 57% selected talent management, and 50% selected succession or leadership pipeline. These are perceptions from 215 respondents out of 1,185 community-hospital CEOs, an 18% response rate. They do not measure COO burnout or a national prevalence estimate.
The data are best read as an operating paradox. Health systems need more cross-domain leadership precisely when leadership attention and succession depth are constrained. A design that responds by assigning more unresolved work to a single executive can intensify the very constraint it intends to solve.
FIGURE 7
The COO agenda converges. Lower mean rank indicates greater concern. The community-hospital CEO survey had an 18% response rate.
Source: ACHE (2026). Published figure.
FIGURE 8
Leadership capacity is an operating constraint. These are CEO-selected concerns, not measured prevalence among COOs.
Source: ACHE (2026). Published figure.
TABLE 2
Why external pressure becomes role-design pressure
| Pressure | Operating consequence | COO design question |
|---|---|---|
| Financial heterogeneity | Margin pressure, capital scarcity, site-level variation | Prioritize across unequal local conditions without masking access or safety harm |
| Care dispersion | More sites, partners, and handoffs | Standardize what must be common while preserving local responsiveness |
| Digital integration | Adoption outpaces reliable use | Convert deployment into safe workflow and measurable benefit |
| Cyber continuity | Technical incidents disrupt clinical operations | Integrate downtime, command, communications, recovery, assurance |
| Consolidation | Transactions create unpriced integration debt | Translate scale into clinical, operational, and community value |
| Leadership capacity | Burnout concern and thin succession pipelines | Distribute integration rather than centralize all dependency |
4. THE CHANGING WORK, NOT THE TITLE
From hospital operator to enterprise integrator
The traditional stereotype of the COO is a technically gifted internal operator: the executive responsible for throughput, facilities, service standards, productivity, and the daily resolution of operating problems. That work remains essential. The contemporary role increasingly combines it with a second layer: making separate executive portfolios converge across a network.
This distinction matters because broader accountability can be misread as functional takeover. The CIO retains technology architecture and security. The CFO retains financial stewardship. The CHRO retains human-capital systems. Clinical executives retain professional standards, quality, safety, and clinical integrity.
Integration is not a softer word for control. It is boundary-spanning work. It requires acquiring information across domains, translating different professional logics, resolving timing and resource conflicts, and representing the enterprise when no single function can optimize the whole. When the work is invisible, it is easily assigned without time, staff, analytics, or authority.
TABLE 3
From department operator to enterprise integrator
| Dimension | Traditional emphasis | Contemporary emphasis |
|---|---|---|
| Primary unit | Hospital or operating division | Interdependent enterprise network |
| Core promise | Reliable daily operations | Organizational coherence under change |
| Technology interface | User of systems, sponsor of implementation | Integrator of platform, workflow, safety, adoption, benefit |
| Decision mode | Direct line authority | Mixed line, matrix, dyad, and escalation authority |
| Failure signature | Local delay or process variance | Enterprise bottleneck, hidden work, repeated escalation |
| Success signature | Stable service performance | Distributed execution, timely decisions, low dependence on heroics |
TABLE 4
Current role-expansion cases, 2026
| System | Observed role pattern | Integration problem |
|---|---|---|
| Ochsner Health | Clinical network integration across 47 hospitals and 370+ centers | Payer, value-based, technology, access, care-setting continuity |
| Dartmouth Health | Consolidation of shared services and clinical operations | Integration, workforce, IT, pharmacy, facilities, supply chain |
| Orlando Health | Portfolio orchestration across states and Puerto Rico | Joint ventures, ambulatory services, remote monitoring, AI |
| UCSF Health Network | Operating specialization within a network model | Network performance plus a separate ambulatory-operations role |
| Baptist Health Jacksonville | COO as enterprise leadership and succession platform | Operations, safety, experience, growth, access, clinical alignment |
5. PRIMARY REVELATORY CASE
Northwell Health: when a technology program becomes an operating model
Northwell Health’s Epic conversion is commonly described as a $1.2 billion program intended to consolidate more than 30 legacy electronic record platforms across 28 hospitals and about 1,000 ambulatory access points.
By August 2026, COO Kevin Beiner reported that about 70,000 users had transitioned, with another major implementation wave planned for November. Northwell’s original 2023 audited estimate was approximately $1.0 billion, while later executive accounts used $1.2 billion, apparently reflecting the evolving program estimate and associated infrastructure, training, and stabilization needs.
These values should not be treated as interchangeable with annual revenue or as a completed return-on-investment calculation. The case was selected because it combines large organizational scale, extensive legacy heterogeneity, a substantial publicly reported capital commitment, and direct COO accountability for cross-functional implementation.
FIGURE 9
Northwell Health transformation scale. Program cost and user counts are current, executive-reported figures and do not constitute an independent ROI evaluation.
Sources: Northwell Health (2023, 2025); Condon (2026); Gamble (2026).
SECTION 5
Scale communicates the problem, not the outcome
The visible asset is the electronic record. The deeper transformation is the redesign of how patients enter the system, how clinicians document and order, how teams communicate, how departments coordinate, how revenue is captured, how quality is measured, and how leaders compare performance across sites.
Technical build, clinical design, workforce preparation, activation sequencing, revenue-cycle conversion, local workflow, patient communication, command response, and post-live stabilization become one tightly coupled program.
The conversion does not occur in isolation. Beiner entered the role during a senior leadership transition while Northwell was integrating seven Nuvance hospitals, expanding into new markets, developing ambulatory services, addressing workforce needs, and protecting financial performance. Each initiative has a nominal executive owner. Their operational consequences converge.
CASE INTERPRETATION
Northwell establishes the scale of the coordination problem. It does not yet establish the size of the clinical, financial, or workforce return.
TABLE 5
What should be measured before declaring success
| Domain | Credible measure | Insufficient proxy |
|---|---|---|
| Adoption | Active use by intended role and site; workarounds; task completion | Provisioned or trained accounts alone |
| Clinical safety | Medication, diagnostic, escalation, and downtime events; near misses | Aggregate incident count without volume context |
| Access and flow | Appointment availability, ED flow, transfer latency, cancellation | One system average that hides local deterioration |
| Workforce | Time on task, after-hours work, turnover, psychological safety | Satisfaction testimonials alone |
| Finance | Verified recurring benefit, revenue integrity, labor and vendor effects | Gross benefit without implementation cost |
| Equity | Stratified digital access, communication, complaints, transitions | Unstratified portal enrollment or global rating |
PART II
Evidence and theory
Management research, role theory, contingency, boundary spanning, and transformation
- 6. What the evidence says, and what it refuses to say
- 7. When integration becomes overload
6. THE TITLE IS NOT THE MECHANISM
The direct healthcare COO evidence gap
The most important finding in the literature is an absence. No strong peer-reviewed healthcare study isolates the causal effect of the COO title, of adding a COO, of expanding the role, or of assigning a particular scope.
Zuchowski et al. (2025) screened 5,430 records and included 60 studies across seven C-suite roles and six performance domains. The review found multifaceted associations and emphasized collaboration, governance, and CEO alignment. Direct COO-specific causal evidence remained thin.
This gap is not a reason to ignore the role. It is a reason to improve the construct. Studies often record whether a title exists while omitting enacted authority, portfolio heterogeneity, direct reports, time allocation, geographic reach, initiative concurrency, decision rights, and executive-team integration.
A COO can be a line operator, deputy CEO, transformation executive, clinical network integrator, successor, or ceremonial title. Treating those designs as a single exposure introduces construct error before analysis begins.
TABLE 6
Selected healthcare management evidence
| Source | Design | Finding |
|---|---|---|
| Zuchowski et al., 2025 | Scoping review, 60 C-suite studies | Effects are multifaceted and contingent; direct COO causal evidence is sparse |
| Ward et al., 2025 | Systematic review, 25 studies, 111 associations | 55 positive, 55 null, 1 negative |
| McConnell et al., 2013 | 597 U.S. cardiac units | Management practices associated with processes and lower AMI mortality |
| Tsai et al., 2015 | 103 matched hospitals | Board attention and management practices associated with quality |
| Asaria et al., 2022 | Panel of 129 English hospitals | Management quantity unrelated to quality or performance |
| Ward et al., 2026 | 36 hospitals, 20,831 neonatal admissions | No mortality association with management score |
SECTION 6
The executive relationship is part of the intervention
The most directly COO-relevant healthcare study is qualitative. Mitra et al. (2019) interviewed physician and nonphysician NHS leaders. Effective partnerships depended on trust, role clarity, interpersonal competence, relational connection, and protected time for shared learning.
A formal charter without shared interpretation will fail in practice. Overlap can be productive when parties know which decisions are shared, which are delegated, and how disagreement is resolved. It becomes corrosive when ambiguity is personalized.
Direct COO research outside healthcare is contradictory
Hambrick and Cannella (2004) found that COOs were more common when CEOs lacked operational experience, but the presence of a COO was substantially negatively associated with performance. Marcel (2009), studying 153 firms, found positive associations contingent on top-management-team characteristics. Their contradiction is the answer: a COO is not a universally beneficial structural form.
Health-system transformation research provides a counterweight to heroic leadership. Best et al. (2012) identified five recurring principles: blend designated and distributed leadership, establish feedback loops, attend to history, engage physicians, and include patients and families.
7. FAILURE MODES
When integration becomes overload
The overload problem usually begins as a success story. A strategic initiative stalls between functions. The COO resolves it. A second initiative follows. The COO becomes the trusted integrator. Boards and CEOs learn that assigning the problem to operations produces motion. Over time, the portfolio accumulates faster than work is removed, delegated, or resourced.
The next stage is attention fragmentation. The COO attends more committees, carries more exceptions, and becomes the default escalation point. Senior deputies receive less developmental space because the highest-consequence decisions migrate upward. The system appears coordinated at the top while decision latency and hidden rework increase below.
The final stage is organizational fragility. A major transformation, safety event, cyber incident, merger, or leadership departure consumes the remaining buffer. The COO becomes indispensable, and that indispensability is misread as excellence rather than a continuity risk.
TABLE 7
Failure modes of an overextended COO design
| Failure mode | Mechanism | Observable signal |
|---|---|---|
| Residual assignment | Unowned or disputed work defaults to the COO | New responsibilities appear without removing old ones |
| Accountability without authority | COO judged on outcomes controlled by peers | Repeated escalation and private negotiation |
| Heroic integration | Relationships substitute for explicit architecture | Progress stops when the COO is unavailable |
| Meeting substitution | Recurring forums replace decisions | Issues return without closure, owner, or deadline |
| Span illusion | Direct-report count hides interdependence | A small span carries enterprise-wide dependency |
| Metric theater | Green status conceals weak definitions | Benefits lack baselines, guardrails, or verification |
| Succession fragility | Integration knowledge lives in one memory | No tested deputy, coverage plan, or decision record |
SECTION 7
The countercase: a broad COO role can be rational
Scope expansion should not be interpreted as failure by default. One accountable integrator can reduce transaction costs, shorten escalation, connect clinical and administrative work, and protect the CEO’s capacity for external strategy, governance, capital, and partnerships. The problem is not breadth itself. It is the mismatch between breadth and the architecture available to carry it.
- Real interdependence that functional coordination has failed to resolve.
- Explicit integration and escalation rights that do not erase clinical or functional authority.
- Capable deputies and value-stream leaders who distribute integration near the work.
- A limit on concurrent transformation and a decision cadence that closes issues.
- Independent assurance that verifies benefits and balancing outcomes.
SYNTHESIS CONCLUSION
The expanding COO role is neither inherently progressive nor inherently dysfunctional. It is an organizational design response whose value depends on fit.
PART III
Original operating model
The Emrick COO Operating System, Executive Operating Load, and decision rights
- 8. The Emrick COO Operating System
- 9. Executive Operating Load and decision rights
8. ORIGINAL CONCEPTUAL FRAMEWORK
The Emrick COO Operating System
E-COS is designed for a health system in which the COO integrates strategies that remain substantively owned by other executives: clinical quality, labor, digital platforms, capital, ambulatory growth, revenue performance, supply resilience, and patient access. The model defines the COO as the enterprise integrator of operating commitments, not the substitute owner of every function.
E-COS extends the Emrick Structure, Process, Outcomes model by specifying the mechanisms through which a health system converts strategic intent into coordinated execution. Its dependent construct is enterprise operability: the capacity to translate priorities into timely, coordinated, safe, equitable, and financially responsible action under changing constraints.
EPISTEMIC STATUS
E-COS has not been psychometrically validated, experimentally tested, or shown to cause better performance. Its constructs and provisional decision rules are hypotheses for governance design and empirical study, not a diagnostic instrument or a basis for judging an individual executive.
FIGURE 10
The Emrick COO Operating System (E-COS). Author synthesis. The framework is conceptual and unvalidated.
Published figure. Author-developed model.
TABLE 8
E-COS architecture and the five design tests
| Component | Required artifact | Proposed mechanism |
|---|---|---|
| Mandate | Purpose, accountability, exclusions, escalation authority, review interval | Bounds the role and prevents residual assignment |
| Authority | Decision-rights inventory using Frame, Decide, Integrate, Execute, Verify | Replaces diffuse consensus with explicit ownership |
| Capacity | Value-stream leaders, analytics, program management, succession coverage | Prevents one-person dependency |
| Cadence | One transformation register; weekly, monthly, quarterly forums | Limits work in process and closes decisions |
| Learning | Defined thresholds, feedback loops, local expertise, issue closure | Shortens sensing-to-action time |
| Assurance | Independent review of safety, adoption, finance, equity, workforce | Detects local optimization and benefit inflation |
SECTION 8
Mechanism of action
E-COS GOVERNANCE RULE
Every material addition to COO accountability requires a corresponding subtraction, a capacity addition, or an authority decision.
E-COS is intended to interrupt a common failure sequence. Strategic priorities accumulate faster than the executive team retires them. Functional executives optimize their domains while dependencies remain unresolved. The COO absorbs those dependencies through informal relationships and repeated escalation.
E-COS substitutes explicit architecture for heroic coordination. The charter bounds accountability. Decision rights separate choosing, integrating, delivering, and verifying. Portfolio control exposes concurrency and resource collisions. Distributed capacity locates integration closer to the work. Cadence creates predictable closure. Assurance tests whether reported gains are real.
The model predicts that COO value is highest when the role increases organizational coherence without increasing organizational dependence on the incumbent.
9. MAKING MISMATCH VISIBLE
The Executive Operating Load heuristic
EOL estimates mismatch between the demands placed on an executive role and the enabling conditions supplied to it. It does not measure individual competence, effort, endurance, or worth. Its unit of analysis is the designed role within a defined period, preferably a quarter.
Demand is the geometric mean of domain breadth, change concurrency, operational interdependence, and volatility. Enablement is the geometric mean of authority clarity, leadership and analytic capacity, and governance coherence. Values above 1 indicate that rated demands exceed rated enabling conditions.
For early pilots only, values below 0.80 may be labeled capacity buffer; 0.80 to 1.00, tight balance; 1.01 to 1.24, structural strain; and 1.25 or greater, overload concern. These bands are illustrative and unvalidated. EOL should never be used alone for compensation, succession, or removal decisions.
FIGURE 11
Executive operating load rises nonlinearly. Values are illustrative and unvalidated. The heatmap makes the demand assumption visible before enablement protections are applied.
Author synthesis. Published figure.
TABLE 9
Frame, Decide, Integrate, Execute, Verify
| Right | Accountability | Design rule |
|---|---|---|
| Frame | Define question, evidence, options, constraints, and delay cost | One owner produces a decision-ready frame, not a presentation |
| Decide | Select course and accept enterprise tradeoff | One person or formal body holds the final right within a threshold |
| Integrate | Resolve dependencies across functions, markets, professions, time | Often, but not automatically, the COO; must include escalation rights |
| Execute | Deliver within approved scope, budget, and timetable | Remains with accountable operational and functional leaders |
| Verify | Test fidelity, safety, equity, workforce effect, controls, benefit | Separate from Decide and Execute as risk and irreversibility increase |
PART IV
Action and validation
Scorecard, implementation roadmap, research propositions, limitations, and conclusion
- 10. A balanced COO scorecard
- 11. Research propositions and empirical validation
- 12. Discussion, limitations, and conclusion
TABLE 10
Balanced COO scorecard
| Domain | Illustrative indicators | Required guardrails |
|---|---|---|
| Access and flow | Third-next-available; ED admit-to-bed median and 90th percentile; O/E length of stay | Readmission, mortality, LWBS; payer, language, geography, site |
| Quality and safety | Serious harm per 10,000 patient days; infection SIR; failure-to-rescue | Case mix, volume, near-miss reporting, throughput balancing |
| Workforce | Critical-role vacancy and turnover; premium labor share; intent to stay | Role, shift, site, tenure, workload, injury, safety outcomes |
| Financial stewardship | Operating margin; labor expense per adjusted unit; verified benefit | Volume and acuity normalization; safety, access, workforce effects |
| Transformation | Intended-user adoption; milestone reliability; continuity recovery time | Post-implementation harm, workload transfer, benefit persistence |
TABLE 11
E-COS implementation roadmap
| Period | Objective | Exit evidence |
|---|---|---|
| Days 0-30 | Make the current system visible: inventory accountabilities, committees, transformations, escalations | Verified scope map, load profile, decision backlog, urgent authority gaps |
| Days 31-90 | Design and ratify: charter and exclusions, decision maps, work-in-process limit, metric dictionary | Signed charter, rights maps, operating calendar, governed metrics, named owners |
| Days 91-180 | Pilot under measurement: integration forums, escalation tests, benefit and guardrail verification | Shorter decision delay, reliable closure, verified data, no guardrail deterioration |
| Days 181-365 | Institutionalize and evaluate: embed review, align incentives, test succession, retire duplicate forums | Sustained results, benefit persistence, succession coverage, board-approved agenda |
11. DESIGNED TO BE TESTED
Research propositions
- Role breadth has a nonlinear association with enterprise performance; high breadth performs worse when authority and distributed capacity are low.
- Concurrency and interdependence predict decision latency, milestone unreliability, and rework.
- Distributed integration capacity mediates the relationship between scope and executive burnout, turnover intention, and reliability.
- Governance coherence weakens the relationship between high operating demand and adverse outcomes.
- Fidelity to the Five-Right protocol shortens decision-cycle time and reduces reopened decisions.
- Quarterly EOL predicts subsequent decision delay, unplanned escalation, and transformation slippage.
A prospective, multisite mixed-method study is proposed: a 36-month panel of roughly 60 to 100 health systems stratified by ownership, teaching status, size, geographic reach, payer integration, and rurality, including systems with centralized COOs, distributed operating models, transformation-specific executives, and no enterprise COO.
12. THE ARCHITECTURE, NOT THE HERO
Discussion
The defining challenge of the contemporary healthcare COO is not breadth alone. It is the fit between enterprise operating load and the architecture available to carry it. A COO can coordinate a broad portfolio when authority is clear, the senior team is deep, governance resolves conflicts quickly, and operating information arrives in a usable cadence.
The same portfolio can become hazardous when responsibility expands faster than decision rights, analytic capacity, or delegation.
For boards and CEOs, the practical implication is to govern role architecture, not merely appoint an extraordinary individual. Scope additions should trigger an explicit review of decision authority, deputy capacity, executive interfaces, committee burden, information flow, assurance, and outcome ownership.
GOVERNANCE STANDARD
A health system should become more capable because of its COO, not more vulnerable to the COO’s absence.
13. LIMITATIONS
What this study cannot conclude
Direct empirical research linking COO role design to health-system outcomes remains sparse. Most available evidence is descriptive, cross-sectional, qualitative, or case-based. This study cannot conclude that expanded COO scope causes better or worse performance.
Public evidence is subject to selection and impression-management bias. Health systems announce appointments and transformations more often than failed implementations. The title chief operating officer is not standardized, so comparing titles without measuring enacted responsibility may create construct error.
The Northwell case is analytically informative but not representative. Its scale, capital resources, market position, and leadership infrastructure may limit transfer to smaller, rural, public, safety-net, or financially distressed systems.
The synthesis is predominantly U.S. focused. Current operating data use mixed years and definitions. E-COS and EOL are conceptual; the formulas, thresholds, and tools should not be institutionalized as validated instruments before content, construct, predictive, and implementation testing.
14. CONCLUSION
Every strategic promise eventually becomes operations
The modern healthcare COO stands at the point where strategy loses its abstraction. Digital investment becomes clinician workflow. Growth becomes geography, staffing, and capital. Access becomes a schedule, a queue, a transfer, and a patient’s ability to arrive. Quality becomes the reliability of thousands of local acts.
That location gives the COO unusual leverage and unusual exposure. A broad mandate can create coherence across functions that no longer operate independently. It can also allow the organization to avoid a redesign by assigning its unresolved interdependence to a single talented person. The distinction is visible in the architecture.
The best future COO will not be the person who personally holds the health system together. The best future COO will help build a system in which clinicians, operators, functional executives, regional leaders, patients, and boards can coordinate consequential work without waiting for one heroic center.
When every strategic problem becomes an operations problem, the answer is not to make the COO infinitely larger. It is to design an operating system worthy of the work.Final finding
APPENDIX A
Twelve questions before expanding the COO charter
- What enterprise problem exists because this COO role exists, and is that problem still current?
- Which accountabilities are explicitly outside the COO charter?
- For every assigned outcome, which decisions can the COO make, stop, integrate, or escalate?
- How many concurrent enterprise transformations depend on the same executives, analysts, or frontline teams?
- Which responsibilities will be removed, delegated, or resourced when new scope is added?
- Where does integration occur when the COO is absent for two weeks or departs unexpectedly?
- Which committees make decisions, which advise, and which should be retired?
- How long do high-consequence decisions take, and how often are they reopened?
- Who independently verifies transformation benefits and unintended harm?
- What evidence would lead the board to narrow, distribute, or redesign the role?
About this report
Author and use
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, is a healthcare executive, author, and registered radiologic technologist whose work addresses imaging service line transformation, operating-model design, workforce capacity, and healthcare leadership.
This report is a critical integrative review and theory-building case synthesis. E-COS and the Executive Operating Load heuristic are original conceptual contributions presented as testable propositions. They are not validated instruments and should not be used for employment, compensation, succession, or performance decisions.
When every strategic problem becomes an operations problem, the answer is not to make the COO infinitely larger. It is to design an operating system worthy of the work.
Click the right page or press the right arrow key to turn forward. Click the left page or press the left arrow key to turn back.
Executive research report, 2026. Figures are reproduced from the published report. E-COS and the Executive Operating Load heuristic are conceptual and unvalidated.
Executive research report · Critical integrative evidence synthesis
Hospitals Are Waiting for Washington
The strategic cost of treating Medicaid retrenchment as a temporary political crisis. A critical integrative evidence synthesis by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
- Medicaid
- Hospital finance
- Safety-net hospitals
- Rural health
- Work requirements
- PRISM
Executive research report|Critical integrative evidence synthesis
Hospitals AreWaiting for Washington
The Strategic Cost of Treating Medicaid Retrenchment as a Temporary Political Crisis
Medicaid policy · Hospital strategy · Patient accessA research project
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
August 2026
Critical integrative evidence synthesis
Relief is upside. Resilience is duty.
What a hospital should do when political relief is plausible, materially valuable, and fundamentally unreliable. Policy status and estimates are current through August 27, 2026.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Keywords
- Medicaid
- hospital finance
- safety-net hospitals
- rural health
- work requirements
- provider taxes
- state-directed payments
- consolidation
Structured abstract
Background and objective
Public Law 119-21, enacted July 4, 2025, introduced the largest ten-year federal Medicaid financing reduction in the contemporary program.
On August 27, 2026, hospital executives were reported to be pursuing a politically difficult rollback while systems varied in their willingness to restructure. The managerial risk is not limited to the size of the reduction. It arises from the interaction of delayed implementation, uneven hospital exposure, state financing constraints, coverage loss, and the temptation to treat political advocacy as a substitute for operating action.
The objective is to determine the most defensible strategic posture for hospitals facing enacted Medicaid retrenchment under political uncertainty, with particular attention to safety-net and rural institutions, essential services, workforce safety, access, and the risk of value-destroying consolidation.
Scope note: this is a critical integrative evidence synthesis, not a primary econometric study or a legal opinion. Modeled estimates are labeled as models and should not be interpreted as observed causal effects.
Decision thesis
A hospital that plans to the statute can survive a failed rescue and can use any later relief as reinvestment capital. A hospital that plans to rescue may discover too late that political hope is not liquidity.
Researcher brief
The decision is not whether Washington might act
On August 27, 2026, STAT reported that hospital leaders and lobbyists were assembling a long-shot campaign to delay, soften, or reverse nearly $1 trillion in enacted Medicaid reductions. Some systems are redesigning care, pursuing mergers, seeking revenue, and preparing for service cuts. Others are postponing structural decisions because they expect Congress to intervene.
It is a matter of whether a hospital can afford to wait.
The answer is dual-track resilience. Hospitals should advocate aggressively for targeted relief while budgeting, governing, and redesigning against the enacted statutory baseline.
Advocacy and adaptation are complements. Treating them as substitutes transfers control of liquidity, workforce, capital, and service-line decisions from the boardroom to an uncertain political calendar. It also invites the most damaging form of delay: a quiet period in which near-term results appear manageable while fixed commitments accumulate before the largest reductions arrive.
Source: Payne (2026).
Researcher brief
Five findings that change the operating posture
- The financing shock is structural, not episodic.
- KFF’s allocation of Congressional Budget Office estimates places the net federal Medicaid reduction at $911 billion from 2025 through 2034, equal to 14% of baseline federal spending. The figure already accounts for estimated interactions among provisions and excludes the $50 billion Rural Health Transformation Program.
- The timing creates strategic complacency.
- Seventy-six percent of the ten-year reduction occurs from 2030 through 2034. That backloading may protect near-term reported performance while increasing the cost of later action, particularly when labor contracts, debt, facilities, and service commitments are difficult to reverse.
- Exposure is concentrated.
- Hospitals with high Medicaid shares had an aggregate operating margin of 2.3% in 2023, compared with 7.0% among hospitals with low Medicaid shares. High-Medicaid rural hospitals were at 1.7%, and independent rural hospitals were at 0.6%. A national average conceals the institutions with the least capacity to absorb a shock.
Sources: Euhus et al. (2025); Levinson et al. (2024).
Researcher brief
Two more that change the response
- Coverage loss and payment compression strike through different channels.
- Work and reporting requirements reduce covered volume and increase uncompensated care. Provider-tax limits and state-directed-payment caps reduce states’ capacity to finance hospital payments. More frequent redeterminations and administrative friction add denials, self-pay balances, and revenue-cycle costs.
- Neither rescue nor retrenchment is sufficient by itself.
- Untargeted hospital subsidies can preserve market power or inefficient capacity without improving quality. Chatterjee et al. (2021) found financial improvement after Medicaid expansion but little differential improvement across multiple hospital quality measures. A credible strategy must protect essential access while requiring transparency, productivity, quality, and local accountability.
Advocacy and adaptation are complements, not substitutes.
Figure 1
Scale and composition of the reset
The $526 billion is a subset of the gross reductions and must not be added to the $911 billion. Sources: KFF analysis of CBO estimates (Euhus et al., 2025); CMS (2026c).
Figure 2
Where the reduction actually lands
Percentages refer to the distribution of the ten-year federal reduction, not annual changes in hospital revenue. Source: KFF allocation of CBO estimates.
Board actions
The next 90 days, part one
- 1Adopt an enacted-law baseline. Require the base budget and long-range financial plan to reflect current statute. Treat any rollback, waiver, favorable rule, or supplemental appropriation as upside until enacted and collectible.
- 2Build a provision-level exposure map. Connect Medicaid enrollment, payer mix, state-directed payments, provider-tax-supported revenue, disproportionate-share payments, and state budget risk to each facility and service line.
- 3Set clinical minimums before cuts are necessary. Define the emergency, obstetric, behavioral health, pediatric, trauma, diagnostic, and transfer capabilities the system will protect, along with safe staffing floors and maximum travel-time consequences.
- 4Stage irreversible decisions. Use option gates for construction, technology, debt, leases, acquisitions, and labor commitments. Every major commitment should specify the policy and cash-flow assumptions that would trigger acceleration, redesign, or pause.
Board actions
The next 90 days, part two
- 1Create a trigger dashboard. Monitor Medicaid redetermination loss, self-pay conversion, days in accounts receivable, uncompensated care, agency labor, vacancy, cash, debt covenant headroom, state budget action, and service access monthly.
- 2Target advocacy with evidence. Seek relief that is time-limited where uncertainty is temporary, targeted where fragility is concentrated, and tied to measurable access, quality, and transformation rather than across-the-board preservation of revenue.
- 3Protect reversibility. If relief arrives, use it to rebuild liquidity, stabilize essential teams, and fund delivery redesign. Do not immediately restore every fixed cost eliminated under the downside plan.
Numbering continues from the previous page. All seven actions are board-level, not delegated operational tasks.
Section one
Waiting is not the absence of strategy
The contemporary hospital dilemma begins with a rational impulse. If a federal policy is unpopular, backloaded, and already attracting bipartisan concern, why commit to painful restructuring before the political system has completed its work?
Federal health policy contains a long history of temporary patches, transition rules, exceptions, special designations, and year-end appropriations. Major hospitals are visible employers, essential infrastructure, and politically connected institutions. A board may therefore conclude that waiting preserves optionality.
That conclusion confuses political optionality with operating optionality. Political optionality belongs to lawmakers. Operating optionality belongs to the institution only while it has sufficient cash, covenant capacity, workforce credibility, and time to act deliberately.
Waiting consumes those resources. What looks like patience can become path dependence.
Section one
The consequences of being wrong are asymmetric
This report does not argue that executives who expect a fix are politically naive. It argues that the organizational consequences of being wrong do not balance.
- A hospital that restructures prudently and later receives relief
- can restore selected capacity, accelerate investment, or rebuild reserves.
- A hospital that delays and receives no relief
- may be forced into rapid layoffs, indiscriminate cuts, distressed financing, closure, or a merger negotiated from a position of weakness.
Hospital strategy must synchronize three clocks. The political clock turns on elections, coalitions, and the willingness to revisit a signature law. The implementation clock turns on regulations, state systems, redeterminations, and grandfathering schedules. The cash-flow clock turns on remittance, denials, uncompensated care, labor expense, and the time needed to redesign clinical delivery.
These clocks do not move together.
Figure 3
Politics, implementation, and cash
Source: Author synthesis of Public Law 119-21 implementation dates, KFF policy analyses, and CMS proposed-rule timing as of August 27, 2026.
Normative premise
The endpoint is not organizational survival for its own sake. It is reliable access to safe, necessary care.
Some institutions and service configurations should change. The obligation is to prevent fiscal change from producing avoidable clinical harm or unaccountable market concentration.
Section two
The evidence hierarchy used in the synthesis
| Tier and evidence type | Use in this report | Primary caution |
|---|---|---|
| 1. Enacted law, CBO estimates, CMS rules and official program documents | Policy status, effective dates, federal estimates, program design | Budget estimates are forecasts and depend on state behavior |
| 2. Systematic reviews and strong quasi-experimental studies | Direction and credibility of coverage, closure, employment, staffing, and consolidation effects | External validity may vary by state, market, and period |
| 3. National cost-report analyses and multi-hospital observational studies | Distribution of margins, uncompensated care, and safety-net exposure | Accounting definitions and reporting lags |
| 4. Transparent simulations and policy models | Magnitude ranges and scenario sensitivity | Results depend on assumptions and are not observed effects |
| 5. Industry reports and current journalism | Operational context, stakeholder positions, and current behavior | Selection, advocacy interests, and incomplete financial detail |
Table 1. Higher tiers were not automatically treated as more relevant. Direct applicability, methods, recency, and source independence were considered together.
Section two
Four adjudication rules
- 1Statutory and proposed regulatory estimates were kept separate. CMS’s estimated $510 billion federal savings from its May 2026 state-directed-payment proposal is not added to KFF’s $911 billion estimate of enacted-law Medicaid reductions, because the proposed rule implements and extends mechanisms that overlap with the law.
- 2Gross provision estimates were not summed to infer a net total when CBO reported interaction effects.
- 3Simulations were reported as ranges with assumptions and were never described as realized losses.
- 4Associations between hospital finances and outcomes were not converted into causal claims without supporting design.
Limitations: no national model can reproduce each state’s Medicaid financing architecture. Work-requirement evidence comes largely from Arkansas, which had a shorter, narrower policy than the national statute. The strategy developed here is conditional by design.
Section three
The headline number, and what it is not
Before interactions, the Medicaid provisions sum to $990 billion. The net figure follows $79 billion in estimated Medicaid interaction effects. Provisions that apply only to expansion states account for $526 billion of gross reductions, and Louisiana, Illinois, Nevada, and Oregon face estimated reductions of at least 19% of baseline federal Medicaid spending.
These are changes in federal spending, not direct estimates of hospital revenue loss.
A federal reduction can be absorbed by several actors. States can replace lost funding, or reduce eligibility, benefits, rates, or supplemental payments. Plans can change networks and contracting. Hospitals can lose payment, receive more uninsured patients, raise commercial prices where market power allows, cut cost, close services, consolidate, or draw reserves. The hospital effect is mediated, delayed, and heterogeneous.
Section three
Four channels that reach the hospital
| Policy channel and timing | Federal estimate or scale | Hospital mechanism |
|---|---|---|
| Work and reporting requirements Required by Jan. 1, 2027, with limited good-faith delay authority | $326B gross federal savings over ten years | Coverage loss, more self-pay, higher uncompensated care, reduced utilization, denials, administrative burden |
| More frequent eligibility determinations Phased under the law | $63B gross federal savings in KFF’s provision allocation | Coverage churn, payer conversion during episodes, retroactive billing risk, revenue-cycle work |
| Provider-tax restrictions Expansion-state safe harbor phases to 3.5% by FFY 2032 | $226B federal reduction for provider-tax policy changes | Reduced state financing capacity, lower supplemental payment capacity, state budget pressure |
| State-directed-payment limits Grandfathered amounts phase down from Jan. 1, 2028 | $149B enacted-law CBO estimate; CMS proposed rule estimates $510B through 2035 | Compression of managed-care supplemental payments toward Medicare-related limits |
| Rural Health Transformation Program $10B annually, FY 2026 to 2030 | $50B temporary program | Transformation grants, infrastructure, workforce, technology; not a revenue replacement |
Table 2. Amounts overlap across mechanisms and horizons. They must not be summed.
Section three
Coverage loss is a financing mechanism
The law requires states to verify, at application and renewal, that affected expansion adults complete 80 hours per month of work or qualifying community engagement, or meet an exemption. When data matching cannot establish compliance, individuals must respond to a notice and document eligibility. After termination, they must reapply and may be barred from subsidized Marketplace coverage when loss is attributed to noncompliance.
KFF reports that CBO expected the provision’s savings to arise largely from coverage loss.
Sources: Burns et al. (2025a); Sommers et al. (2019, 2020); Gangopadhyaya and Karpman (2025). The national policy is broader and implemented across different systems, so magnitude remains uncertain. The direction of administrative coverage loss is strongly supported.
Section three
State financing compression
Provider taxes and state-directed payments determine how states assemble and distribute the nonfederal share of Medicaid. All states but Alaska used provider taxes in fiscal year 2025, including hospital taxes in 47 states. CMS estimated $98.6 billion in provider-tax revenue for calendar year 2026, with $61.8 billion from hospitals.
The law effectively prohibits new taxes or increases to taxes that existed at enactment. In expansion states, the safe-harbor ceiling declines by 0.5 percentage points per year from federal fiscal year 2028 until it reaches 3.5% in 2032.
State-directed payments are capped at 100% of published Medicare rates in expansion states and 110% in nonexpansion states. Grandfathered amounts above the limits phase down beginning in 2028. CMS estimates $510 billion in federal savings through 2035 from its May 2026 proposal, against CBO’s earlier $149 billion through 2034.
KFF attributes the gap to newer payment data, an additional year, and broader proposed policies. Sources: Burns et al. (2026, 2026b).
Section three
The rural fund is transformation capital, not replacement revenue
The Rural Health Transformation Program provides $50 billion over five fiscal years, half distributed equally among approved states and half using CMS factors. Every state received a first-year award, ranging from $147 million in New Jersey to $281 million in Texas. Texas has roughly 30 times New Jersey’s rural population but received only about twice the award, because equal allocation dominates part of the formula.
KFF estimated that rural areas would experience $137 billion in federal Medicaid reductions over the decade. That is not directly comparable with a five-year transformation program, but it establishes scale.
Using the money to sustain a structurally unsupported fixed cost converts a five-year opportunity into a later cliff.
Source: Levinson et al. (2026).
Section four
Backloading produces false stability
Seventy-six percent of the ten-year Medicaid reduction is expected to occur between 2030 and 2034. When early income statements do not show the ultimate magnitude, leaders can interpret the absence of immediate crisis as evidence that the threat was overstated, that state mitigation is working, or that Congress will act before the pain becomes visible. Each interpretation may be correct. None is a substitute for downside readiness.
The cost of delay is nonlinear.
A service line can often be redesigned over two or three years through referral agreements, ambulatory migration, staffing pipelines, tele-specialty capacity, and transfer protocols. The same service, facing a sudden cash shortfall, may be closed within weeks. A capital project can be staged before debt is drawn; after construction begins, it becomes sunk cost. Vacancy management can gradually reduce labor expense; an emergency reduction can damage trust and increase turnover among the clinicians the system most needs to retain.
Time is therefore an asset that should be spent on redesign, not reassurance.
Section four
The rescue case is real, but not bankable
The argument for waiting is not frivolous. Implementation extends across election cycles. Democratic leaders have discussed rollback, and several Republican senators have expressed concern. Senator Josh Hawley introduced legislation to repeal provider-tax and state-directed-payment reductions and double the rural fund. Targeted rule changes, transition relief, rural exceptions, state waivers, or base-rate adjustments may be more politically achievable than full repeal.
The obstacles are equally concrete. Reversing a signature tax-and-spending law requires a governing coalition, a legislative vehicle, fiscal offsets or a willingness to increase the deficit, and presidential approval or veto-proof support.
Hospital political capital has weakened as lawmakers scrutinize high commercial prices, executive compensation, 340B practices, consolidation, and payment opacity.
Even allies distinguish fragile rural hospitals from large systems, and some congressional staff expect little to materialize. Sources: Hawley (2025); Payne (2026).
Figure 4
Advocacy and readiness are independent
Source: Original author framework.
Section four
Path dependence and irreversible loss
Hospital capabilities are not inventories that can be replenished immediately when funding returns. An obstetric unit depends on nursing, anesthesia, blood access, neonatal response, call coverage, transfer reliability, and sufficient volume to maintain competence. A behavioral program depends on licensed clinicians, referral relationships, payer authorizations, community trust, and step-down capacity.
Once a team dissolves, reopening can require years even when money becomes available.
A similar asymmetry applies to mergers. Satiani et al. (2025) reviewed 37 studies and found that 93% of price studies reported increases, 81% of cost or spending studies reported increases or no change, and 77% of quality studies reported reductions or no change.
Merger may still be the least harmful alternative in a failing market, but waiting until distress eliminates negotiating leverage increases the risk that the transaction solves the balance-sheet problem by creating a community value problem.
Strategic rule
Do not ask whether an action assumes repeal or assumes retrenchment. Ask whether it improves clinical value, liquidity, or reversibility in both states of the world.
A no-regrets action creates value across multiple policy outcomes: automated eligibility verification, denial reduction, regionalized low-volume specialty coverage, a clinical-minimum framework, and staged capital.
Section five
A balance-sheet problem arrives through an income-statement story
Medicaid retrenchment is commonly described as a reimbursement problem. That description is incomplete. The immediate income statement effects include lower Medicaid revenue, higher bad debt and charity care, payer-conversion friction, and higher administrative costs. The strategic effects appear in cash conversion, borrowing capacity, capital replacement, supplier terms, workforce retention, and the timing of decisions that cannot be easily reversed.
A hospital can report a positive margin and still lose the flexibility required to absorb a delayed shock.
- 1Coverage fails. Eligible people fail verification, churn at renewal, or lose coverage.
- 2The payer source changes. Hospitals continue treating emergency and clinically necessary patients while the payer becomes self-pay, retroactively uncertain, or administratively contested.
- 3State capacity falls. States have less capacity to finance supplemental and directed payments.
- 4Cash and capital tighten. Reduced margins weaken days cash on hand and capital spending.
- 5Boards respond. Labor, service, facility, contract, pricing, or transaction decisions follow, each affecting the next period.
Figure 5
Capacity differs by payer mix and place
Categories overlap and are not mutually exclusive. Source: KFF analysis of RAND Hospital Data and AHA data for 2023 (Levinson et al., 2024).
Section five
Why one strategy cannot fit the sector
For-profit hospitals had an aggregate operating margin of 14.0%, compared with 4.4% for nonprofit and 3.4% for government hospitals. System affiliation, a higher commercial share, and higher commercial prices were associated with stronger margins.
- A large system with strong commercial contracts
- can absorb temporary Medicaid compression, cross-subsidize essential services, and invest in automation.
- A government safety-net hospital with high Medicaid concentration
- may have little uncommitted cash and no commercial base to tax through price.
- An independent rural hospital
- may have a positive total margin due to nonoperating support while patient care remains structurally negative.
Sector-level advocacy that treats all hospitals as equally fragile undermines credibility and risks diverting scarce relief from the institutions whose closure would cause the greatest harm to access.
Section five
The work-requirement channel, modeled
The Commonwealth Fund commissioned a hospital finance simulation using 2023 Medicare cost reports for 2,958 acute-care hospitals in expansion states and Urban Institute coverage-loss estimates.
| Modeled effect | Range |
|---|---|
| Decline in Medicaid hospital revenue | $12.2B to $13.8B |
| Offsetting additional commercial revenue | $1.3B to $1.5B |
| Increase in uncompensated care expense | $7.0B to $8.0B |
| Modeled decline in hospital expense | $5.5B to $6.3B |
| Net decline in revenue | $10.9B to $12.4B |
Because people use less hospital care when uninsured, modeled expense also falls. The expense relief is insufficient to protect net income. Source: Haught et al. (2025).
Figure 6
The modeled margin effect, by hospital type
A simulation, not an observed causal effect. The range reflects coverage-loss assumptions and does not model all managerial responses or later public support. Source: Haught et al. (2025).
Section five
Build exposure from the state ledger upward
CBO assumed that states would replace half of certain lost federal financing in aggregate, while CMS used a lower replacement assumption in a 2026 provider-tax proposal. Neither assumption predicts a particular state.
The essential questions are concrete:
- Which payments are financed with provider-tax revenue?
- Which directed payments exceed the new Medicare-related limits?
- What is grandfathered, for how long, and under which rating period?
- Which state agencies are likely to protect base rates?
- What share of each payment is recurring, approved, pending, or subject to reconciliation?
- Which public hospitals supply intergovernmental transfers?
A provision-level map often reveals that two hospitals with identical Medicaid payer mix face different cash risk.
Section five
Cost growth narrows the adaptation window
The AHA also reports that 56% of hospital costs were tied to service lines where reimbursement fell below cost. As an industry association it has an advocacy interest, and its estimates should be interpreted accordingly. The direction of input-cost pressure is nevertheless consistent with the managerial challenge.
If Medicaid revenue is flat while labor, drugs, supplies, and administrative costs rise, the real value of payments declines before a statutory cut is apparent.
A downside plan must distinguish permanent productivity from deferred maintenance, clinical risk transfer, and costs that will return with interest. Source: AHA (2026).
Financial doctrine
A favorable margin is not resilience if it depends on delayed capital, fragile staffing, disputed receivables, or a payment stream that statute is phasing out.
Boards should govern the transition through cash and optionality, not EBITDA alone. The key question is the number of decision cycles remaining before the organization loses the ability to choose its mode of adaptation.
Section six
The hospital response is part of the policy effect
Federal estimates describe spending and coverage. They do not determine which unit closes, which shift goes unfilled, which capital project is canceled, or which patient travels farther. Those outcomes emerge from state choices and hospital responses.
Strategy is not downstream housekeeping. It is a mediator between policy and patient outcome.
Blavin (2016) found that Medicaid expansion was associated with a $2.8 million relative reduction in annual uncompensated care per hospital in the early implementation period, a 30.3% reduction from the pre-expansion mean. Lindrooth et al. (2018) associated expansion with stronger financial performance and a substantially lower likelihood of closure, particularly in rural areas.
Gaffney and Michelson (2023) found that hospitals in the highest quintiles of undercompensated care, uncompensated care, and neighborhood disadvantage had operating margins 3.4 to 6.2 percentage points lower than comparison hospitals after adjustment.
Section six
Access deteriorates before a facility disappears
A hospital can eliminate obstetrics, inpatient psychiatry, surgery, or intensive care while maintaining an emergency department or outpatient footprint. Service reduction transfers volume to neighboring facilities, lengthens transfer times, raises ambulance unavailability, and changes where clinicians practice.
The mean increase in transport time of 2.6 minutes did not reach conventional statistical significance. These are not universal mortality estimates, but they demonstrate that closure externalizes cost to patients, emergency systems, receiving hospitals, employers, and families. Sources: Miller et al. (2020); GAO (2021).
Section six
Remove work before removing clinical capacity
Labor is the largest hospital cost and therefore the first target in many contingency plans. Clinical labor is also a measure of productive capacity and a determinant of safety. Dall’Ora et al. (2023) systematically reviewed multidisciplinary staffing and found that higher physician and registered-nurse staffing was associated with lower hospital mortality.
A rapid labor reduction that does not account for acuity, skill mix, missed care, throughput, and transfer burden may turn a financial problem into a quality problem that further erodes revenue and trust.
The defensible sequence is to standardize documentation, automate eligibility and authorization tasks, reduce avoidable variation, consolidate duplicative administrative structures, redesign schedules, improve length of stay, and eliminate non-value-added processes. Only then should labor reductions be considered, with explicit staffing floors and prospective surveillance.
Even this sequence requires caution, because administrative roles often support care coordination, infection prevention, quality reporting, appeals, language access, and discharge.
Section six
Clinical-minimum tests for service-line decisions
| Domain | Board question | Guardrail |
|---|---|---|
| Emergency and trauma | What is the time-to-definitive-care effect? | Maximum transport and transfer interval; monitored diversions |
| Obstetrics and neonatal | Can births and emergencies be managed safely across the region? | No closure without a verified receiving pathway and maternal transport plan |
| Behavioral health | Where will crisis, inpatient, partial, and intensive outpatient demand go? | Track boarding hours, repeat crisis use, suicide follow-up, failed handoffs |
| Pediatrics | Are substitutes age-appropriate and geographically reachable? | Preserve stabilization, transfer, medication, and safeguarding capability |
| Diagnostics and pharmacy | Does removal delay treatment or discharge? | Backup service level and recovery-time objective |
| Workforce | Does the plan maintain safe skill mix? | Predefined staffing floors and stop-work triggers |
Table 3. The clinical minimum is a governance standard, not a promise to preserve every incumbent service in its current form.
Section six
A one-time rescue can preserve care but not structure
Alameda Health System illustrates both the value and the limitation of interim support. In 2026, the public safety-net system reversed 92 planned layoffs after Alameda County approved up to $19.3 million in one-time funding. The allocation also expanded behavioral health programs and funded an external performance audit.
Reporting identified a much larger budget deficit and a prior workforce-reduction plan tied in part to anticipated federal changes. The county action protected people and services. It did not, by itself, establish a recurring operating model.
Interim relief is most valuable when it delivers measurable transformation, not when it merely pushes the same decision to the next budget.
Equity is produced by distribution, not rhetoric. A strategy that protects enterprise margins by withdrawing from low-margin communities can improve consolidated financial performance while worsening access for the population. Every material contingency should report who loses local access, whose travel and unpaid caregiving increase, and which receiving facility absorbs the volume.
Source: Bender (2026).
Section seven
The strongest countercase
| Counterargument | What the evidence supports | Strategic implication |
|---|---|---|
| Provider taxes and directed payments can be opaque or circular | Financing complexity and weak public accountability are legitimate concerns | Advocate for transparent, access-linked, quality-linked payments rather than defending every incumbent arrangement |
| More hospital money does not guarantee better quality | Expansion improved finances without broad differential quality gains in one safety-net study | Tie support to measurable clinical capability, workforce, access, and transformation |
| Not every hospital is fragile | Margins vary sharply by ownership, commercial share, price, rural status, and affiliation | Target relief using need, alternatives, and community function |
| Some capacity should move to lower-cost settings | Low-volume inpatient models may be unsustainable or clinically weak | Right-size with verified transportation, diagnostics, escalation, and receiving capacity |
| Mergers may avert closure | Affiliation can supply capital and operating capability | Use a structured alternatives analysis and enforceable public-interest terms |
| Federal fiscal discipline matters | Medicaid financing contributes materially to federal outlays | Pair relief proposals with accountability, outcomes, and credible offsets |
Table 4. The purpose of the countercase is to improve the design of adaptation and advocacy, not to minimize the access risk created by poorly targeted retrenchment.
Section seven
Reconfiguration must work in the actual geography
State-directed payments expanded from two states in 2016 to 41 by 2026 and accounted for more than one-quarter of Medicaid managed-care spending in fiscal year 2025. CMS argues that some arrangements direct high payments to providers capable of supplying the nonfederal share without a consistent connection to beneficiary access or quality.
Rural access cannot be reduced to keeping every inpatient bed open. Ambulatory care, freestanding emergency capabilities, regional specialty networks, transportation, mobile services, telehealth, hospital-at-home, and rural emergency hospital models may improve access for some communities.
Telehealth without broadband, transportation, diagnostics, local hands, pharmacy, and an escalation pathway is not access. A regional model that does not receive capacity or have a transfer obligation is not a network.
Sources: CMS (2026a); Carroll et al. (2023).
Section eight
Strategic archetypes
| Archetype | No-regrets strategy | High-risk response |
|---|---|---|
| Independent rural hospital Low volume, thin cash, distance | Regionalize coverage, protect emergency capability, stage capital, secure transport and transfer, use rural funds for recurring-cost reduction | Wait for relief while maintaining unsupported fixed capacity; accept a distressed merger without access terms |
| Urban public safety net High Medicaid and uncompensated care | Eligibility support, state-financing map, denial control, workforce floors, city and county partnership | Across-the-board labor cuts or a one-time bailout without redesign |
| Children’s or specialty hospital Complex care, limited payer substitution | Quantify service-specific directed payments, document regional uniqueness, negotiate carveouts tied to outcomes | Assume high clinical reputation alone will produce political exemption |
| Academic health system Research and teaching cost, administrative scale | Separate vulnerable mission assets from duplicative overhead, stage capital, partner regionally | Use market power or acquisitions as the primary substitute for productivity |
| Commercially strong integrated system Reserves and price leverage, high scrutiny | Self-fund transition, target community investment, share services with fragile partners, offer transparent commitments | Seek blanket relief while preserving high prices, weak accountability, or local extraction |
Table 5. Each system should calculate its own archetype at facility and service-line level. Enterprise averages can hide internal safety-net exposure.
Section eight
Three scenarios, one base posture
Scenario A is statutory implementation. Scenario B is targeted mitigation. Scenario C is material rollback. The base operating plan should be viable under A, preserve enough reversibility to accelerate under B, and reinvest under C.
| Decision domain | A: statute proceeds | B: targeted mitigation | C: material rollback |
|---|---|---|---|
| Budget | Downside revenue and higher uncompensated-care assumptions | Release only the mitigated provision after enactment and state confirmation | Restore selected investment after cash receipt and board review |
| Capital | Pause or stage nonessential fixed commitments | Advance projects with strong cross-scenario return | Prioritize deferred safety, access, and productivity capital |
| Workforce | Protect clinical floors; redesign administrative capacity | Retain hard-to-rebuild teams where mitigation is durable | Rebuild selectively; do not recreate duplicative structure |
| Services | Execute preplanned redesign where triggers fire | Use transition time to improve reliability | Reinvest where unmet need is demonstrated |
| Transactions | Maintain alternatives and early partner dialogue | Use improved leverage to negotiate public commitments | Avoid acquisitions justified only by a temporary funding window |
Table 6. Cash receipt, not political announcement, determines when a financial assumption can be released.
Figure 7
Five linked disciplines
Source: Original author framework.
The model
What each discipline requires
- P: Plan to enacted law
- Separate enacted provisions, proposed rules, guidance, state choices, litigation, bills, and political statements. Each belongs in a different probability class. Every material assumption needs an owner, source, effective date, confidence rating, cash-flow window, and decision consequence.
- R: Retain liquidity and reversibility
- Liquidity is time translated into strategic freedom. Price reversibility explicitly: the decision rule is return on investment adjusted for reversibility, access consequence, and time to recover capability.
- I: Insulate essential access and workforce
- The clinical minimum is the set of capabilities a community must reach within clinically defensible time and reliability. Aggregate headcount targets obscure the causal chain between staffing and outcomes.
- S: Strengthen revenue and state alignment
- The fastest available revenue is often revenue already earned. Eligibility support becomes a clinical-access intervention under work reporting, because a patient may remain substantively eligible while losing coverage for documentation failure.
- M: Mobilize targeted advocacy and measurement
- Advocacy is most credible when specific about mechanism, beneficiary, geography, timing, and accountability. Measurement closes the loop through trigger discipline, not dashboard abundance.
Governance
The PRISM accountability map
| Discipline | Board question | Primary safeguard |
|---|---|---|
| Plan | What is law, what is scenario, and what assumption could fail first? | No political relief in the base budget until enacted and operationally confirmed |
| Retain | Which decisions consume option value or covenant capacity? | Protect time to act before distress |
| Insulate | Which capabilities cannot fail without material patient harm? | No service reduction without a verified substitute |
| Strengthen | Which lawful revenue and state levers remain controllable? | Do not shift financing risk to vulnerable patients |
| Mobilize | What precise change is requested, for whom, and with what accountability? | Relief tied to access, quality, and transformation |
Table 7. The compact turns policy resilience into ordinary governance. It should be reviewed at least quarterly and whenever a material trigger fires.
Roadmap
Move before certainty, but move through gates
| Window | Required decisions | Evidence of completion |
|---|---|---|
| 0 to 90 days | Reconcile exposure by provision, state, facility, and service; define three scenarios; establish clinical minimums; freeze assumptions that depend on unpassed relief | Signed exposure register, scenario budget, clinical-minimum map, assumption log |
| 3 to 12 months | Build eligibility and work-reporting support; reduce denials; stage capital; negotiate managed-care and state issues; model monthly cash | Readiness testing, denial trend, cash runway, capital gates, partner term sheets |
| 12 to 24 months | Execute preplanned service redesign where triggers fire; protect scarce teams; secure transport and receiving capacity; publish access commitments | Service-level access measures, staffing floors, transfer agreements, project milestones |
| 2028 to 2029 | Rebase after directed-payment and tax changes; renegotiate state strategy; validate whether temporary funds reduced recurring cost | Audited payment bridge, recurring savings, quality and access evaluation, decision memorandum |
| 2030 to 2034 | Operate the durable model; revisit ownership and regional structure; reinvest verified relief | Sustainable cash, regional access, outcome trend, compliance with public commitments |
Table 8. Relief changes a gate only after the legal and cash consequence is verified.
Scorecard
A red trigger must activate a protocol, not a meeting
| Domain | Leading indicator | Preauthorized response |
|---|---|---|
| Coverage | Renewal or reporting failure among patients otherwise likely eligible | Deploy outreach and navigation; escalate system defects to plans and state; track reinstatement |
| Revenue integrity | Medicaid clean-claim rate, denial rate, days in receivables | Root-cause team, payer escalation, coding and authorization correction, cash contingency |
| Financing | Supplemental-payment and provider-tax net cash bridge | Reforecast monthly cash, gate capital, activate state negotiation and liquidity plan |
| Workforce | Vacancy, turnover, agency share, scarce-competency coverage | Retention pool, cross-coverage, regional staffing compact, service-volume gate |
| Access | Transfer rejection, boarding, travel time, service days available | Pause service change, add transport or receiving capacity, activate regional escalation |
| Liquidity | Days cash, covenant headroom, monthly burn, capital commitments | Stage capital, preserve revolver access, pursue structural alternatives before distress |
| Policy | Probability-weighted value and timing of pending relief | Execute the base-plan gate while continuing targeted advocacy |
Table 9. Illustrative rules require local validation. Red should never be defined so late that only emergency action remains.
Governance
The last-responsible decision date
Every reversible decision has a last-responsible date: the latest point at which action can occur without avoidable loss of liquidity, bargaining leverage, staffing reliability, regulatory compliance, or patient safety.
Government affairs should report political probability against that date. A 40% chance of relief may be worth vigorous advocacy, yet still be too low to delay a decision that requires twelve months to implement safely.
The calculation should include implementation duration, notice and consultation, licensing, labor obligations, patient transition, vendor termination, receiving capacity, capital availability, and the time required to replace a failed strategy.
This temporal discipline prevents both panic and drift.
Governance
Relief as governed reinvestment
If Congress, CMS, or a state provides relief, the board should not allow the money to disappear into the prior cost base. Relief should move through a reinvestment waterfall.
- 1Restore breached clinical minimums and patient access.
- 2Repair liquidity and deferred safety capital.
- 3Retain scarce workforce and fund validated transformation.
- 4Reduce recurring administrative friction. Remove the work before restoring the headcount.
- 5Consider growth only if it remains viable after the relief expires.
Temporary money should not support a recurring model without a credible replacement source.
Board resolution
Advocate as though relief depends on the quality of the evidence. Operate as though survival depends on the statute. Reinvest as though the public will ask what the money protected.
Research agenda
What decision-makers still need to know
- Estimate the causal effect of each major provision separately and in interaction.
- Track administrative loss of coverage as distinct from substantive ineligibility.
- Link state financing reform to provider-level net resources, traced to the facility that retains funds.
- Evaluate the Rural Health Transformation Program against both its own goals and the larger financing environment.
- Test which hospital adaptations protect outcomes, from eligibility navigation to staged capital.
- Measure merger counterfactuals against a realistic closure, downsizing, subsidy, or independent-recovery alternative.
- Develop patient-centered measures of service retrenchment beyond travel distance.
- Study organizational waiting itself, and whether early dual-track preparation changes survival and access.
A minimum national data architecture would link Medicaid enrollment and claims, cost reports, encounter and payment data, provider taxes, directed payments, workforce data, EMS, facility and service changes, outcomes, and travel networks.
Conclusion
Relief is upside. Resilience is duty.
The political campaign described by STAT is rational. The enacted reductions are large, backloaded, uneven, and likely to place the greatest stress on hospitals that already carry high Medicaid, rural, safety-net, behavioral-health, pediatric, and uncompensated-care burdens. Hospitals should make that case with precision and persistence.
But a hospital cannot convert political possibility into booked revenue, cash, a staffed unit, or a safe transfer network.
Aggregate profitability is not protection for a high-Medicaid rural facility. Nor does financial support automatically produce quality. The ethical and strategic problem is targeting: which capabilities must be protected, which structures should change, which relief is warranted, and what public return should accompany it.
Under PRISM, failed relief does not become organizational surprise, and successful relief does not become permission to restore inertia.
Final perspective
Hospitals are justified in asking Washington to change course. Their boards are obligated to ensure that, while Washington decides, the mission is not left waiting.
KE
About this report
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Healthcare executive, author, and registered radiologic technologist. This critical integrative evidence synthesis draws on enacted law and CBO estimates, KFF and CMS analyses, MACPAC, MedPAC, Commonwealth Fund and AHA reporting, and peer-reviewed evidence on coverage, hospital finance, closure, staffing, quality, and consolidation. Policy status is current through August 27, 2026.
Suggested citation
Emrick, K. (2026). Hospitals are waiting for Washington: the strategic cost of treating Medicaid retrenchment as a temporary political crisis. Executive research report.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Copyright © 2026 Kelly Emrick. All rights reserved.
Use the arrows, the left and right keys, or the section menu. 56 pages on 28 leaves.
Health policy · Executive perspective
The Uninsured Surge Is Not a Revenue-Cycle Problem. It Is a Capacity Crisis.
Why coverage loss is already reshaping hospital demand, margins, and community access. A policy and operations perspective by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
- Uninsured patients
- Hospital capacity
- Uncompensated care
- Affordable Care Act
- Payer mix
- Rural hospitals
Health policy|Executive perspective
The Uninsured SurgeIs Not aRevenue-Cycle Problem.
It Is a Capacity Crisis.
Why coverage loss is already reshapinghospital demand, margins, and community access
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
August 2026
Policy and operations perspective
A capacity cascade, not a collections problem
A synthesis of 2026 Marketplace enrollment data, second-quarter hospital disclosures, and evidence on coverage, access, hospital finance, and rural vulnerability.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Keywords
- uninsured patients
- hospital capacity
- uncompensated care
- Affordable Care Act
- payer mix
- rural hospitals
Abstract
An incomplete framing
The rapid rise in uninsured hospital patients in 2026 is primarily framed as a payer mix and uncompensated care problem. That framing is incomplete.
Current Marketplace enrollment data and second-quarter hospital disclosures show a simultaneous decline in insured exchange volume, growth in uninsured care, and weakening demand for elective procedures. The result is not merely lost reimbursement.
It is a capacity cascade in which delayed ambulatory care returns as higher-acuity demand, emergency departments become default access points, financially productive service lines soften, and hospitals have fewer discretionary resources to sustain staffing, capital investment, and community access.
This perspective argues that leaders should manage coverage instability as an enterprise capacity risk, while recognizing that enrollment assistance and revenue-cycle interventions can mitigate but not neutralize a coverage-policy shock.
Central argument
Hospitals cannot collect their way out of a coverage shock. When insurance disappears, the operational problem shifts upstream to access and downstream to acute-care capacity.
Section one
The signal arrived in earnings before it arrived in the national rate
A hospital chief financial officer can see a coverage reversal before a national survey can confirm it. The early signal appears in exchange admissions, self-pay encounters, emergency visits, elective cancellations, charity-care demand, and the share of scheduled procedures that never reach the operating room.
The underlying disclosures were not isolated anecdotes. They formed a consistent operating signal across several of the country’s largest hospital companies.
The enrollment backdrop is now measurable. Effectuated Affordable Care Act Marketplace enrollment fell from 21.8 million in February 2025 to 19.2 million in February 2026. Every state except New Mexico experienced a decline, and states using the federal Marketplace had larger losses than states running their own enrollment platforms.
These figures do not prove that every person leaving a Marketplace plan became uninsured, but they establish the size and direction of the coverage shock. Source: Lo et al. (2026).
Figure 1
The coverage shock is measurable
Observed enrollment data. Chart constructed for this flipbook from values reported in Lo et al. (2026).
Section one
How quickly the shock migrated into operations
HCA Healthcare reported that a payer-mix shift driven primarily by patients who lost exchange coverage reduced second-quarter pretax income by approximately $400 million.
Same-facility emergency visits increased 3.6 percent, while inpatient surgeries declined 2.3 percent and outpatient surgeries declined 3.4 percent.
Those figures describe a dual pressure: more acute demand for access from people with little ability to pay, and less scheduled care from people who no longer have coverage.
Company-reported figures. Source: HCA Healthcare (2026).
Figure 2
Unscheduled demand up, scheduled demand down
Company-reported same-facility figures. Chart constructed for this flipbook from values reported in HCA Healthcare (2026).
Figure 3
The guidance moved during the quarter
Company guidance, not a realized total. Chart constructed for this flipbook from values reported in HCA Healthcare (2026).
Section one
The same direction across different portfolios
- Community Health Systems
- Reported a 2.9 percent increase in same-store adjusted admissions but a 13.2 percent year-over-year decline in adjusted EBITDA, with unfavorable payer mix among the identified causes. Executives later reported uncompensated care patients were approximately 20 percent higher than the prior year, with more than half of the admission growth attributable to uninsured patients.
- Tenet Healthcare
- Reported that lower exchange admissions partially offset revenue growth and favorable acuity, even as the company maintained strong overall performance.
- Universal Health Services
- Increased its expected 2026 exchange-related loss from about $75 million to $85 million after management concluded that virtually everyone losing exchange coverage in its markets appeared to become uninsured.
The companies differ in geography, portfolio, and financial resilience. The direction of the signal does not. Sources: Community Health Systems (2026); Tenet Healthcare (2026); Halleman (2026a, 2026b).
Editorial photograph
Where the shock lands hardest
Editorial photograph from the published report.
Section two
Coverage loss changes the care pathway
Treating uninsurance as a revenue-cycle category begins too late in the causal chain. By the time a self-pay account reaches billing, the consequential decisions have already occurred.
A patient may have postponed a diagnostic scan, stretched medication, skipped a specialty visit, deferred a procedure, or lost access to a usual source of care.
Coverage loss changes what care is sought, when it is sought, and where it can be obtained.
The most recent national access data make this behavior predictable, and because they predate the 2026 Marketplace contraction they represent a baseline vulnerability rather than a forecast.
Figure 4
What being uninsured already does to access
Observed survey data predating the 2026 contraction. Chart constructed for this flipbook from values reported in Tolbert et al. (2026).
Section two
What the research supports
A substantial body of research supports the mechanism. Medicaid expansion increased insurance coverage and access to care, particularly after implementation had matured.
A broader review concluded that insurance improves access to outpatient, preventive, and pharmaceutical care, strengthens financial security, and is associated with better health and survival.
Reversing coverage should not be expected to produce a simple mirror image in utilization. People do not stop being ill when a policy lapses.
They ration earlier, lower-cost care until symptoms, functional decline, or fear force a return to the system.
Sources: Miller and Wherry (2017); Sommers et al. (2017).
Section two
The emergency department becomes the default door
That return often occurs through the emergency department, because federal law, clinical ethics, and community obligation preserve emergency access even when payment is uncertain.
The immediate effect is visible as uncompensated care. The more important effect is substitution: care that could have been longitudinal, scheduled, and clinically coordinated becomes episodic, urgent, and operationally expensive.
Some patients will also forgo care entirely, which suppresses elective volume and creates a second source of financial loss. The hospital is therefore exposed on both sides of the access curve.
Section three
Why the balance sheet understates the damage
Uncompensated care is the most visible measure because it can be assigned a dollar value. It is not the whole loss.
- 1Demand becomes less schedulable. Preventive visits, diagnostic workups, and elective procedures weaken while unscheduled acute demand becomes more prominent. This erodes the ability to plan staffing, rooms, and throughput around predictable case volume.
- 2Service mix deteriorates. Insured surgical and procedural cases often subsidize less remunerative but essential services. When elective demand falls and self-pay acute care rises, average revenue per unit of clinical capacity declines even if total encounters remain stable.
- 3Discharge becomes more fragile. Uninsured patients may lack access to affordable medications, follow-up care, home health, durable medical equipment, or post-acute placement. The hospital must build a safer transition with fewer funded options.
Section three
Two more mechanisms the ledger misses
- 1Workforce pressure becomes harder to finance. Nurses, physicians, technicians, case managers, and financial counselors still perform the work. A coverage shock does not reduce labor intensity in proportion to payment. It reduces the discretionary margin available to recruit, retain, and redesign that labor.
- 2Capital capacity narrows. Lower cash generation competes with cybersecurity, facility renewal, clinical technology, debt service, and growth. The result may be deferred investment long before an organization reports an access crisis.
Coverage is not simply a payment source attached to a claim. It is part of the infrastructure that keeps clinical capacity economically usable.
Section three
The historical evidence points the same way
Medicaid expansion was associated with a $6.4 million decline in mean hospital uncompensated-care costs and a 2.6 percentage-point reduction in uncompensated care as a share of total expenses, with improved operating and excess margins in 2017 relative to the pre-expansion period.
Medicaid expansion was also associated with a substantially lower likelihood of hospital closure, particularly in rural markets and counties with high pre-expansion uninsurance.
Sources: Blavin and Ramos (2021); Lindrooth et al. (2018).
Framework
One signal, two operating lenses
| Observed signal | Revenue-cycle interpretation | Capacity interpretation |
|---|---|---|
| Uninsured ED growth | Higher charity care and bad debt | Acute access is replacing longitudinal access |
| Exchange admissions fall | Adverse payer-mix variance | A covered demand channel has contracted |
| Elective surgery softens | Lost contribution margin | Scheduled capacity is underused while urgent capacity tightens |
| Eligibility work rises | More account conversion effort | Clinical and administrative workflows carry new friction |
| Cash flow weakens | Less collectible revenue | Staffing, capital, and community-service options narrow |
Table 1. Both questions are legitimate. Only the second is large enough for the problem. Author synthesis of 2026 Marketplace enrollment, hospital disclosures, and coverage literature.
Section four
The capacity cascade
The emerging pattern can be understood as a six-stage cascade. Coverage loss leads to delayed or foregone ambulatory care. Delayed care changes acuity and timing. The emergency department absorbs a larger share of access.
Uncompensated treatment and weaker elective demand compress operating margin. Margin compression constrains workforce and capital choices. Those constraints reduce community capacity, which feeds back into delayed access.
A revenue-cycle framing asks how much of this self-pay balance can be converted, collected, or classified. A capacity framing asks which clinical services, locations, and patient pathways will become less stable as coverage erodes.
Not every patient will follow every stage, and local effects will vary. The cascade is an enterprise risk model, not a claim that all coverage loss results in hospitalization.
Figure 5
Six stages, and a feedback loop
Original framework developed for this report. Diagram constructed for this flipbook.
Figure 6
What the national modeling projected
Modeled estimates, not realized 2026 totals. Their value lies in showing that the expected provider impact includes both lost paid care and increased unpaid care. Source: Blavin and Simpson (2025).
Section five
The shock will be geographically unequal
The phrase hospital industry can obscure enormous variation in exposure. State Marketplace policy, Medicaid expansion status, payer mix, service portfolio, population health, and local labor supply will determine how the same federal change translates into local capacity.
In 2024, 42 percent of uninsured people younger than 65 lived in the ten states that had not expanded Medicaid. The expiration of enhanced premium tax credits was projected to have larger effects in nonexpansion states, the South, and rural communities.
A national average is an inadequate planning benchmark. Each health system needs a local exposure map.
Sources: Tolbert et al. (2026); Blavin and Simpson (2025).
Figure 7
Exposure is concentrated, not spread
Uninsured rates are observed; rural figures are projections. Chart constructed for this flipbook from values reported in Tolbert et al. (2026) and Haught et al. (2025).
Section five
The risks interact
A July 2026 Urban Institute review further widened the risk picture. It concluded that recent federal changes affecting Medicaid eligibility, Marketplace coverage, state financing, immigration, and rural transformation funding will interact to increase uninsurance, uncompensated care, workforce pressure, and financial uncertainty.
Rural and Medicaid-dependent hospitals face the greatest risk of service reductions or closure.
The early 2026 experience may therefore be a leading indicator rather than the full effect. Additional Medicaid eligibility and financing changes are scheduled to be phased in over several years, and implementation will vary by state and by litigation.
Source: Allen et al. (2026).
Editorial photograph
Mitigation is necessary and insufficient
Editorial photograph from the published report. Estimate from Tolbert et al. (2026).
Section six
Where mitigation reaches a hard boundary
A hospital that waits until discharge to discover a coverage lapse has missed opportunities for both the patient and the organization. Administrative friction creates avoidable uninsurance, and closing that gap is real work with real returns.
Yet eligibility teams cannot enroll a patient into a subsidy that no longer exists. Financial counselors cannot manufacture an affordable premium.
Charity care can protect a patient from a bill, but it does not restore paid primary care, specialty access, outpatient pharmacy coverage, or a durable network for follow-up.
Efficiency programs may partially offset margin pressure, but they can also become counterproductive if they reduce the access points most likely to prevent higher-cost acute care.
The executive question changes
A revenue-cycle framing asks how much of this balance we can convert or collect. A capacity framing asks which clinical services, locations, and patient pathways will become less stable as coverage erodes.
Both questions are legitimate. Only the second is large enough for the problem.
Executive agenda
Six actions, first three
- 1Build a coverage-loss early-warning system. Integrate Marketplace status, Medicaid redeterminations, self-pay registrations, eligibility conversion, ED utilization, scheduled-procedure cancellations, charity-care applications, and point-of-service estimates into one weekly view. Lagging bad-debt reports should confirm the signal, not originate it.
- 2Move eligibility conversion into the care pathway. Screen at scheduling, preregistration, bedside, discharge, and follow-up. Give financial counselors and case managers shared protocols so that coverage, medication access, post-acute placement, and financial assistance are addressed as a connected workflow.
- 3Protect ambulatory access during the shock. Avoid solving a near-term margin problem by shrinking the clinics, navigation, and chronic-care infrastructure that keep people out of the emergency department. Use targeted access preservation where coverage losses are concentrated.
Executive agenda
Six actions, second three
- 1Model exposure by market and service line. Separate exchange-sensitive elective demand from emergent uncompensated demand. Identify hospitals, ambulatory sites, specialties, and zip codes where a decline in coverage is most likely to destabilize staffing, throughput, or capital commitments.
- 2Align financial assistance with access strategy. Treat charity care as a clinical access policy, not only a billing policy. Simplify presumptive eligibility, reduce the documentation burden, and ensure patients are connected to a realistic follow-up source rather than discharged with an unaffordable plan.
- 3Govern the cascade at the board level. Report uninsurance alongside ED boarding, elective conversion, length of stay, discharge barriers, avoidable readmissions, workforce vacancies, and deferred capital. The board should see the relationship between coverage, operations, and community capacity.
Editorial photograph
Govern it with capacity metrics, not bad debt alone
Editorial photograph from the published report.
Measurement
What leaders should track now
- Share of self-pay encounters that were insured within the prior 12 months
- Conversion to Medicaid, Marketplace, or other coverage
- Uninsured emergency visits and admissions
- Scheduled procedures canceled for financial reasons
- Service-line contribution margin changes
- Charity-care approvals and patient liability after assistance
- Discharge delays linked to medication, equipment, home health, or post-acute funding
- Capacity decisions: closed beds, reduced clinic sessions, delayed hiring, deferred capital
Trend interpretation should be disciplined. A rise in uninsured admissions can coexist with overall volume growth. Supplemental Medicaid payments can temporarily offset a coverage-related loss without eliminating the underlying demand shift. A stable operating margin can mask reduced investment.
Governance
Be precise about what is known
The hospital response cannot remain solely defensive. Boards and executives should translate local operating evidence into policy-relevant capacity consequences: emergency access, maternity care, oncology pathways, behavioral health, rural travel distance, and the viability of community clinics.
- Observed
- The 2.6 million decline in effectuated Marketplace enrollment.
- Company-reported
- HCA’s $400 million pretax second-quarter impact.
- Modeled
- Urban Institute estimates of national spending and uncompensated-care effects.
Keeping those categories separate strengthens credibility and makes the case more persuasive.
Conclusion
A reallocation of care, not a receivables problem
The 2026 uninsured surge is not simply a rise in accounts receivable that will be solved by better collection. It is a reallocation of care from covered, coordinated, and schedulable pathways toward delayed, episodic, and often uncompensated pathways.
The immediate accounting effect is an adverse shift in the payer mix. The strategic effect is weaker clinical capacity.
Revenue-cycle excellence still matters. Eligibility conversion, financial assistance, and patient navigation will reduce preventable harm. But the executive task is larger: manage coverage instability as a system-level threat to access, workforce, capital, and community resilience.
The appropriate stance is neither certainty nor passivity. It is structured preparedness: identify exposures, preserve high-leverage access points, create rapid enrollment pathways, and define the capacity thresholds that would trigger board action.
Final perspective
Hospitals that recognize the cascade early can protect the services that absorb it. Hospitals that see only bad debt may discover the capacity crisis only after their options for responding have narrowed.
KE
About this report
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Healthcare executive, author, and registered radiologic technologist. This policy and operations perspective synthesizes 2026 Marketplace enrollment data, second-quarter hospital disclosures, and peer-reviewed evidence on insurance coverage, access to care, hospital finances, and rural vulnerability.
Suggested citation
Emrick, K. (2026). The uninsured surge is not a revenue-cycle problem. It is a capacity crisis: why coverage loss is already reshaping hospital demand, margins, and community access. Policy and operations perspective.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Copyright © 2026 Kelly Emrick. All rights reserved.
Use the arrows, the left and right keys, or the section menu. 36 pages on 18 leaves.
Executive research report · Perspective
Healthcare’s Workforce Crisis and Financial Crisis Are the Same Operating Failure
What labor and management owe patients when healthcare workers organize. A rapid evidence synthesis by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
- Collective bargaining
- Nurse staffing
- Patient safety
- Hospital finance
- Workforce markets
- Strikes
Executive research report · Perspective
Healthcare’s Workforce Crisis and Financial Crisis Are the Same Operating Failure
What labor and management owe patients when healthcare workers organize
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
The sustainable institution is the one that protects both the next patient and the people who will care for that patient.
Perspective article
Two crises, one operating failure
A rapid evidence synthesis on healthcare labor, patient safety, workforce markets, and the economics of keeping care available.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Keywords
- healthcare labor
- collective bargaining
- strikes
- nurse staffing
- workforce markets
- patient safety
- hospital finance
- organizational sustainability
Abstract
A framing that fails on both sides
Healthcare labor disputes are commonly framed as a contest between employees seeking fairness and executives protecting a budget. This perspective argues that the framing is analytically weak and operationally dangerous.
Workforce distress and organizational financial distress are interacting failures within the same care-delivery model. Healthcare workers have legitimate claims involving workload, compensation, violence, professional voice, and safety. Many healthcare organizations simultaneously face narrow operating margins, fixed or regulated payments, rising pharmaceutical and supply costs, payer administrative burden, and the obligation to maintain unprofitable but essential services.
Yet financial pressure is highly uneven, and sector averages cannot establish what a specific employer can afford. Drawing on current labor actions, national financial reports, and peer-reviewed evidence on staffing, burnout, turnover, contingent labor, strikes, and patient outcomes, this report proposes a balanced labor compact grounded in entity-level transparency, acuity-based staffing, market-responsive compensation, shared productivity gains, and measurable continuity protections.
Central thesis
The workforce crisis and the financial crisis are not rival explanations. They are two expressions of a care-delivery system that has failed to reconcile clinical need, labor scarcity, payment constraints, and operating capability.
Sustainable bargaining must treat patient outcomes, workforce conditions, and financial viability as a single operating obligation.
Executive summary
Four propositions
- The false choice must end.
- Supporting healthcare workers and protecting organizational solvency are not mutually exclusive positions. The organization that cannot retain qualified people will lose clinical capacity. The organization that cannot finance its obligations will eventually lose services, jobs, and community trust.
- Financial stress is real, but unevenly distributed.
- Total hospital expenses rose 7.5% in 2025 while hospital price growth was 3.3%. MedPAC reported a 6.5% all-payer operating margin in 2024 alongside a negative 12.1% aggregate Medicare fee-for-service margin. The meaning of financial stress depends on payer mix, efficiency, liquidity, and the specific entity being negotiated with.
- Staffing belongs in the quality infrastructure.
- Better staffing coverage has been associated with lower rates of mortality, readmissions, adverse events, and burnout. Labor demands concerning staffing cannot be dismissed as ordinary compensation claims, but fixed ratios alone are not a complete operating model.
- Bargaining must become a form of operating design.
- A contract should not maximize either wages or margin. It should preserve safe care, a stable workforce, service continuity, and the capacity for reinvestment.
Section one
The conflict is not labor versus patients
A hospital is one of the few institutions asked to be ready before demand is known.
Its doors must open at three in the morning. Its pharmacy must stock medications whose prices it does not control. Its intensive care unit must remain prepared even when not every bed is occupied. Its clinicians must make consequential decisions under pressure, and its staff must continue working when the rest of the community is at home. That readiness is expensive, but it is also the moral promise at the center of healthcare.
A strike vote enters this environment as both an economic action and a moral indictment. Workers are not simply saying that they want more. They often say that the institution has asked them to carry too much risk for too long: too many patients, too few colleagues, too much violence, too little recovery time, or too little influence over the conditions under which professional judgment must be exercised.
These complaints deserve to be heard as evidence about the operating system, not dismissed as rhetoric.
Section one
Two kinds of evidence
The balance sheet presents a different kind of evidence. It records debt service, payer mix, cash reserves, pension obligations, drug expense, capital replacement, and the cumulative effect of recurring wages. It has no language for exhaustion or moral injury. Yet ignoring it does not make its constraints disappear.
A health system can report billions of dollars in revenue and still have only a narrow operating margin available for reinvestment. Conversely, a large consolidated system can invoke industry hardship even when the local or parent entity has substantial liquidity, investment income, or strategic flexibility.
The central error is to force a choice between these realities. Workforce well-being, patient safety, and financial sustainability are not three independent priorities that leaders can rank according to the crisis of the week. They are mutually dependent conditions.
The practical question is not which side has the more emotionally persuasive story. It is whether labor and management can reach an agreement that recognizes all costs, including those that do not initially appear in payroll.
Section two
Organizing momentum reflects a crisis of voice
The resurgence of healthcare organizing is not confined to nurses. Resident physicians, attending physicians, pharmacists, technologists, therapists, laboratory professionals, and support-service employees are increasingly using collective bargaining to recover influence within organizations that have become larger, more consolidated, and more administratively distant.
Schulman and Richman (2024) describe physician unionization as a natural consequence of consolidation and corporatization. The observation matters because unionization is often treated as an external threat when it may be an internal diagnostic signal: employees no longer believe that ordinary governance channels can improve working conditions.
Among 986 respondents at nonunionized institutions; 1,235 respondents overall. Source: Barger et al. (2025).
Figure 1
Organizing interest among resident physicians
Source: Barger et al. (2025). These data do not establish that unionization will solve every problem. They establish that the employment model is producing broad demand for collective voice.
Section two
Larger, longer, and more operationally consequential
Compensation remains central, but current disputes increasingly include enforceable staffing, workplace violence, benefit protection, artificial intelligence, missed breaks, scheduling, and professional authority. These are operating-model demands rather than isolated wage requests.
| Setting | Scale | Outcome |
|---|---|---|
| Kaiser Permanente California and Hawaii, 2026 | About 31,000 nurses and other professionals; open-ended action from January 26 | 21.5% across-the-board increases over four years. Kaiser estimated strike-related costs above $1 billion. |
| New York City Three major systems, 2026 | About 15,000 nurses; the final NewYork-Presbyterian group remained out for 41 days | Raises of more than 12% over three years, staffing improvements, benefit protections, and AI safeguards. |
| Providence Oregon Eight hospitals, 2025 | About 5,000 nurses and physicians; more than six weeks | Immediate increases up to 22%, missed-break penalties, and acuity-based workload provisions. |
| Cleveland and Connecticut Active signals, August 2026 | Strike authorization at Cleveland Clinic Lutheran Hospital; continued picketing at Lawrence + Memorial Hospital | Organizing pressure persists even where employers report material financial constraints. |
Sources: Reuters (2026a, 2026b, 2026c), Associated Press (2025, 2026), Axios (2026), CT Insider (2026). Employer and union financial claims are reported as claims unless independently audited.
Section three
The financial argument is real, but generic claims are not enough
Healthcare labor organizations need to understand the peculiar economics of institutions that cannot close at night, cannot choose only profitable patients, and cannot raise most prices whenever costs increase.
The American Hospital Association estimates that workforce spending accounted for about $1.009 trillion in 2025, or approximately 60% of the spending categories included in its analysis. Because the AHA is an industry advocacy organization, these estimates should be treated as sector evidence rather than neutral proof of any employer position.
A wage increase is not a one-time purchase. It becomes the new base for overtime, retirement contributions, payroll taxes, differentials, paid leave, and the next bargaining cycle.
Source: American Hospital Association (2026), using industry benchmarks and BLS data.
Figure 2
Every major input grew faster than price
Source: American Hospital Association (2026).
Figure 3
Workforce is the largest single line
Source: American Hospital Association (2026). A recurring labor settlement enters an expense structure already affected by faster growth in medications, supplies, technology, insurance, and payer administration.
Section three
Reimbursement does not rise when labor costs rise
Medicare and Medicaid rates are established through public payment systems, while commercial contracts may be fixed for multiple years. Payer denials, prior authorizations, and delayed payments can add administrative expenses after care has already been delivered.
Even if the estimate is debated, the underlying point is operationally familiar: revenue-cycle friction consumes cash and labor that cannot be deployed at the bedside.
MedPAC provides a useful counterweight to a one-sided financial story. A hospital may lose money on Medicare services while generating a positive overall operating margin through commercial payments, outpatient drugs, scale, efficiency, or other revenue.
Sources: American Hospital Association (2026); Medicare Payment Advisory Commission (2026).
Figure 4
Medicare margins vary by efficiency and remain negative
Source: Medicare Payment Advisory Commission (2026).
Section three
The distribution of risk is profoundly unequal
Kaufman Hall reported an adjusted year-to-date operating margin of 1.7% through March 2026, up from 1.3% in February but below the prior year pace. It identified payer-mix erosion, bad debt, charity care, and drug expense as continuing pressures.
Rural hospitals face a still more fragile position. In a study of 2,311 rural hospitals:
The financial risk is therefore genuine. Its distribution is not.
Sources: Kaufman Hall (2026); Malone et al. (2025).
Section three
Financial credibility requires entity-level evidence
Sector averages cannot decide a local contract.
Kaiser Permanente earned $9.3 billion in net income in 2025, even as the organization argued that union proposals created unsustainable recurring costs. Yale New Haven Health, by contrast, recorded a $196.8 million operating loss for fiscal year 2025 while bargaining with workers at Lawrence + Memorial Hospital.
Neither number settles the dispute. Net income can include investment results and nonrecurring events, while a consolidated loss can obscure the performance and resource flows of individual entities.
The lesson is that both parties need the same, appropriately disaggregated financial picture.
Sources: Gooch (2026); CT Insider (2026).
Framework
Five tests of a credible financial constraint
- Entity specificity
- Hospital, subsidiary, and parent-system results are separated; intercompany transfers and overhead allocation are visible.
- Liquidity
- Days cash on hand, unrestricted cash, debt covenants, pension obligations, and near-term capital needs are disclosed.
- Recurring economics
- The full multiyear cost of wages, benefits, overtime, differentials, and compression adjustments is modeled.
- Allocation consistency
- Executive compensation, consulting, acquisitions, growth capital, agency labor, and workforce requests are evaluated under consistent standards.
- Avoided-cost recognition
- Vacancy, turnover, overtime, agency premiums, closed capacity, recruitment, and orientation are included in the affordability analysis.
Author-developed financial credibility framework. A claim of financial hardship should be assessed using audited or verifiable entity-level information whenever possible.
The governing principle
Workers should not be asked to accept generic industry distress as proof that their employer cannot afford change. Leaders should not be asked to negotiate as though reimbursement, liquidity, debt, and essential-service obligations do not exist.
Section four
Staffing is a clinical intervention, not merely a labor expense
The strongest argument for workforce investment does not begin with employee satisfaction. It begins with the work that must be completed for a patient to remain safe. Observation, medication administration, mobilization, escalation, documentation, discharge preparation, infection prevention, and communication are time-dependent clinical processes.
When labor capacity falls below patient need, work is not simply delayed. Some of it is missed, compressed, transferred, or performed by people without the same familiarity, competency, or recovery time.
Muir et al. (2025) studied 14,518 nurses across 463 hospitals. Nurses in California reported a mean patient-to-nurse ratio of 3.8 compared with 4.7 in the comparison states. Every additional patient in a nurse workload was associated with 12% higher adjusted odds of burnout, and approximately half of the California lower-burnout advantage was attributed to better staffing.
The study is cross-sectional and uses pre-pandemic survey data, so it does not establish a universal causal effect.
Figure 5
Outcomes associated with staffing coverage above 90%
Source: Juve-Udina and Adamuz (2025). Observational evidence from Catalonia should not be interpreted as a randomized causal estimate.
Section four
Filling a vacancy is not the same as solving it
Pittman et al. (2025) examined 70 U.S. hospitals and found that agency nurse hours were associated with pressure ulcers and perioperative hemorrhage or hematoma. Exceeding estimated breakpoint thresholds for pressure ulcers was associated with a 6.44% increase in agency hours and a 2.09% increase in overtime hours.
Contingent labor and overtime are necessary safety valves, but sustained reliance may indicate an unstable staffing system.
A systematic review by Li et al. (2024) included 85 studies and 288,581 nurses from 32 countries. Burnout was associated with poorer safety climate, more missed care, more medication errors, more adverse events, lower patient satisfaction, and lower nurse-assessed quality.
Mauricio et al. (2025) analyzed 8,584 medical and surgical units and found that higher RN and advanced-practice nurse turnover was associated with more patient falls. Turnover changes the composition and continuity of the team even when the schedule appears numerically filled.
Figure 6
Burnout is linked to the reliability of care
Source: Li et al. (2024). High heterogeneity and predominantly cross-sectional designs limit causal inference. The evidence justifies treating burnout as an organizational risk signal with consequences beyond morale.
Section four
Safe staffing requires more than a fixed ratio
Labor organizations are right to seek enforceable boundaries when internal escalation has repeatedly failed. Management is right to worry that a rigid ratio may not account for competency, patient turnover, admissions, discharges, acuity, available supply, support staff, technology, and seasonal volatility.
The false choice is between no enforceable standard and one inflexible number.
A stronger contract combines minimum protections with:
- acuity-based workload measurement
- skill-mix rules that recognize competency and familiarity
- surge protocols agreed before the surge
- real-time escalation with a named response owner
- an independent resolution mechanism when the parties disagree
Section five
Strike evidence is mixed, and neither side should weaponize it
A work stoppage is designed to impose operational costs, but the service being disrupted may be urgent, intimate, and irreplaceable. Workers may believe that short-term disruption is necessary to correct chronic conditions that already harm patients. Employers may believe that disruption itself creates unacceptable risk. Both claims can be true in different settings.
| Evidence | Finding | Interpretive limit |
|---|---|---|
| Essex et al. (2022) Systematic review and meta-analysis; 17 observational studies | No statistically significant pooled increase in in-hospital mortality; RR 0.91 (95% CI, 0.63 to 1.31) | Very low certainty, high heterogeneity, possible protection from contingency plans or reduced admissions |
| Essex et al. (2023) Scoping review; 43 studies | Presentations and admissions commonly declined; multiple services were disrupted | Mortality alone cannot capture delayed, displaced, or canceled care |
| Gruber & Kleiner (2012) New York hospital analysis | 18.3% higher in-hospital mortality and 5.7% higher 30-day readmission among strike-period admissions | Older, setting-specific observational evidence does not establish a universal strike effect |
The defensible conclusion is that strikes reliably disrupt access and operations, while the measured mortality effect is context-dependent and uncertain.
Shared ethical obligation
Labor should design emergency coverage and continuity protections that preserve essential care. Management should address chronic safety concerns before a strike notice makes patient safety politically useful.
Section six
What labor organizations need to understand
- 1Recurring compensation compounds. A 5% increase on a $500 million payroll begins as a $25 million annual decision, but it does not end there. It resets overtime, payroll taxes, retirement contributions, differentials, paid leave, and the base for future increases.
- 2Revenue does not reprice itself. Medicare and Medicaid rates are not reopened when a contract is ratified. A financially responsible wage proposal should be accompanied by a credible plan for revenue, productivity, or cost reallocation.
- 3Essential services are cross-subsidized. Trauma readiness, obstetrics, behavioral health, teaching, infectious disease, and emergency care may not recover their full costs. Not every profitable segment represents discretionary cash.
- 4Capital is necessary, but its purpose matters. Distinguish maintenance, safety, regulatory, and growth capital. Some funds may be donor-restricted or debt-financed and unavailable for salaries.
- 5A staffing rule cannot create professionals who do not exist. BLS projects about 189,100 registered-nurse openings per year from 2024 through 2034, most driven by replacement rather than growth. Pair the standard with pipeline commitments and agreed capacity actions.
- 6Strike costs are borne by the whole system. Replacement labor, canceled care, security, transportation, housing, legal expense, and lost revenue may exceed the cost of an earlier compromise, and those dollars do not return.
Section six
What healthcare leaders need to admit
A request for labor to understand hospital finance is credible only if leadership accepts reciprocal obligations.
- Staffing is quality infrastructure.
- Labor cannot be evaluated only as an expense when staffing affects surveillance, throughput, missed care, capacity, and retention.
- Transparency precedes sacrifice.
- Leaders asking employees to moderate demands should disclose the entity-level facts needed to evaluate the request.
- Market rate is a lagging measure.
- Compensation surveys may overlook local housing costs, scheduling burdens, violence, complexity, and emerging competitors.
- The cost of vacancy belongs in the model.
- Turnover, agency premiums, overtime, recruitment, orientation, lost capacity, and management time are real economic costs.
- Safety cannot appear only during a strike.
- Chronic unsafe-assignment reports and frontline escalation should receive the urgency later applied to strike contingency planning.
- Credibility is an allocation decision.
- Executive pay, consulting, duplicated administration, acquisitions, and discretionary expansion shape whether employees believe constraints are shared.
Section six
The measurement asymmetry
Organizations often measure labor costs more precisely than the value of a stable team.
Payroll is visible every two weeks. The avoided fall, the retained expert, the prevented vacancy, the reopened bed, and the shorter stay are dispersed across accounts, cost centers, and reporting periods.
When accounting recognizes every dollar of compensation but only fragments the return from workforce stability, the organization systematically underinvests in people.
This is not an argument that every workforce request is affordable. It is an argument that the affordability model most organizations use is incomplete on one side of the ledger and precise on the other, and that the imbalance is not neutral.
Figure 7
A shared operating model for bargaining
Original framework developed for this report.
The compact
Six domains, three columns of obligation
| Domain | Workforce obligation | Employer obligation |
|---|---|---|
| Patient safety and access | Safety planning, competency standards, emergency coverage, continuity protocols | Safe capacity, supplies, supervision, escalation pathways, timely corrective action |
| Staffing and workload | Acuity-based deployment, appropriate flexibility, evidence-based skill mix | Enforceable minimums, filled vacancies, limits on unsafe assignment, funded support roles |
| Compensation | Recognize recurring cost, internal equity, and compression implications | Current local market evidence; correct persistent inequity or wage compression |
| Financial viability | Consider payer mix, liquidity, debt, capital, essential-service obligations | Audited entity-level transparency and consistent allocation logic |
| Productivity and redesign | Participate in appropriate work redesign and technology adoption | Remove unnecessary work, protect professional judgment, return verified gains to care teams |
| Dispute continuity | Provide notice and preserve agreed emergency coverage | Avoid retaliation, prepare safe contingencies, use expedited resolution |
Author-developed framework based on the evidence synthesized in this report. Shared measures for each domain are specified in the full report.
The compact
Seven mechanisms make it operational
- 1Create a shared financial workbook. Agreed definitions for margin, liquidity, agency expense, vacancy cost, debt, capital commitments, and recurring contract cost. Reconcile competing projections before positional bargaining hardens.
- 2Reset labor-market evidence annually. Current regional data, not only lagging salary surveys. Include schedule burden, scarce competencies, geography, and competing employment settings.
- 3Give staffing councils authority. Frontline clinicians should help define acuity, skill mix, surge thresholds, escalation, and corrective action. Advisory councils without enforceable response timelines will not restore trust.
- 4Share verified gains. When reduced turnover, agency use, documentation burden, or length of stay yields measurable savings, allocate a defined portion to workforce investment.
- 5Use contingent economics carefully. Increases tied to revenue or margin triggers must be protected from unilateral reclassification or discretionary manipulation.
- 6Publish a quarterly scorecard. Joint reporting on staffing, safety, access, workforce stability, and finance makes accountability reciprocal and reduces information shock.
- 7Pair continuity with expedited resolution. No-strike or extended-notice provisions are defensible only alongside rapid mediation, interest arbitration, and enforceable safety escalation.
Governance
No single metric may win at everyone else’s expense
The scorecard should detect whether financial improvement is achieved through unsafe workloads, whether staffing improvement is achieved through unsustainable contingent expense, and whether productivity is achieved through hidden care delays.
- Patient outcomes and access
- Risk-adjusted mortality, falls, infections, medication events, readmission, ED boarding, canceled or delayed procedures.
- Workforce stability
- Vacancy, turnover, time to fill, agency share, overtime, missed breaks, injuries, violence, burnout, intent to leave.
- Financial resilience
- Operating margin, days cash, labor share, denial expense, payer mix, debt service, capital backlog.
- Operating capability
- Adjusted volume per FTE, length of stay, discharge delay, capacity closure, documentation time, skill mix.
- Dispute continuity
- Diversion hours, cancellations, temporary labor cost, grievance age, time to settlement, post-dispute recovery.
Every measure requires a numerator, a denominator, a time window, a data owner, and a threshold. A metric without a shared definition becomes another bargaining position rather than a source of truth.
Recommendations
For unions, executives, and boards
- 1Begin before bargaining begins. Establish data definitions, the scorecard, staffing governance, and financial review at least six months before contract expiration.
- 2Model three futures, not one. Compare the employer proposal, the union proposal, and a jointly engineered option over at least three years, including turnover, agency labor, closed capacity, and productivity effects.
- 3Separate inability from unwillingness. When a request is rejected, specify whether the barrier is liquidity, recurring affordability, labor supply, operational feasibility, legal restriction, or strategic choice.
- 4Protect the patient during escalation. Agree in advance on emergency services, transfers, notice, communication, and rapid mediation. Continuity must not be improvised after trust collapses.
- 5Audit whether management systems are producing the dispute. Examine spans of control, scheduling, workload allocation, grievance delay, supervisor capability, workplace violence, documentation burden, and the credibility of employee voice.
- 6Make the board accountable for both. Boards should review the labor compact scorecard with the seriousness they apply to bond covenants, quality events, and strategic capital.
- 7Treat settlement as implementation. Ratification should trigger a 30, 90, and 180 day implementation review covering staffing, compensation accuracy, grievances, access, quality, and financial performance.
Discipline
Questions that should remain open
- Which financial entity is actually responsible for funding the agreement, and what resources can it legally and practically deploy?
- What proportion of vacancies reflects inadequate pay, poor management, scheduling, workload, location, training capacity, or an absolute labor shortage?
- Which staffing measure best captures patient need in each clinical setting, and how will frontline disagreement be resolved in real time?
- What work can be eliminated, automated, delegated, or redesigned without transferring hidden burden to clinicians or patients?
- How will the parties determine whether a contract improved patient care rather than merely changing labor costs?
- What evidence would cause either party to revise its current position?
Caveats: this is a rapid evidence synthesis and perspective, not a formal systematic review. Most studies on staffing, burnout, turnover, and strikes are observational. AHA estimates describe the sector and should not be used as proof of a particular employer affordability. The report addresses acute-care hospitals most directly.
Conclusion
Sustainability with accountability
The durable healthcare contract is not the agreement that extracts the greatest possible concession from the other side.
It is the agreement that leaves enough skilled people at the bedside, enough trust in the institution, enough operating discipline in the workflow, and enough financial capacity to serve the next patient.
Healthcare workers should understand that an organization cannot spend indefinitely beyond what it receives. Leaders should understand that an organization cannot indefinitely balance its finances by asking fewer people to do more difficult work under deteriorating conditions. Both strategies eventually destroy capacity.
The deepest obligation in healthcare labor relations is not to labor or management as categories. It is to the continuity of care itself.
Final perspective
A financially insolvent hospital cannot protect its workforce. An understaffed hospital cannot protect its patients. A bargaining process that ignores either fact will eventually damage both.
KE
About this report
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Healthcare executive, author, and registered radiologic technologist. This rapid evidence synthesis draws on current labor actions, national financial reporting, and peer-reviewed evidence on staffing, burnout, turnover, contingent labor, strikes, and patient outcomes.
Suggested citation
Emrick, K. (2026). Healthcare’s workforce and financial crises are the same operating failure: what labor and management owe patients when healthcare workers organize. Executive research report.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Copyright © 2026 Kelly Emrick. All rights reserved.
Use the arrows, the left and right keys, or the section menu. 38 pages on 19 leaves.
Independent executive journal report · health policy, August 2026
Costco and the Destabilization of Medicare Advantage
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R · how a trusted retail ecosystem could reorder competition, choice, and accountability in private Medicare
Executive journal report · health policy
Costco and the destabilization of
Medicare Advantage
How a trusted retail ecosystem could reorder competition, choice, and accountability in private Medicare.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Publication information
Disclosures and interpretive standard
This report is an independent evidence synthesis. It was not commissioned by Costco Wholesale Corporation, SCAN Group, a provider, a broker, or an advocacy group. The products announced on August 18, 2026 remain subject to regulatory review, and important commercial and clinical details have not been disclosed.
Destabilization is used as an analytical hypothesis, not as an established outcome.
Claims are labelled as observed, announced, inferred, or testable. Costco is treated as a retailer and distribution partner, not as the licensed insurer or the Medicare contract holder. No inference here should be read as a finding about plan performance before a plan is approved, offered, or used.
Sources reviewed through August 24, 2026. Every figure derived rather than published is labelled as author-derived at the point of use.
Contents
What is inside
Abstract
A credible test, not a finished outcome
On August 18, 2026, Costco and the nonprofit SCAN Health Plan announced jointly branded Medicare Advantage products in two states and a Medicare supplement product in a third. States, timing, premiums, networks, formularies, and final benefits were not disclosed.
The announcement does not establish that Costco has destabilized Medicare Advantage. It does provide a credible test of whether a trusted, high-frequency retail ecosystem can serve as a new organizing layer for private Medicare.
The policy objective is not to prevent retail entry. It is to ensure that a simpler front door delivers a better coverage experience rather than a stronger sales channel.
Keywords: Costco; SCAN Health Plan; Medicare Advantage; retail health; plan choice; market concentration; forced disenrollment; consumer trust.
Author summary
Six findings leaders should carry forward
- 1The announcement is not a live national product. Medicare Advantage in two states and Medicare supplement in a third, across markets with roughly five million Medicare enrollees. States, timing, networks, premiums, and benefits remain undisclosed.
- 2Medicare Advantage is large and nationally resilient. In 2026, 35.2 million people, or 55% of eligible beneficiaries, were enrolled. More than 99% had access to a plan and 75% paid no supplemental premium.
- 3The program is also locally unstable. Individual plans fell 9%, average MA-PD options declined for a second year, and forced disenrollment reached 10% in the studied HMO and PPO population.
Author summary
Findings four to six
- 4Costco’s advantage is a consumer operating system. Membership, renewal rates, warehouses, pharmacy, optical, hearing, and digital operations create repeated touchpoints most insurers cannot reproduce. SCAN provides the regulated capabilities Costco lacks.
- 5A simpler storefront can conceal a complex product. Medicare Advantage still turns on county-level networks, formularies, prior authorization, cost sharing, and annual benefit changes. Brand confidence is not a substitute for plan-specific fit.
- 6Destabilization should be measured, not declared. The decisive evidence will be conversion, retention, network breadth, continuity, total member cost, benefit use, complaints, risk selection, rural reach, and competitor response.
The bottom line: the destabilizing force is not a warehouse logo on an insurance card.
01
Consequential, but not yet destabilizing
What is known, what is unknown, what cannot yet be claimed
The partnership creates a credible experiment in retail-led Medicare distribution. It does not yet provide evidence of enrollment, quality, access, savings, or changes in market share.
Costco is not becoming a Medicare Advantage insurer. SCAN is the insurance partner and is expected to hold the regulated relationship with CMS. Federal rules prevent a Costco membership from being bundled as a plan benefit.
Figure 1
What the announcement establishes
The design is intentionally modest. The partnership can matter strategically before it matters statistically, but market impact cannot be inferred from an approval-dependent pilot.
Table 2
Known, unknown, and why it matters
| Element | Status | Why it matters |
|---|---|---|
| Product scope | Announced | Separates managed-care risk from standardized supplemental coverage |
| Geography | Not disclosed | County economics, provider supply, and rurality determine impact |
| Premiums | Not disclosed | Retail value cannot be judged without total member cost |
| Networks | Not disclosed | Network fit is often more consequential than supplemental perks |
| Formulary | Not disclosed | Pharmacy integration does not establish drug access |
| Membership | Cannot be bundled | Plan eligibility cannot depend on retail membership |
The absence of plan-specific detail is not evidence of weakness. It is a boundary on present-tense claims.
“
The destabilizing force is not a warehouse logo on an insurance card. It is the possibility that consumer trust, distribution, and everyday health services become more decisive than the traditional insurer brand.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
02
The market Costco is entering
National resilience and local volatility coexist
Medicare Advantage is not collapsing. It is a mature, concentrated program undergoing a difficult reset in plan economics, benefit generosity, and geographic participation.
Enrollment reached 35.2 million, or 55% of eligible beneficiaries, against a 64.2 million eligible population. That reproduces the reported share at 54.8% (author-derived check).
Figure 2
Choice contracted going into 2026
Figure 3
Forced disenrollment rose sharply
Figure 4
Replacement options after termination
Table 3
Two truths about the 2026 market
Resilience
- 35.2 million enrolled, 55% of eligible
- Enrollment grew about 1.1 million, or 3%
- More than 99% had access to a plan
- 75% paid no supplemental premium
- 98.9% of terminated members had a 2026 option
Instability
- Forced disenrollment reached 10%
- Individual plan offerings fell by 346
- 122 counties had no individual MA plan, up from 81
- Average supplemental premium rose from $13 to $15
- Rural members were overrepresented among those affected
Both columns are accurate because they describe different denominators. Counties without a plan rose 41, a 50.6% increase (author-derived).
“
The program is stable from far away and disruptive up close.
Two truths, one market
Costco’s opportunity arises precisely from this gap: a market can remain nationally robust even as consumers experience repeated local disruptions and decision fatigue.
03
Why Costco changes the competitive logic
Retail loyalty, repeated contact, and health-adjacent services
Traditional Medicare Advantage competition begins with a plan and adds benefits. Costco can begin with an existing consumer relationship and attach a plan to it.
SCAN brings what cannot be improvised inside a retail chain: a Medicare contract, actuarial pricing, regulated benefit design, provider contracting, care management, grievances, appeals, risk adjustment, quality reporting, and compliance. The partnership is asymmetric by design.
Figure 5
Three footprints, three denominators
Figure 6
From insurance product to consumer ecosystem
Mechanisms
Four mechanisms of disruption
- 1Trust transfer. Confidence in a retailer becomes a shortcut for evaluating a complex plan. That lowers anxiety and search cost, and can also create a brand halo that overwhelms plan-specific tradeoffs.
- 2Distribution advantage. Warehouses, digital traffic, agents, and existing health interactions lower acquisition friction and create enrollment opportunities at moments of immediate need.
- 3Benefit integration. Pharmacy, optical, hearing, OTC, and food touchpoints make supplemental benefits easier to use. Better utilization matters only if the benefits are clinically relevant.
- 4Competitive response. Incumbents may improve service, copy the partnership, change broker strategy, consolidate, or retreat. This is the pathway from local innovation to market destabilization.
The beneficiary paradox
Retail integration can reduce the friction of finding and using benefits while increasing the risk that beneficiaries underweight networks, utilization controls, and annual plan changes.
Peer-reviewed research has repeatedly shown that Medicare plan choice is not a frictionless market. Complex choice sets may overwhelm older adults, particularly those with impaired cognition. Beneficiaries enroll in dominated choices. Trust in agents, information gaps, and the effort of switching contribute to plan stickiness.
A high-quality retail experience does not automatically establish the clinical quality of the insurance product.
04
Market structure and real destabilization
Entry can increase competition or deepen dependence
In a concentrated county, a credible new nonprofit-backed plan can pressure incumbents on service, benefit usability, and price. It can also struggle to build a sufficiently broad provider network where hospitals and physician groups have their own market power.
If a few dominant retailers become gateways, insurer concentration could be replaced by platform concentration.
The rural question is especially important. Costco’s warehouse footprint is concentrated in population centers, while recent plan exits disproportionately affected rural beneficiaries.
Figure 7
A new channel enters a concentrated market
Figure 8
Four scenarios define what destabilization would mean
Table 6
The Retail-Medicare Disruption Test
- 1Trust transfer Did consumers understand the product or rely on the brand alone?
- 2Distribution advantage Did the channel lower friction and cost, or only shift sales?
- 3Benefit integration Were retail-linked benefits actually used, and were they valuable?
- 4Clinical access Did members retain and reach the clinicians they needed?
- 5Accountability Can members identify who owns each failure?
- 6Market structure Did entry improve durable competition, or precede consolidation?
A plan is disruptive only if it changes value, access, accountability, or market behavior. Enrollment alone measures commercial traction, not public benefit. Stratify every measure by county, product, channel, age, disability, dual eligibility, language, rurality, and clinical need.
“
The future of Medicare Advantage may be decided not only by who bears insurance risk, but by who owns the trusted relationship before the risk is ever discussed.
Author conclusions
Costco is not yet the destabilizer. It is the most visible sign that the next phase of competition may be organized around consumer ecosystems rather than insurance brands alone.
References
Selected sources
- Freed, M., Fuglesten Biniek, J., Damico, A., Ochieng, N., & Neuman, T. (2026). Medicare Advantage in 2026: Enrollment update and key trends. KFF. kff.org
- Freed, M., et al. (2025, December 9). Medicare Advantage 2026 spotlight: A first look at plan offerings. KFF. kff.org
- Fuglesten Biniek, J., et al. (2026, March 13). Most Medicare beneficiaries affected by plan terminations in 2025 have robust options in 2026. KFF. kff.org
- Meiselbach, M. K., Lavallee, M., Zahn, M., Xu, J., & Polsky, D. (2026). Forced disenrollments among Medicare Advantage beneficiaries following 2026 plan exits. JAMA, 335(10), 907-909. 10.1001/jama.2026.0028
- McWilliams, J. M., Afendulis, C. C., McGuire, T. G., & Landon, B. E. (2011). Complex Medicare Advantage options may overwhelm seniors. Health Affairs, 30(9), 1786-1794. 10.1377/hlthaff.2011.0132
- Rivera-Hernandez, M., et al. (2021). Plan switching and stickiness in Medicare Advantage. Medical Care Research and Review, 78(6), 693-702. 10.1177/1077558720944284
- SCAN Health Plan. (2026, August 18). SCAN and Costco launch landmark partnership to reimagine healthcare for older adults. scanhealthplan.com
- Costco Wholesale Corporation. (2025). Annual report on Form 10-K, fiscal year ended August 31, 2025. sec.gov
Also cited: CMS (2025), Adrion (2019), Afendulis et al. (2015), Feyman et al. (2024), Meyers et al. (2019), Miller (2024), Newhouse & McGuire (2014), Reid et al. (2016), Sen et al. (2021), Tucher et al. (2024), and Weaver (2026). Derived figures are labelled at the point of use.
About the author
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, is a healthcare executive, researcher, and author whose work focuses on radiology leadership, clinical capacity, workforce strategy, operational transformation, and population health.
Interactive dashboards, assessments, and executive research briefs are published free to healthcare leaders across three platforms.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Use Turn the page, the arrow keys, or the page selector to read. Twenty-eight pages, eight figures, every derived value labelled. Now showing: two-page spread.
Independent executive journal report · health policy, August 2026
If 67% of Appealed Medicare Advantage Denials Are Overturned, What Was Wrong With the First Decision?
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R · a quality-control analysis of prior authorization, appeals, and Medicare Advantage accountability
Executive journal report · health policy
If 67% of appealed Medicare Advantage denials are overturned,
what was wrong with the first decision?
A quality-control analysis of prior authorization, appeals, and Medicare Advantage accountability.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Publication information
Disclosures and interpretive standard
This report is an independent evidence synthesis. It was not commissioned by a Medicare Advantage organization, provider, utilization-management vendor, or patient advocacy group. No original patient-level data were analyzed and no causal estimate is claimed.
The 67% statistic is treated as a quality signal that requires denominator discipline, not as proof that 67% of all denials were wrongful.
Policy and operational facts were drawn primarily from CMS, the HHS Office of Inspector General, and KFF. Clinical and workflow implications were evaluated against recent peer-reviewed literature. Because the new public reporting system remains aggregated and incomplete, conclusions are graded as established, supported, or plausible rather than presented with false precision.
Scope: Medicare Advantage medical-item and service prior authorization, with selected cross-market and peer-reviewed context. Sources reviewed through August 24, 2026. Every figure derived rather than published is labelled as author-derived at the point of use.
Contents
What is inside
Abstract
A conditional outcome, not a verdict
Medicare Advantage insurers overturned 67% of appealed standard prior authorization denials in a KFF analysis of insurer-posted 2025 metrics. The figure applies only to denials that reached an appeal, a selected subset shaped by clinical confidence, administrative capacity, patient persistence, and the availability of additional documentation.
Even with that limitation, the reversal rate is a consequential quality-control signal. It indicates that many initial decisions were not reliably final, and that the appeal supplied something the first review lacked.
Current oversight measures transaction speed more consistently than decision validity.
Keywords: Medicare Advantage; prior authorization; denial; appeal; utilization management; patient access; artificial intelligence; quality control.
Executive summary
Five findings leaders should carry forward
- 1The denominator changes the claim. The 67% applies to appealed standard denials in an enrollment-weighted sample covering 69% of Medicare Advantage enrollment. Counts were optional and service categories were not disclosed.
- 2A reversal can reflect more than one defect. Some denials are clinically incorrect. Others become approvable only after documentation is added or a more experienced clinician reviews the case.
- 3Variation argues against a single explanation. Appeal overturn rates ranged from 40% to 93% among large insurers in the same reporting sample.
- 4Speed is not accuracy. Median standard response times were about one day, yet high reversal rates persisted.
- 5The next frontier is first-decision integrity. Service-level and contractor-level outcomes, auditable counts, root-cause codes, and measures of delay and abandonment.
Table 1
At-a-glance evidence
| Signal | Value | What it does not mean |
|---|---|---|
| Standard requests denied | 12% | Not adjusted for service mix, case mix, or benefit design |
| Appealed denials overturned | 67% | Not 67% of all denials, and not a validated wrongful-denial rate |
| Insurer overturn range | 40% to 93% | Does not isolate criteria, documentation, staffing, or contractor effects |
| SNF appealed denials overturned | 95% | Not representative of all services or all calendar-year requests |
| Median standard response | 1 day | Does not measure time to clinically final resolution after a denial |
Values draw on KFF’s 2025 insurer-metric analysis and HHS OIG’s 2026 reports. The sources use different populations and should not be combined into a single trend.
01
What the 67% actually measures
The claim is powerful only when the denominator is preserved
The reported rate is a conditional outcome among appealed standard denials. It cannot identify how many of all denials were clinically inappropriate.
KFF collected the first wave of metrics insurers were required to post under the 2024 federal interoperability and prior authorization rule. The analysis covered six large insurers and about 25 million enrollees, or 69% of Medicare Advantage enrollment, which implies a program of roughly 36.2 million (author-derived).
Figure 1
The denominator guardrail
State it precisely
State precisely
Among appealed standard Medicare Advantage denials in KFF’s 2025 reporting sample, 67% were approved after appeal.
Do not claim
Sixty-seven percent of all Medicare Advantage denials were wrongful. The public data do not support that conclusion.
That description contains three selection steps. A service had to require prior authorization. The request had to receive an initial denial. The patient or provider had to appeal.
The appeal population is especially selective, because challenging a denial requires time, expertise, documentation, and a judgment that the case is worth pursuing. A high success rate among appealed cases can coexist with a large pool of unappealed denials whose validity is unknown.
Precision protects the argument from a denominator challenge while preserving the real accountability question: why did most challenged first decisions change?
Figure 2
Initial denial rates by market
Medicare Advantage had the lowest of the three enrollment-weighted standard denial rates in KFF’s sample, but a 12% rate still affects a large volume of requests.
“
Every overturned denial is evidence of decision change. A pattern of reversals demands root-cause analysis.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
A process that repeatedly relies on appeals to assemble routine clinical facts has moved quality control downstream.
Figure 3
Overturn rates by market
HHS OIG’s 2022 medical-record review found that 13% of sampled Medicare Advantage denials met Medicare coverage rules and would likely have been approved in traditional Medicare.
Table 2
Why an initial denial may later be approved
- 1Clinical judgment error The same record is reinterpreted and the service approved.
- 2Documentation interface failure Notes, results, or severity details are supplied.
- 3Policy or criteria mismatch The appeal applies Medicare rules or a different guideline.
- 4Administrative or technical error Member, provider, code, or routing data are corrected.
- 5Automation or contractor failure A human overrides a model, rule engine, or delegated reviewer.
- 6Request modification A different site, technique, dose, or level of care is approved.
More than one defect can occur in the same case. A useful audit assigns a primary cause, contributing causes, preventability, and patient impact.
02
The evidence does not describe one uniform problem
Market averages conceal plan, contractor, and service variation
When denial and reversal rates vary this widely under the same federal program, leaders should investigate decision architecture before attributing differences to beneficiary need.
The same insurer can produce different results across markets. KFF reported that UnitedHealth denied 17% of standard requests in Medicare Advantage, 11% in Medicaid managed care, and 21% in the ACA Marketplace. A corporate average is not a sufficient control.
Figure 4
Observed insurer variation
The largest range is the Medicare Advantage appeal overturn rate, a 53-point span. Public data do not adjust these endpoints for case mix, so they should trigger investigation rather than ranking.
Figure 5
Selected post-acute outcomes
Why these estimates cannot be trended
HHS OIG examined post-acute admission requests submitted to 19 Medicare Advantage organizations in June 2024. For skilled nursing facility admission, organizations denied 12% of requests, with organization-level rates from 0.4% to 23%. Requests involving nursing home residents were denied at 40%, against 11% for other beneficiaries. NaviHealth processed roughly half of the requests, denied 14%, and saw 97% of its appealed denials overturned.
Organization-level inpatient rehabilitation overturn rates ranged from 14% to 86%, a 72-point span (author-derived). A single market-wide appeal statistic can conceal service-specific decision rules, contractor effects, and patient vulnerability.
“
A quick initial denial can shorten a measured transaction while lengthening the patient’s time to a clinically final answer.
Speed is not accuracy
Median response was about one day for standard requests and 0.4 days for expedited requests, yet high reversal rates persisted.
03
What was wrong with the first decision?
Four mechanisms account for the evidence
Figure 6, in detail
The four mechanisms
Clinical judgment
Test agreement between independent reviewers using the same record and criterion, with specialty escalation for nuanced judgment.
Documentation interface
A denial overturned after predictable information is added is preventable when that information could have been captured at intake.
Policy alignment
Internal criteria must sit within federal boundaries, be evidence based, publicly accessible, and applied consistently.
Review system
The question is not whether a human touched the case. It is whether that human had the information, authority, time, and independence to disagree.
A 2025 radiation oncology study integrated payer requirements, automated retrieval, and missing-information checks into the clinical workflow. Across 6,551 cases, denial rates fell from 7.6% to 2.6%. The printed rates give a 65.8% relative reduction; the paper states 65.4%, which implies unrounded underlying rates (author-derived check).
Harm occurs before the appeal succeeds
An eventual approval can still be a poor outcome if the patient waits, deteriorates, abandons treatment, changes therapy, or bears avoidable burden.
A 2026 systematic review of 25 studies found prior authorization associated with delayed care and adverse outcomes, including disease exacerbation, preventable hospitalization, longer inpatient stays, and lower disease-free survival.
More than a quarter of those cases were authorized only after a payer-mandated change to technique or dose. Beneficiaries in the highest quartile of prior authorization exposure were 4.7 percentage points more likely to disenroll, a 44% relative difference, which implies about 10.7% against 15.4% (author-derived).
The non-appeal problem
A low appeal rate does not validate the unappealed denials. Non-appeal may indicate agreement with the decision. It may also reflect clinician triage, patient fatigue, deadline pressure, transfer to a different service, abandonment, or the cost of pursuing review.
Oversight that looks only at successful appellants risks survivorship bias.
Plans should sample unappealed denials for external clinical review and track what happened to the requested care.
04
First-decision integrity
A replacement quality standard
Table 4
Eight metrics worth building
- 1Clinically validated inappropriate-denial rate Sampled denials judged approvable by independent reviewers using the complete record.
- 2Overturn root-cause rate Every reversal assigned a primary cause and a preventability judgment.
- 3First-pass finality Requests resolved without avoidable resubmission, peer-to-peer review, or appeal.
- 4Time to clinically final decision From complete request to approval, upheld external decision, or documented abandonment.
- 5Criteria concordance Agreement among independent reviewers using the same record and criterion version.
- 6Documentation preventability Whether the missing element was available or could have been prompted before submission.
- 7Automation safety Machine-assisted decisions tested against blinded human audit, including false-denial patterns.
- 8Patient consequence Delay days, treatment modification, abandonment, out-of-pocket exposure, and complaint.
Table 5
The action agenda
| Owner | First 90 days |
|---|---|
| CMS | Standardize definitions, require numeric counts, service categories, and contractor identifiers. Preserve raw public files. |
| Plans | Code every reversal cause, identify high-volume services and contractors, and review 90th-percentile resolution time. |
| Providers | Create a denial registry, preserve criterion and submission detail, and prioritize radiology and post-acute pathways. |
| Boards | Ask for counts, not percentages alone. Separate first response from final resolution. Demand contractor stratification. |
For radiology and diagnostic services
The strongest approach is not a larger appeals team. It is order-entry alignment: expose current payer criteria to the ordering clinician, pull existing signs, symptoms, conservative-treatment history, and prior imaging from the record, and flag missing decision facts before submission. Pair technical success with medical-director review, so a workflow tool does not simply normalize a restrictive policy.
“
Appeals should function as a safety net, not as the ordinary place where the system finally becomes accurate.
Conclusion
The central defect is not that every denial was wrong. It is that the system cannot yet show whether the first decision was clinically valid, informationally complete, consistently reasoned, and safe for the patient.
References
Selected sources
- Chen, W. C., et al. (2025). Integrating prior authorization into clinical workflows for care access and practitioner experience. JAMA Network Open, 8(12), e2549093. 10.1001/jamanetworkopen.2025.49093
- Gupta, R., Fein, J., Newhouse, J. P., & Schwartz, A. L. (2024). Comparison of prior authorization across insurers. BMJ, 384, e077797. 10.1136/bmj-2023-077797
- Long, M., et al. (2026, August 13). Prior authorization metrics provide new insights into insurer practices, but gaps remain. KFF. kff.org
- Marr, J., Trivedi, A. N., & Meyers, D. J. (2026). Prior authorization in Medicare Advantage. Health Affairs, 45(7), 756-763. 10.1377/hlthaff.2025.01373
- Mello, M. M., et al. (2026). The AI arms race in health insurance utilization review. Health Affairs, 45(1), 6-13. 10.1377/hlthaff.2025.00897
- Murphy, J., et al. (2026). Adverse effects of health plan prior authorization on clinical effectiveness and patient outcomes. American Journal of Medicine, 139(1), 24-32.e1. 10.1016/j.amjmed.2025.08.018
- Shin, J. Y., et al. (2024). Insurance denials and patient treatment in a large academic radiation oncology center. JAMA Network Open, 7(6), e2416359. 10.1001/jamanetworkopen.2024.16359
- HHS Office of Inspector General. (2022; 2026a; 2026b). Medicare Advantage prior authorization denial reviews. oig.hhs.gov
Also cited: CMS-4201-F, CMS-0057-F, KFF (2026, January 28), Bingham et al. (2022), Chino et al. (2023), Kyle & Keating (2024), Osmundson et al. (2025), Rooke-Ley et al. (2025), and Schwartz et al. (2021).
About the author
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, is a healthcare executive, researcher, and author whose work focuses on radiology leadership, clinical capacity, workforce strategy, operational transformation, and population health.
Interactive dashboards, assessments, and executive research briefs are published free to healthcare leaders across three platforms.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Use Turn the page, the arrow keys, or the page selector to read. Twenty-eight pages, seven figures, every derived value labelled. Now showing: two-page spread.
Healthcare leadership review · cost and value analysis, August 2026
Healthcare Is About to Cost More Than $19,000 Per Employee. What Are Employers Actually Buying?
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R · a demand-side warning for health system leaders
The $19,000 question
Healthcare is about to cost more than $19,000 per employee.
What are employers actually buying?
A demand-side warning for health system leaders.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
About this brief
Why the next cost reckoning will be led by employers
A new 9.5% cost projection is not merely another difficult benefits cycle. It is a warning that the purchasers financing American healthcare are reaching the limits of what they are willing and able to absorb.
American employers have spent years treating healthcare inflation as an unavoidable annual expense. Premiums increased, deductibles rose, benefit plans were redesigned, and employees assumed more of the financial burden.
That strategy is approaching its practical limit.
Evidence base: Aon’s 2027 projection, PwC’s independent medical cost trend, a Business Group on Health viewpoint, and three peer-reviewed studies. Full sources appear on page 21.
Every figure derived rather than published is labelled as author-derived at the point of use.
Contents
What is inside
01
The $19,000 question
A demand-side warning for health system leaders
Aon projects that employer-sponsored healthcare costs will rise 9.5% in 2027, pushing average total plan costs above $19,000 per employee.
The forecast is a fourth consecutive year of cost growth approaching double digits. It draws on more than 1,100 U.S. employers representing 7.9 million employees and about $135 billion in 2026 healthcare spending, and it is a baseline projection before plan changes or care-management interventions.
Figure 1
The 2027 arithmetic
Applying the projected 9.5% trend to the reported 2026 total plan cost of $17,562 adds about $1,668 per employee in a single year.
Two forecasts, one conclusion
PwC independently projects a 9% commercial medical cost trend for 2027, the highest in 17 years, informed by health plans covering more than 103 million people in employer-sponsored plans.
The agreement between these forecasts matters. This is not an isolated actuarial estimate. It is evidence of a structural trend.
“
This is no longer simply a benefits problem. It is a capital-allocation, workforce, and healthcare legitimacy problem.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
The arithmetic should command the attention of every executive team.
The cost problem is changing
For years, healthcare inflation was described predominantly as a price problem. Aon now estimates that utilization and the mix of services account for approximately 60% of healthcare cost growth.
Patients are using more services, chronic conditions are becoming more prevalent, high-cost claims are increasing, and treatment is shifting toward more expensive pharmaceuticals and higher-intensity settings.
Price nevertheless remains critical. Hospital consolidation strengthens provider negotiating leverage, and site-of-care differences remain substantial.
Figure 2
Same procedure, different building
A 2025 Health Affairs analysis found commercial insurers paid an average of $1,489 more for 13 common procedures in hospital outpatient departments than in ambulatory surgery centers, a 78% difference.
02
Cost shifting is not cost control
Redistribution dressed as savings
The traditional employer response is predictable: increase payroll contributions, raise deductibles, narrow coverage, or move employees into leaner plans.
Those measures may reduce an employer’s immediate exposure, but they do not lower the underlying cost of care. They redistribute it.
Figure 3
Who actually pays
Employers still absorb about 82% of plan costs, but employees experience the pressure through direct payments and through compensation displaced by rising benefits expense.
The burden lands on families
Research in JAMA Network Open estimated that growth in employer-sponsored premiums between 1988 and 2019 was associated with approximately $125,340 in cumulative lost earnings for the median family, in 2019 dollars. The burden was proportionally greater among Black and Hispanic families and workers with lower earnings.
A national study of privately insured families found that healthcare expenses consumed 26.4% of postsubsistence income among lower-income families by 2019, roughly four times the burden on higher-income families.
Patients do not reliably distinguish low-value from high-value services when facing a large deductible. They defer preventive care, diagnostics, and chronic disease management alongside the services that offer little benefit.
“
Cost shifting redistributes spending. It does not lower the underlying cost of care.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Raising a deductible can create apparent short-term savings while disease progresses until treatment becomes more complicated and more expensive.
Expensive innovation is not necessarily low-value care
Specialty medications and GLP-1 therapies are frequently named as major cost drivers. The concern is understandable. Simply eliminating coverage, however, would confuse price with value.
A medication that reduces cardiovascular events, prevents kidney disease progression, improves diabetes control, or lowers future hospitalization may create substantial clinical and economic value.
The question is not whether a therapy is expensive. It is whether the right patients receive it and whether the spending produces measurable health.
Clinical innovation deserves access when it produces meaningful benefit. It does not deserve exemption from evidence, transparency, or performance accountability.
Artificial intelligence introduces another question
AI can strengthen documentation, increase coding specificity, identify previously unbilled services, and maximize reimbursement. Aon and PwC both identify AI-enabled documentation and revenue optimization as possible contributors to higher billed charges.
Improved documentation is not inherently inappropriate. The concern arises when the financial benefits of AI are realized faster than its clinical and operational benefits.
Read as inflation
- More revenue from the same encounters
- Higher coding yield as the headline metric
- Charges rise, capacity does not
Read as value
- Complications prevented
- Capacity released, delay reduced
- Care moved to lower-cost settings
The distinction is not technological. It is a governance decision.
03
Employers are becoming active healthcare purchasers
The demand side stops being passive
Organizations facing repeated near-double-digit increases will scrutinize where their healthcare dollars go and what those expenditures produce.
The Business Group on Health reports that employers experiencing lower cost trends are more likely to use value-based strategies, and recommends moving successful models beyond pilot status, directing employees toward high-quality providers, and holding partners accountable for clinical and financial outcomes.
What to expect
Eight moves employers will make
- 1High-performance and narrower networks based on measurable cost and quality
- 2Direct contracting with health systems and physician organizations
- 3Centers of excellence for complex and high-cost procedures
- 4Advanced primary care and chronic disease management
- 5Site-of-care optimization for imaging, infusion, surgery, and diagnostics
- 6Transparent pharmacy contracts based on net cost rather than rebates
- 7Outcome guarantees and financial accountability for vendors
- 8Tougher review of hospital pricing, coding intensity, and avoidable utilization
Commercially insured patients have historically been the most financially valuable population. Employers that become selective about networks, sites, and provider performance can redirect considerable volume.
Healthcare organizations have double exposure
Hospitals and health systems are both providers of care and large employers purchasing coverage for their own workforces. Every increase in employer-sponsored healthcare cost therefore reaches them twice.
It raises the cost of employing nurses, technologists, physicians, therapists, and support staff while increasing pressure from commercial purchasers to constrain reimbursement.
A health system that cannot improve healthcare value for its own workforce will struggle to convince outside employers that it can do so for theirs.
Examine total-cost trends, chronic disease prevalence, pharmacy spending, avoidable emergency utilization, care-site variation, diagnostic access, employee navigation, and vendor performance as one integrated operating model.
Executive leadership implications
Four moves for health systems
Make price defensible
Connect commercial price differences to measurable access, quality, safety, experience, and episode outcomes.
Redesign the care site
Create reliable pathways that move imaging, infusion, surgery, and diagnostics to the most appropriate setting.
Govern AI for value
Measure whether AI lowers total cost, prevents delay, releases capacity, and removes administrative work, not only whether it improves coding yield.
Build purchaser evidence
Give employers transparent reporting on utilization, outcomes, avoidable cost, network performance, and employee access.
The objective is not indiscriminate spending reduction. It is to separate spending that creates health from spending generated by fragmentation, excessive prices, preventable disease progression, inappropriate settings, administrative complexity, and weak accountability.
“
The $19,000 figure should be treated as a deadline.
A warning for executive leadership
The next stage of healthcare transformation may not be led by government, insurers, or providers. It may be led by employers who finally decide that spending more is not the same as purchasing better health.
References
Sources and further reading
- Aon. (2026, August 20). U.S. employer health care costs continue multi-year climb, projected to rise 9.5% in 2027. aon.mediaroom.com
- Business Group on Health. (2026). Taking action on health care costs: A Business Group on Health viewpoint. businessgrouphealth.org
- Hager, K., Emanuel, E., & Mozaffarian, D. (2024). Employer-sponsored health insurance premium cost growth and its association with earnings inequality among U.S. families. JAMA Network Open, 7(1), e2351644. doi.org/10.1001/jamanetworkopen.2023.51644
- Maughan, M. P., Ryan, A. M., Radhakrishnan, N., & Whaley, C. M. (2025). Commercial insurers paid more for procedures at hospital outpatient departments than at ambulatory surgical centers. Health Affairs.doi.org/10.1377/hlthaff.2025.00297
- PwC. (2026, June 11). Behind the numbers 2027: Medical cost trend is expected to hit 9%, highest in 17 years. pwc.com
- Shashikumar, S. A., Zheng, Z., Joynt Maddox, K. E., & Wadhera, R. K. (2024). Financial burden of health care in the privately insured U.S. population. JAMA Internal Medicine, 184(7), 843-845. doi.org/10.1001/jamainternmed.2024.1464
Derived figures shown in this brief: the $1,668 increase and $19,230 total (from $17,562 and 9.5%), the 16.7 million at 10,000 employees, the 17.8% payroll share, the $1,966 implied 2025 out-of-pocket, and the two site-of-care payment levels implied by the published gap and percentage.
About the author
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, is a healthcare executive, researcher, and author whose work focuses on radiology leadership, clinical capacity, workforce strategy, operational transformation, and population health.
Interactive dashboards, assessments, and executive research briefs are published free to healthcare leaders across three platforms.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Use Turn the page, the arrow keys, or the page selector to read. Twenty-two pages, six sources, every derived figure labelled. Now showing: two-page spread.
Executive leadership research · 2026
Leadership Rounding Works. Why Do Employees Still Feel Unheard?
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R · the science of executive presence, employee voice, and the feedback loop that converts listening into engagement
Executive leadership research
Leadership rounding works.
Why do employees still feel unheard?
The science of executive presence, employee voice, and the feedback loop that converts listening into engagement.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
About this brief
A leadership warning for health system executives
This brief examines a specific inconsistency in the leadership rounding evidence. The quantitative literature associates rounding with better safety culture, stronger engagement, and lower burnout. The qualitative literature records employees who experience frequent executive visits and still conclude that leaders are not listening.
Both findings can hold at once, because they measure different things. One measures leader presence. The other measures organizational responsiveness. The distance between them is where rounding programs succeed or quietly fail.
Evidence base: seven peer-reviewed sources published between 2006 and 2025, spanning cross-sectional survey research, a quasi-experimental trial, a systematic review, and qualitative studies of mature WalkRounds programs. Full citations appear on page 21.
Association does not establish causation. Where a figure in this brief is derived rather than published, it is labelled as author-derived at the point of use.
Contents
What is inside
01
The evidence
A leadership warning for health system executives
The science increasingly supports leadership rounding. Yet the same evidence shows why employees can experience frequent executive visits and still conclude that leaders are not truly listening.
That is not a trivial inconsistency. It exposes the difference between leader presence and organizational responsiveness. Rounding creates access. Engagement rises only when employees can speak candidly and see that what they disclose influences decisions, resources, or work design.
The quantitative case is compelling
A 2021 study of 10,627 clinical and nonclinical healthcare workers across 396 work settings compared units with high versus low exposure to Positive Leadership WalkRounds.
In high-exposure settings, staff reported better patient safety norms, teamwork, leadership accessibility, quality improvement readiness, and work-life balance.
Source: Sexton et al. (2021), Joint Commission Journal on Quality and Patient Safety. Cross-sectional design; association does not establish causation.
Figure 1
Outcomes by rounding exposure
Figure 1. Percent reporting favorable outcomes by WalkRounds exposure. Emotional exhaustion is inverse: lower is better. Source: Sexton et al. (2021).
Why a short conversation can change the work
Healthcare is status-intensive. Nurses, technologists, medical assistants, support staff, and early-career clinicians may see operational risks that senior leaders cannot, yet remain silent when the interpersonal cost of speaking up seems high.
In 23 neonatal intensive care units, Nembhard and Edmondson found that leader inclusiveness predicted psychological safety, and that psychological safety mediated engagement in quality improvement work.
Effective rounds temporarily compress hierarchy. The leader explicitly invites contribution, signals appreciation rather than punishment, and turns local knowledge into legitimate organizational evidence.
“
The executive act is not walking the unit. It is making the truth safe to tell and consequential after it is told.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Presence is the precondition. Consequence is the intervention.
The mechanism is psychological safety, not visibility
Rounding works through a social and operational chain. Visible leader presence makes access possible. An inclusive invitation reduces the perceived risk of speaking. Psychological safety releases employee voice. What the organization does next decides the rest.
The feedback loop is the active ingredient
In a 2018 analysis of 16,797 survey responses from 829 work settings, 32.7% of respondents reported participating in Leadership WalkRounds, but only 24.3% reported that the rounds were accompanied by feedback on actions taken.
Work settings with more rounds and feedback had safety-culture scores approximately 15% to 27% higher across domains, along with stronger engagement and lower burnout.
Source: Sexton et al. (2018), BMJ Quality & Safety.
The escalation and resolution gap
A 2025 mixed-methods evaluation of Staff Wellness Rounding across six acute hospitals and 14 community health centers found that informal conversations fostered trust and active listening supported well-being. The gap between escalation and perceived resolution was nonetheless striking.
Staff who felt able to escalate a concern against staff who said the concern was addressed. The 45.0-point difference is arithmetic on the two published rates, not a published figure. Source: Iqbal et al. (2025).
Raising a concern and resolving it are separate organizational capabilities. Only the first is a function of the round.
Positive inquiry matters, but it is not positivity theater
Asking what is working well identifies adaptive capacity, exemplary teams, and practices worth spreading. It also restores meaning and recognition.
The strongest rounds balance appreciative inquiry with disciplined questions about hazards, burden, unfairness, workarounds, and resources.
A leader who asks only for good news creates reassurance, not intelligence.
Appreciative half
- What is helping this team provide excellent care?
- What deserves recognition?
- What practice should spread?
Disciplined half
- Where are people working around a broken process?
- What concern is hardest to raise?
- What creates an avoidable burden?
02
Why leadership rounds fail
Observation without protection
Rounding fails when employees experience it as observation without protection, listening without response, or executive theater without operational consequence.
Qualitative research is especially instructive. Studies of mature WalkRounds programs found that seemingly small design choices can make staff interpret rounds as monitoring or evaluation.
Some senior leaders entered the conversation believing they already understood frontline risks, spoke more than they listened, or redirected concerns into explanations. Increased visibility under those conditions can worsen, rather than strengthen, safety perceptions.
Six conditions
What breaks a round
- 1The purpose is vague or changes from visit to visit.
- 2Managers or peers are present when confidentiality is needed.
- 3Leaders correct, defend, minimize, or problem-solve before understanding.
- 4The same issue is repeatedly raised without an owner, deadline, or explanation.
- 5Only easy quick wins are closed while systemic workload and workflow defects remain untouched.
- 6Rounding is used as a substitute for staffing, process redesign, or accountable local leadership.
Each of these is a design choice, not an accident of personality. Every one of them is correctable before the next round.
“
A round that collects concerns but does not visibly process them teaches employees that speaking up produces exposure without influence.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Silence after a round is not neutral. It is instruction.
What the science can and cannot prove
The direction of the evidence is favorable, but its certainty is uneven. A systematic review of nurse-manager intentional rounding found only seven eligible studies and judged their methodological quality generally low.
Much of the WalkRounds literature is cross-sectional, so reverse causation stays plausible: stronger units may both round more effectively and report better cultures.
A 2023 quasi-experimental study across 10 primary care clinics is stronger. Private leader-employee check-ins were associated with a large drop in emotional exhaustion after one year (d = -0.71), but the difference was no longer statistically significant after two years.
Rounding can
- Reveal weak signals early
- Restore human connection
- Improve the capacity to learn
Rounding cannot
- Offset chronically unsafe staffing
- Repair punitive supervision
- Substitute for redistributing resources
03
The evidence-based rounding operating model
Steps one to four
- 1Define the purpose. Specify whether the round is designed to surface safety risks, workforce strain, workflow barriers, recognition opportunities, or all four. Keep the scope stable.
- 2Protect candor. Use small groups or private conversations when needed. Separate rounds from performance evaluation and explain confidentiality limits before asking sensitive questions.
- 3Ask balanced questions. What is helping this team provide excellent care? Where are people working around a broken process? What concern is hardest to raise? What creates an avoidable burden? What deserves recognition?
- 4Listen before solving. Ask follow-up questions, test your understanding, and resist the urge to explain why the problem exists. Aim for employees to do most of the talking.
The operating model
Steps five to eight
- 5Triage transparently. Classify each issue as a local quick win, a cross-functional dependency, an enterprise policy or capital decision, or a non-actionable request requiring an explanation.
- 6Assign accountability. Record an owner, next action, and response date before the issue disappears into a spreadsheet or inbox.
- 7Close the loop rapidly. Respond within 7 to 14 days, even if the final solution will take longer. Silence should never be the default status update.
- 8Return and measure. Revisit prior concerns in the next round and review patterns monthly. The unit of success is not the number of rounds completed. It is trust earned and barriers removed.
Steps 5 through 8 are the half most programs omit. They are also the half the evidence identifies as the active ingredient.
Measure the system, not the executive’s activity
| Domain | Recommended measures |
|---|---|
| Process | Coverage, concerns logged, owner assigned, median days to response, closure rate |
| Experience | Psychological safety, leadership accessibility, perceived responsiveness, engagement |
| Workforce | Emotional exhaustion, intent to leave, absence, regrettable turnover |
| Care | Safety climate, near-miss reporting, operational delays, patient experience |
The most dangerous rounding metric is the number of rounds completed. It measures leader activity while revealing almost nothing about the employee experience or the organization’s response.
“
Employees do not judge leadership rounding by the warmth of the conversation. They judge it by what becomes safer, easier, fairer, or more effective afterward.
A warning for executive leadership
The lasting engagement signal is not that an executive visited. It is that the organization learned and acted.
References
Selected peer-reviewed evidence
- Bayram, A., et al. (2023). Nurse manager intentional rounding and outcomes: Findings of a systematic review. Journal of Advanced Nursing, 79(3), 896-909. https://doi.org/10.1111/jan.15307
- Hurtado, D. A., et al. (2023). Promises and perils of leader-employee check-ins in reducing emotional exhaustion. Mayo Clinic Proceedings, 98(6), 856-867. https://doi.org/10.1016/j.mayocp.2022.12.012
- Iqbal, U., et al. (2025). Using conversations, listening, and leadership to support staff wellness: The CALM Framework. International Journal of Environmental Research and Public Health, 22(10), 1558. https://doi.org/10.3390/ijerph22101558
- Nembhard, I. M., & Edmondson, A. C. (2006). Making it safe: Leader inclusiveness, professional status, and psychological safety. Journal of Organizational Behavior, 27(7), 941-966. https://doi.org/10.1002/job.413
- Rotteau, L., Shojania, K. G., & Webster, F. (2014). I think we should just listen and get out: Patient Safety Walkrounds. BMJ Quality & Safety, 23(10), 823-829. https://doi.org/10.1136/bmjqs-2012-001706
- Sexton, J. B., et al. (2018). Providing feedback following Leadership WalkRounds is associated with better culture, engagement, and burnout outcomes. BMJ Quality & Safety, 27(4), 261-270. https://doi.org/10.1136/bmjqs-2016-006399
- Sexton, J. B., et al. (2021). Safety culture and workforce well-being associations with Positive Leadership WalkRounds. Joint Commission Journal on Quality and Patient Safety, 47(7), 403-411. https://doi.org/10.1016/j.jcjq.2021.04.001
About the author
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, is a healthcare executive, researcher, and author whose work focuses on radiology leadership, workforce strategy, operational transformation, and population health.
Interactive dashboards, assessments, and executive research briefs are published free to healthcare leaders across three platforms.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Use Turn the page, the arrow keys, or the page selector to read. Twenty-two pages, seven peer-reviewed sources. Now showing: two-page spread.
Healthcare leadership review · workforce analysis, August 2026
Physician Burnout Is Declining. Why Are More Doctors Considering Early Retirement?
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R · a workforce warning for health system leaders
Workforce analysis
Physician burnout is declining.
Why are more doctors considering early retirement?
A workforce warning for health system leaders.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
About this brief
Two true trends, pointing opposite ways
Physician burnout prevalence has fallen for four consecutive years. Over roughly the same period, the share of overworked physicians who say the work has made them consider retiring early has more than doubled.
Both findings can be true. Read together they expose a measurement problem: burnout prevalence and workforce commitment are related constructs, and health systems have been treating them as one.
Evidence base: American Medical Association burnout prevalence 2022 to 2025, a 2026 Doximity physician poll, and a 2026 national study of nearly 20,000 family physicians. Sources appear on page 21.
The two headline series come from different organizations with different sampling years. Where this brief compares them directly it does so only at the years both report, and every derived figure is labelled as author-derived at the point of use.
Contents
What is inside
01
The contradiction in the data
A workforce warning for health system leaders
At first glance, the latest physician workforce data appear encouraging. Yet a second trend suggests that the apparent improvement may conceal a deeper threat to clinical capacity.
Physician burnout is declining. The proportion of physicians considering early retirement is increasing. The contradiction does not mean one dataset is wrong. It exposes a serious limitation in how health systems measure physician workforce stability.
The burnout numbers are real
The American Medical Association reports that the percentage of physicians experiencing at least one symptom of burnout fell from 53.0% in 2022 to 48.2% in 2023, 43.2% in 2024, and 41.9% in 2025.
After years of worsening distress, a four-year decline suggests that investments in physician well-being, workflow redesign, leadership development, and organizational support may finally be producing measurable progress.
Yet another trend tells a far less reassuring story.
Figure 1
The divergence
Burnout prevalence (AMA, 2022 to 2025) against the share of overworked physicians considering early retirement (Doximity, 2021 to 2026). Different sources, different sampling years. Only 2023 and 2025 are reported by both.
Four numbers that do not agree
In a 2026 Doximity poll, 82% of participating physicians described themselves as overworked. More strikingly, 46% reported that overwork had caused them to consider retiring early. That proportion rose from 21% in 2021 to 30% in 2023, 37% in 2025, and 46% in 2026.
Early-retirement consideration rose 25 points across the polled years, a 119.0% relative increase, while burnout prevalence fell 20.9% (author-derived).
“
A declining burnout rate should be recognized as progress. It should not be interpreted as proof that the physician workforce has stabilized.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Progress on one measure is not stability on the measure that decides capacity.
Both findings can be true
Burnout prevalence, early-retirement consideration, professional fulfillment, engagement, intent to reduce clinical hours, and intention to leave are related constructs. They are not interchangeable.
A physician may no longer meet a formal burnout threshold while remaining chronically overworked, dissatisfied with organizational leadership, frustrated by administrative demands, or uncertain that continuing full-time practice is personally sustainable.
Burnout can improve statistically while workforce commitment continues to deteriorate.
A physician may report less emotional exhaustion than during the height of the pandemic and still conclude that another five or ten years under the current operating model is untenable. From the physician’s perspective that decision may be rational. From the health system’s perspective it is a substantial loss of capacity.
Burnout is not the entire workforce story
Burnout is commonly characterized by emotional exhaustion, depersonalization or cynicism, and diminished professional efficacy. These are important indicators of occupational distress.
They do not capture every reason a physician may choose to reduce clinical effort or leave practice earlier than planned.
What burnout measures
- Emotional exhaustion
- Depersonalization or cynicism
- Diminished professional efficacy
What it misses
- Whether the work is sustainable for a career
- Confidence that leadership will change it
- Financial readiness to leave
The physician who has stopped expecting improvement can score below the burnout threshold and still be the next departure.
Cumulative exposure
Ten conditions behind the exit
- 1Persistent workload exceeding available clinical capacity
- 2Inadequate staffing and unstable care teams
- 3Excessive documentation and inbox responsibilities
- 4Loss of control over scheduling and clinical decisions
- 5Repeated organizational change without meaningful physician participation
- 6Compensation models that prioritize volume over professional sustainability
- 7Increasing call intensity and inadequate recovery time
- 8Declining confidence in executive leadership
- 9Reduced professional meaning
- 10Financial readiness to leave an increasingly difficult work environment
None of these reliably appears in a conventional burnout measure.
02
Actual turnover evidence raises the stakes
The concern is no longer limited to stated intentions
A 2026 national study involving nearly 20,000 family physicians found that physicians reporting burnout were approximately 1.5 times more likely to change practices or stop practicing medicine than physicians who did not.
Among physicians reporting burnout, 4.8% changed practices and 5.4% stopped practicing. Among physicians without burnout, the corresponding rates were 3.4% and 3.7%.
Figure 2
Departure rates by burnout status
Family physicians, 2026 national study. The absolute differences look modest. Across a large workforce they are not: a 3.1-point spread in annual departure risk is 31 additional departures for every 1,000 physicians (author-derived).
Turnover is not merely a human-resources event
- Recruitment, credentialing, and onboarding expenses
- Temporary coverage and locum-tenens costs
- Lost clinical revenue during vacancies
- Longer appointment waits and disrupted patient continuity
- Greater workload for the physicians who remain
- Referral-network instability and delayed service-line growth
- Reduced teaching, mentorship, and leadership capacity
The consequences are self-reinforcing.
“
How many physicians believe their current work environment is sustainable for the remainder of their careers?
The question that replaces the burnout count
The more consequential leadership question is not how many physicians meet the threshold for burnout.
Health systems may be measuring the wrong finish line
A broader framework evaluates burnout alongside professional fulfillment, organizational engagement, perceived leadership support, staffing completeness, workload and capacity alignment, schedule control, administrative burden, after-hours electronic health record work, intent to reduce clinical hours, intent to leave, early-retirement consideration, and actual voluntary turnover.
These measures should be segmented by specialty, department, career stage, clinical effort, call burden, and immediate leadership unit. A systemwide average can conceal severe problems inside individual departments.
Where 2025 burnout remained elevated
The national average does not describe the daily experience of every physician.
03
The response must move beyond wellness programming
Services help; they do not restructure the work
Health systems have invested in resilience education, peer support, mindfulness resources, counseling, and well-being applications. These can be valuable, especially when physicians need confidential support.
They cannot compensate for a structurally unsustainable operating model.
A physician cannot meditate away an understaffed clinical team. Resilience training will not correct an excessive patient panel, an unstable call schedule, a poorly designed electronic workflow, or the absence of influence over clinical decisions.
Strategies that address the work itself
One to five
- 1Develop accountable local physician leaders.
- 2Align workload with clinical capacity.
- 3Fully staff interdisciplinary care teams.
- 4Eliminate or redesign low-value administrative work.
- 5Give physicians meaningful voice and decision authority.
Each of these changes what the job is, not how the physician is expected to feel about it.
Strategies that address the work itself
Six to ten
- 6Increase schedule control and career flexibility.
- 7Protect professional meaning and development.
- 8Strengthen collegiality and psychological safety.
- 9Improve fairness, recognition, and transparency.
- 10Measure physician experience and visibly act on the findings.
These are not peripheral wellness initiatives. They are operating-model requirements.
“
The most dangerous retention risk may not be the visibly distressed physician. It may be the experienced, high-performing physician who has stopped believing that the organization will make the work sustainable.
A warning for executive leadership
That physician may not complete another wellness module, file a complaint, or request an intervention. He or she may simply leave.
References
Sources and further reading
- American Medical Association. (2026, April 16). Physician burnout rates are falling, specialty gaps remain. ama-assn.org
- Doximity. (2026, August 3). Health Systems Are Finally Redesigning Work to Fight Physician Burnout. opmed.doximity.com
- Weill Cornell Medicine. (2026, March 30). Burnout May Lead Family Doctors to Leave Medicine. news.weill.cornell.edu
A note on the derived figures
The report states that burned-out family physicians were approximately 1.5 times more likely to leave. Recomputed from the printed rates, the ratio is 1.41x for changing practices and 1.46x for stopping practice, or 1.44x for either outcome combined. The published approximation is reasonable; the exact figures are shown so the arithmetic is visible rather than assumed.
The two headline series are reported by different organizations. Figure 1 compares them only at 2023 and 2025, the years both report.
About the author
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R, is a healthcare executive, researcher, and author whose work focuses on radiology leadership, clinical capacity, workforce strategy, operational transformation, and population health.
Interactive dashboards, assessments, and executive research briefs are published free to healthcare leaders across three platforms.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Use Turn the page, the arrow keys, or the page selector to read. Twenty-two pages, three primary sources, every derived figure labelled. Now showing: two-page spread.
Executive research report · Perspective
Hospitals Say They Have a Workforce Shortage. Why Are They Laying People Off?
The hidden difference between labor scarcity, organizational restructuring, and the destruction of care capacity. A rapid evidence synthesis by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R.
- Hospital layoffs
- Care capacity
- Support roles
- Hospital finance
- Burnout
- Organizational redesign
Executive research report · Perspective
Hospitals Say They Have a Workforce Shortage. Why Are They Laying People Off?
The hidden difference between labor scarcity, organizational restructuring, and the destruction of care capacity
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
A hospital can eliminate a position. It cannot eliminate the work by removing the person.
Perspective article · Health policy and management
Payroll is not the operating system
A rapid evidence synthesis on labor scarcity, hospital restructuring, and the hidden economics of care capacity.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Keywords
- hospital layoffs
- workforce shortage
- care capacity
- staffing
- support roles
- patient safety
- burnout
- hospital finance
- organizational redesign
Abstract
The answer begins with a category error
Hospitals publicly describe a workforce shortage while simultaneously eliminating positions. If labor is scarce, why would an institution dismiss people it may soon struggle to replace?
Hospitals do not operate in one undifferentiated labor market. They can lack bedside nurses, respiratory therapists, imaging professionals, laboratory personnel, pharmacists, and support staff in specific locations or shifts while carrying duplicated management, shrinking health-plan operations, outsourced transaction work, or positions tied to services and funding that no longer exist.
Yet this explanation is not a defense of every layoff. Hospitals are tightly coupled production systems. A scheduler, transporter, environmental services worker, laboratory technologist, care coordinator, coder, or nurse manager may be labeled nonclinical while still protecting clinician time, patient throughput, diagnostic reliability, and revenue integrity.
This report argues that layoffs should be judged by their effect on care capacity rather than by job title, and introduces a Workforce Capacity Impact Assessment, a shadow profit-and-loss ledger, and a governance scorecard designed to distinguish necessary restructuring from organizational self-harm.
Central thesis
Hospitals can face genuine labor scarcity and still need to restructure. The decisive question is whether eliminated work has truly ended.
If demand remains and tasks shift to an already constrained team, the layoff has not reduced costs. It has moved cost into the operating system.
Researcher summary
What the evidence supports
- The answer begins with a category error.
- A hospital workforce shortage is not a shortage of every kind of labor. Scarcity varies by occupation, competency, location, shift, compensation, and patient demand. An organization can be desperately short of bedside capacity while eliminating other roles.
- Shortage and layoff data can both be true.
- Forty-five hospitals and health systems announced reductions from January through July 2026, involving about 5,800 disclosed positions. National hospital employment nevertheless stood near 5.77 million and rose by roughly 12,300 between April and July. Local contraction can coexist with national growth.
- Financial pressure is credible, but sector averages do not decide local necessity.
- Hospital expense growth outpaced price growth in 2025. Kaufman Hall reported a 1.7% adjusted year-to-date operating margin through March 2026. MedPAC documented deep aggregate Medicare losses alongside positive all-payer results. These figures establish strain and variation, not the wisdom of a specific layoff.
Researcher summary
What follows from it
- The label nonclinical is an unreliable proxy for dispensability.
- Recent evidence links incomplete team staffing and work overload with burnout, reduced clinical hours, and intent to leave. Support staff expenses can be modest relative to provider expenses while materially affecting provider productivity. Removing support often transfers work rather than eliminating it.
- Every workforce reduction needs a shadow profit-and-loss statement.
- Direct salary savings should be tested against overtime, agency use, vacancy duration, severance, unemployment costs, outsourcing, lost throughput, delayed discharge, safety events, revenue leakage, and the cost of rebuilding capacity.
- Redesign should precede reduction.
- Eliminate unnecessary work, simplify management, standardize fragmented processes, redeploy talent, retrain staff, reduce premium contingent labor, and use natural attrition where feasible. Involuntary layoffs should follow only when demand, work, and capacity have been explicitly redesigned.
Editorial illustration
One operating system, three vantage points
Original illustration prepared for this report.
Section one
The contradiction is real, but the category is wrong
The headline is irresistible because it appears to expose institutional hypocrisy. Hospitals say they cannot find enough workers, then show workers the door. In a moral narrative, the contradiction proves that the shortage was exaggerated or that the layoffs are reckless. In a financial narrative, the reductions are evidence of discipline.
Both stories are too simple.
There is no single hospital workforce market. The labor pool for a night-shift intensive care nurse in rural Kansas is not the market for a corporate communications manager in Boston. The availability of a radiologic technologist trained in magnetic resonance imaging is not interchangeable with that of a revenue-cycle analyst, whose work may be consolidated across a system.
Nor does the existence of applicants prove usable capacity. Licensure, competency, geography, schedule, compensation, immigration rules, training capacity, and patient acuity shape whether labor can be converted into safe care.
Section one
A layoff count is an incomplete measure
The same distinction applies to headcount. A hospital may reduce funded positions in one area while hiring in another. It may eliminate vacant positions that are financially real but do not represent a person losing a job.
It may close a service because demand has declined, outsource a function, merge duplicated departments, or leave a payer market. Each decision changes the organization differently.
A layoff count without the affected work, patient demand, and replacement plan is an incomplete measure.
That is why the same announcement can be, at once, honest about scarcity and honest about reduction. The two statements describe different parts of the same institution, and neither one settles what the change will do to care capacity.
Framework
Six conditions lumped together as shortage
| Condition | Illustrative setting | Operating consequence |
|---|---|---|
| Occupational shortage Too few qualified workers for a role | Respiratory therapists, bedside RNs, pharmacists | Vacancy, wage pressure, closed capacity |
| Geographic shortage Workers exist nationally but not locally | Rural and underserved markets | Travel labor, recruitment delay, access loss |
| Schedule shortage Coverage fails at particular times | Nights, weekends, holidays, call | Overtime, fatigue, service bottlenecks |
| Team capacity shortage Clinicians lack supporting roles | Scheduling, transport, environmental services, care coordination | Work transfer, lower productivity, burnout |
| Budgeted headcount shortage Needed roles are not funded | Hiring freezes or vacancy controls | Demand remains, capacity is administratively constrained |
| Design failure Work is fragmented or duplicated | Merger overlap, excessive layers, manual processes | High cost despite vacancies and staff overload |
Table 1. Author-developed taxonomy. Six different conditions are commonly collapsed into the single phrase workforce shortage.
Section two
What the current layoff wave actually shows
Reported reductions were less frequent and less deep in the first seven months of 2026 than during the same period in 2025. Forty-five employers announced workforce reductions from January through July 2026, compared with 59 in the comparable period in 2025.
The average disclosed reduction also fell, from about 235 positions to about 140. These counts are useful signals, not a national census. Some systems did not disclose headcount, the tracker can miss events, and the 2026 list remains open.
Source: Becker’s Hospital Review tracking.
Figure 1
Fewer employers, fewer disclosed positions
The data are a media tracker, not a comprehensive census. Source: Becker’s Hospital Review.
Section two
Illustrative 2026 workforce actions
| Organization | Reported action | Stated scope or driver |
|---|---|---|
| UPMC June 2026 | About 200 layoffs plus 300 vacant positions | Primarily nonclinical or non-member-facing roles |
| University of Vermont Health June 2026 | 76 eliminated; 66 restructured | More than $300 million projected three-year gap; some clinical roles affected |
| Presbyterian Healthcare Services June 2026 | About 150 | Health-plan and administrative roles linked to exiting most Medicare Advantage plans in 2027 |
| Baptist Health, Fort Smith June 2026 | About 70 | Sustained volume decline and duplicated roles |
| Centra Health June 2026 | About 90 | Roughly 1% of the workforce, mainly administrative and management |
| PeaceHealth, Washington February 2026 | 94 | Included nurses and other clinical staff across three medical centers |
| Baystate Health February 2026 | 117 | Corporate positions, less than 1% of the system workforce |
| Alameda Health System 2026 reversal | Plan reduced, then the remaining 92 rescinded | A $19.3 million county allocation preserved positions and extended behavioral health services |
Table 2. Illustrative, not exhaustive. Announcements can change after publication.
Section two
The composition changed, not just the count
Closures drove some of the largest 2025 events, including full-facility workforces. In 2026, announcements more often concentrated on management, administration, information technology, revenue cycle, health-plan operations, and shared services.
That pattern is consistent with systems attempting to preserve bedside roles.
It does not prove that the reductions are clinically neutral.
Whether a reduction is clinically neutral is an empirical question about work, not a claim that can be settled by the department name on an organizational chart. The Alameda reversal is instructive for the opposite reason: an external allocation changed the arithmetic, and the plan changed with it.
Evidence boundary
A current layoff announcement states that the employer will reduce positions and explains its rationale. It does not establish the causal effect on safety, access, workload, or long-term cost.
Those outcomes require prospective measurement and later evaluation.
Section three
The national labor market did not contract
Federal data reinforce the difference between local restructuring and national employment. Seasonally adjusted hospital employment rose from approximately 5.758 million in April 2026 to 5.770 million in July, although June and July were preliminary.
The July gain was slower than the average monthly increase of about 36,000 over the preceding year, but it was growth. An individual system can contract while the sector expands.
Source: U.S. Bureau of Labor Statistics (2026).
Figure 2
Hospital employment rose while systems announced layoffs
June and July values are preliminary. Source: U.S. Bureau of Labor Statistics (2026).
Figure 3
Hospital-dependent occupations are projected to grow
BLS projects 189,100 registered-nurse openings annually from 2024 to 2034. HRSA projects a national shortage of roughly 109,000 registered-nurse full-time equivalents by 2038, with a much larger percentage gap in nonmetropolitan areas.
Section four
Why cut jobs when the people are still needed?
Hospital finance is a contest between recurring obligations and uneven revenue.
Labor is usually the largest controllable expense. Wages recur every pay period, and a reduction produces an immediate accounting effect. Revenue, in contrast, comes from a complex mix of government payments, commercial contracts, patient responsibility, supplemental funding, and philanthropy.
Claims can be delayed or denied. Payer mix can deteriorate. A service may be clinically necessary but financially negative. Drug and supply expenses can rise faster than payments.
Payroll can be reduced faster than care can be redesigned. That asymmetry, not malice, explains most of the pattern.
Figure 4
Cost growth outpaced price growth in 2025
The values are industry advocacy estimates and should not be treated as proof of a particular hospital’s affordability. Source: American Hospital Association (2026).
Framework
One driver, two very different responses
| Driver | Potentially rational response | Strategic hazard |
|---|---|---|
| Demand loss Volume or service mix no longer supports prior staffing | Reduce the work and the capacity together | Cutting staff while demand remains |
| Merger duplication Two systems retain parallel functions | Standardize, consolidate, and clarify accountability | Removing local coordination without redesigning the enterprise workflow |
| Payer or product exit Health-plan operations shrink | Align staffing with the discontinued product | Transferring unresolved member or claim work to care teams |
| Funding expiration Grant, subsidy, or allocation ends | Close, redesign, or replace the funding transparently | Pretending mandated or essential work has ended |
| Technology or outsourcing A vendor replaces part of a workflow | Validate capability, quality, control, and total cost | Counting theoretical automation before it reliably performs |
| Liquidity crisis Cash or covenant pressure | Use staged, reversible steps and board oversight | Across-the-board cuts unrelated to demand or interdependence |
Table 3. The same financial driver can produce disciplined redesign or indiscriminate capacity loss.
Section four
Entity-level evidence, not sector averages
A national expense chart cannot tell a board which role to eliminate. A consolidated health-system margin may obscure the condition of a safety-net subsidiary.
A hospital can be cash-constrained despite reported net income, because investment gains, restricted funds, debt terms, and capital obligations are not interchangeable with operating cash.
Conversely, leaders cannot invoke sector hardship as a substitute for disclosing the local assumptions behind a reduction.
Both directions of the error are common. Employees point to system revenue as though it were unrestricted cash. Executives point to national benchmarks as though they settled a local decision. Neither is evidence about this hospital, this service line, or this position.
Section five
Nonclinical does not mean nonessential
The distinction between clinical and nonclinical work is administratively convenient and operationally crude. Care is produced by a chain of dependent tasks. A physician can diagnose, but the diagnosis does not become care unless someone schedules, authorizes, transports, cleans, stocks, documents, codes, coordinates, and follows through.
When those tasks lose an owner, they either remain undone or are picked up by the most conscientious person nearby. That person is often a nurse, physician, therapist, or frontline supervisor.
The smaller share does not imply smaller importance. This is the economic logic of complementarity: a lower-cost role can increase the usable output of a higher-cost role. Source: Kaufman Hall (2026).
Framework
What each role is actually holding up
| Role | Work transferred if removed | Likely consequence |
|---|---|---|
| Scheduling and registration | Calls, messages, registration correction, idle capacity | Clinician time, no-shows, delayed care, denials |
| Environmental services | Cleaning delay or task shifting | Closed beds, boarding, infection risk |
| Transport | Nurses or departments coordinate movement | Delayed tests, longer stay, missed throughput |
| Laboratory and imaging support | Rework, interruptions, queue growth | Slower diagnosis and treatment |
| Care coordination and social work | Physicians and nurses chase post-acute arrangements | Avoidable days, readmission risk, moral distress |
| Health information and revenue cycle | Clinical documentation queries and uncompensated work | Revenue leakage, administrative burden |
| Unit leadership and education | Frontline escalation without an owner | Variation, turnover, delayed corrective action |
Table 4. Illustrative role interdependencies. Local workflow observation is required before any position-specific decision.
Section five
Incomplete teams are a retention risk
Rotenstein and colleagues studied 970 physicians across 15 U.S. organizations. Nearly half reported working with an incompletely staffed team more than one-quarter of the time.
Intent to reduce clinical hours was 31.2% against 22.0%. Adjusted odds remained elevated.
Because the study was cross-sectional, it cannot prove that incomplete staffing caused the outcomes. It does show that team completeness is not a peripheral employee-satisfaction issue. Source: Rotenstein et al. (2025).
Figure 5
Working short is not a satisfaction problem
Associations are cross-sectional. Source: Rotenstein et al. (2025).
Section six
The shadow profit-and-loss statement
A restructuring proposal usually begins with a visible ledger: salary, benefits, severance, and implementation cost. That is the direct profit-and-loss statement. It is necessary, but not sufficient.
The governing identity is simple: eliminating headcount does not eliminate work.
Demand may fall, a process may be automated, or duplicated activity may truly disappear. If none of those conditions are present, the remaining workers inherit the task.
The cost can then reappear as longer shifts, missed breaks, sickness absence, vacancy, overtime, agency staffing, queue growth, delayed discharge, canceled procedures, revenue leakage, safety events, or management time spent repairing the change.
Figure 6
How a reduction becomes a loop
This is a conceptual synthesis, not a causal estimate. Original framework developed for this report.
Framework
The shadow ledger, eight domains
| Ledger domain | Costs or effects to count | Measurement rule |
|---|---|---|
| Direct labor | Salary and benefits removed; severance; accrued leave | One-time and recurring values are separated |
| Replacement labor | Overtime, agency, float pool, contractors, outsourcing | Compare 12 months before and after |
| Workforce stability | Vacancy, turnover, sickness absence, time to fill, retirement acceleration | Affected teams and adjacent roles |
| Capacity and access | Closed beds, clinic templates, imaging slots, cancellations, boarding | Volume and wait time normalized for demand |
| Flow and productivity | Length of stay, discharge delay, turnaround time, clinician administrative work | Track work transfer explicitly |
| Quality and safety | Falls, pressure injuries, infections, medication events, complaints | Use control limits and risk adjustment |
| Revenue integrity | Denials, missed charges, authorization delay, documentation rework | Net collections, not gross charges |
| Recovery option | Recruitment, onboarding, training, technology reversal, vendor exit | Estimate time and cost to restore capacity |
Table 5. Every item should have a baseline, an owner, a reporting cadence, and a decision threshold.
Section seven
What the evidence can and cannot establish
There is limited peer-reviewed evidence on the causal consequences of the specific layoffs announced in 2026, because those events are recent, heterogeneous, and incompletely observed. It would be irresponsible to claim that every administrative reduction causes patient harm.
The stronger evidence concerns the pathways through which a poorly designed reduction can become harmful: incomplete teams, workload, sickness absence, turnover, and dependence on overtime or agency labor.
In a national survey of 43,026 healthcare workers, work overload was associated with more than twice the adjusted risk of burnout across physicians, nurses, other clinical staff, and nonclinical staff. The association with intent to leave was also elevated across all groups.
Workload is not confined to the bedside. Nonclinical workforce distress destabilizes the same system clinicians depend on.
Figure 7
Overload travels across every role group
Survey data were collected in 2020 and are cross-sectional. The consistency across roles is what matters here. Source: Rotenstein et al. (2023).
Evidence
What each study can carry
| Study | Decision-relevant finding | Primary limitation |
|---|---|---|
| Rotenstein et al., 2025 970 physicians, 15 organizations | Incomplete teams associated with burnout and intentions to reduce hours or leave | Cross-sectional; team completeness self-defined |
| Rotenstein et al., 2023 43,026 healthcare workers | Work overload associated with burnout and intent to leave across all role groups | 2020 survey; cross-sectional |
| Dall’Ora et al., 2025 2.69 million shifts, 4 English hospitals | Long shifts, lower RN staffing, and lower RN skill mix associated with sickness absence | Observational; UK setting |
| Mauricio et al., 2025 8,584 units, 42,929 falls | Higher RN and APRN turnover associated with more patient falls | Cross-sectional; small effect estimate |
| Pittman et al., 2025 70 U.S. hospitals, 2019 to 2022 | High agency and overtime exposure associated with pressure injuries | Associations found for only selected safety indicators |
| Bae, 2022 Systematic review of 16 studies | Nurse turnover was costly; evidence of negative operational and patient effects was partial | Heterogeneous nonexperimental literature |
Table 6. None of these studies proves that a particular layoff will produce a particular adverse event. Together they make the shadow-cost pathway plausible enough to measure prospectively.
Decision tests
Defensible restructuring or capacity destruction
| Test | Potentially defensible | Warning of strategic failure |
|---|---|---|
| Demand | Volume, service, product, or duplicated work has demonstrably ended | Demand remains, but fewer people absorb it |
| Workflow | Future-state work is mapped, owned, timed, and tested | The plan names positions, not tasks or handoffs |
| Interdependence | Adjacent clinical and operational roles validate the change | Job title is used as a proxy for importance |
| Economics | Full-cost model remains favorable under sensitivity analysis | Only salary and benefit savings are counted |
| Safety and access | Measures, thresholds, and mitigation defined before implementation | Leaders promise no impact without a measurable basis |
| Reversibility | The organization can restore capacity quickly if assumptions fail | Skills, teams, or local knowledge will be hard to rebuild |
| Governance | Board, clinical, workforce, and finance leaders share oversight | The same sponsor designs, approves, and evaluates the plan |
| Distribution | Burden and benefit assessed across patients, sites, shifts, and groups | Cuts fall on low-power teams or underserved communities without review |
Table 7. Decision tests for distinguishing restructuring from capacity destruction.
A practical rule
If leaders cannot identify which work has stopped, who owns each remaining task, what capacity could be lost, and how failure will be detected, they do not yet have a workforce-reduction plan.
They have a payroll target.
Figure 8
The Workforce Capacity Impact Assessment
The assessment applies to layoffs, vacancy suppression, outsourcing, automation, shared-services consolidation, and service closure. Original framework developed for this report.
The assessment
Seven gates before a position is eliminated
- 1Define demand and obligation. Specify current and forecasted volume, acuity, wait time, regulatory requirements, and community obligations. If demand continues, show how equal or better capacity will be delivered.
- 2Map role interdependence. Observe the work, not only the job description. Identify upstream inputs, downstream users, failure modes, informal coordination, and tasks clinicians currently do not see.
- 3Name every work transfer. For each recurring task, identify whether it is stopped, automated, moved to a vendor, or moved to another employee. Unassigned work is not eliminated work.
- 4Model safety and access. Select measures that could deteriorate, define baseline variation, set control limits, and name a mitigation owner. Include closed beds, canceled sessions, wait times, and turnaround times.
- 5Build the full-cost case. Add severance, unemployment, outsourcing, implementation, overtime, agency, recruitment, rework, revenue loss, and recovery cost. Test optimistic and adverse scenarios over at least 24 months.
- 6Protect reversibility. Stage the change, retain critical knowledge, preserve credentialing pipelines, establish vendor exit rights, and define the point at which capacity will be restored.
- 7Separate sponsorship from evaluation. Use independent finance validation, frontline workflow review, clinical safety oversight, workforce input, and board monitoring. The sponsor should not be the sole judge of success.
Sequence
Redesign before reduction
The most durable savings come from removing low-value work and stabilizing the operating system before removing people. A blunt reduction reverses that logic.
| Order | Action | Proof of benefit |
|---|---|---|
| 1 | Stop unnecessary demand. Retire obsolete reports, meetings, approvals, duplicate documentation | Hours removed from the system |
| 2 | Simplify structure. Clarify accountability, reduce redundant layers, standardize processes | Cycle time, spans of control, decisions per layer |
| 3 | Redesign workflow. Move work to the right skill level, standardize handoffs, reduce rework | Task time, queue, first-pass quality |
| 4 | Redeploy and retrain. Move capable employees to shortage roles or adjacent functions | Vacancies filled, training completion, retention |
| 5 | Reduce premium labor. Convert agency and overtime dependence into stable core capacity | Agency share, overtime, total labor cost |
| 6 | Use attrition and voluntary options. Vacancy management, retirement incentives, internal placement | Net headcount and regrettable loss |
| 7 | Implement targeted layoffs. Only after demand, work transfer, safety, full cost, and reversibility are validated | Realized net savings and capacity outcomes |
Table 8. The sequence can be accelerated in a liquidity emergency, but the decision tests should not be abandoned.
Governance
Govern it like a clinical intervention
Clinical interventions begin with indication, contraindication, informed consent, dose, monitoring, and rescue. Workforce interventions deserve comparable discipline, because they alter the environment in which care is produced.
- Financial
- Net recurring savings, implementation cost, agency and overtime, outsourcing, denial and collection effects. Trigger: net savings below threshold for two periods.
- Workforce
- Vacancy, turnover, sickness absence, engagement, time to fill, internal transfers, grievance volume. Trigger: material rises beyond baseline control limits.
- Capacity
- Staffed beds, closed beds, clinic and procedure sessions, wait time, boarding, turnaround time. Trigger: sustained access or throughput deterioration.
- Quality and safety
- Falls, pressure injuries, infections, medication events, complaints, near misses. Trigger: sentinel event or adverse control signal.
- Work transfer
- Clinician administrative time, missed breaks, queue age, rework, supervisor workload. Trigger: unplanned transfer exceeds approved capacity.
- Equity and mission
- Effect by site, service, payer, language, disability, and geography. Trigger: disproportionate access loss without mitigation.
Table 9. Local definitions, baselines, control limits, and accountable owners are required before use.
Implementation
A 90-day architecture
- 1Days 0 to 30: validate the hypothesis. Freeze avoidable expansion, reconcile data definitions, map affected workflows, interview frontline users of the roles, define demand, model the shadow ledger, and identify redeployment options.
- 2Days 31 to 60: test the future state. Pilot automation or consolidation, run a failure-mode review, confirm vendor capability, evaluate workload transfer, finalize safety and access measures, and establish stop rules.
- 3Days 61 to 90: stage and monitor. Implement in waves where possible, protect knowledge transfer, provide dignified communication and placement support, publish the baseline scorecard, and schedule executive and board reviews.
- 4After day 90: decide again. Do not treat completion of notifications as proof of success. Compare actual net savings and capacity outcomes with the approved case at 30, 90, 180, and 365 days. Restore capacity when thresholds fail.
Section eight
What each party should understand
- Employees and labor representatives
- Financial stress should not be dismissed because a system is large or reports substantial revenue. Revenue is not margin, and margin is not unrestricted cash. The strongest challenge to a layoff is not that every job must remain. It is evidence that demand persists, that work transfer is hidden, that a lower-risk alternative exists, or that the full-cost case does not hold.
- Executives and operational leaders
- The phrase non-patient-facing should never be used as a safety conclusion. It is an administrative description. Promising that patients will feel no effect before workflow testing is complete is not reassurance. It is an unfalsifiable claim.
- Boards
- Approving a salary target without reviewing service demand, workload transfer, safety thresholds, and recovery cost is equivalent to approving a capital project without a business case. Ask who independently validated the savings, which roles were redeployed, how clinicians participated, and how quickly capacity can be restored.
Discipline
Questions that should remain open
- Which announced 2026 reductions will produce sustained net savings after replacement labor, outsourcing, revenue, and recovery costs are counted?
- Which nonclinical roles most strongly increase clinician productivity, throughput, patient access, or revenue integrity in different hospital settings?
- How should systems value organizational memory, informal coordination, and surge capability before eliminating experienced teams?
- How do workforce reductions affect rural, safety-net, pediatric, behavioral-health, and academic systems differently?
- Which forms of artificial intelligence truly eliminate work, and which merely move exception handling, verification, and liability to clinicians?
- What evidence would cause leaders to restore a position, or labor representatives to support a reduction?
Caveats: this is a rapid evidence synthesis and perspective, not a formal systematic review. Media trackers are not comprehensive, and organizations can revise, delay, or rescind plans. Financial sources use different populations, periods, and definitions. The peer-reviewed studies summarized here are mostly observational.
Conclusion
The payroll is not the operating system
Hospitals can truthfully report a workforce shortage while laying people off, because the shortage is not a uniform scarcity of all labor. It is a misalignment among demand, skill, place, schedule, payment, and organizational design.
Some reductions are necessary. A discontinued insurance product does not need the same workforce. A merged enterprise does not need every duplicated layer. A service with permanently lower demand should not preserve a structure designed for a different reality.
But the logic becomes dangerous when it moves from specific work to broad labels. Nonclinical does not mean nonessential. A vacant position does not mean the work is absent.
The standard is neither job preservation at any cost nor payroll reduction at any cost. It is capacity stewardship: remove work before removing the people who perform it, account for costs beyond the labor ledger, protect reversibility, and stay accountable after the announcement fades.
Final perspective
A hospital can survive a smaller payroll. It cannot fulfill its mission with less capacity than its patients require. The difference between those two conditions is the work of leadership.
KE
About this report
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
Healthcare executive, author, and registered radiologic technologist. This rapid evidence synthesis draws on 2026 layoff announcements, federal workforce data, current hospital finance benchmarks, and peer-reviewed evidence on incomplete teams, workload, sickness absence, turnover, contingent staffing, and patient safety.
Suggested citation
Emrick, K. (2026). Hospitals say they have a workforce shortage. Why are they laying people off? The hidden difference between labor scarcity, organizational restructuring, and the destruction of care capacity. Executive research report.
- kellyemrick.com
- kellyemrick.org
- kellyemrick.blog
Copyright © 2026 Kelly Emrick. All rights reserved.
Use the arrows, the left and right keys, or the section menu. 44 pages on 22 leaves.