Investigative report dashboard | November 2026
The scan is complete. The care is not.
Radiology results backlogs and leadership accountability, built around the Emrick Diagnostic Completion and Accountability Model.
Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R
- 8 model components
- 4 backlogs
- 6 transitions
- 11 measures
- 90-day plan

The centerpiece
The Emrick Diagnostic Completion and Accountability Model
Eight components that turn a completed scan into completed care. Select a segment of the wheel to open its operating product, accountable role, design choices, measures, and failure conditions. Rate its three controls and the outer band of the wheel records where the organization stands.
- Not rated
- Not in place
- Partly
- In place, with evidence
Select any segment. Keyboard: Tab to a segment, then Enter.
Component 1 of 8
Reconcile unfinished work into one accountable record
- Required operating product
- Auditable episode registry with owners and exception states.
- Accountable leadership role
- Imaging operations leader with informatics lead.
Missing data are a management exception, not evidence that no patient is waiting.
Design choices
- Reconcile scheduling, acquisition, interpretation, communication, and follow-up records at the episode level. Do not assume the worklist is complete because it is operational.
- Set explicit counting rules. Several accessions can belong to one episode, and one report can generate more than one action.
- Merge duplicates with traceability. Require a reason for every cancellation and, where clinically relevant, confirmation of an alternative plan.
- Keep missing timestamps visible as missing. Never replace them with assumed completion times.
Minimum registry fields
- Stable episode identifier
- Patient identifier
- Examination and accession identifiers
- Modality and location
- Acquisition timestamp
- Preliminary and final report timestamps
- Urgency
- Responsible clinician
- Required communication
- Action description
- Action due date
- Current owner
- Current status
- Reason for any hold
Hypothetical example
A team discovers that several examinations disappeared from the routine reading list after a routing change. The fix is to restore the cases, review their urgency, identify the interface defect, and test whether the same defect affected other patients. Adding the discovered examinations to a spreadsheet recovers the visible cases and leaves the mechanism intact.
All examples are hypothetical illustrations.
Success measures
- Source-to-registry reconciliation
- Number of unassigned episodes
- Duplicate rate
- Share of records with valid timestamps
- Daily reconciliation equation: opening cases, additions, valid closures, closing cases
Failure conditions and tradeoffs
The registry becomes unsafe when staff must maintain it through extensive parallel manual entry, or when a reassuring dashboard obscures failed data feeds. Sustain it with a named data steward and sampled comparisons against source records.
Self-check: three controls
The registry reconciles scheduling, acquisition, interpretation, communication, and follow-up records at the episode level.
A daily reconciliation equation is published: opening cases, additions, valid closures, closing cases.
A named data steward samples the registry against source records.
Component 2 of 8
Prioritize clinical risk and prevent indefinite waiting
- Required operating product
- Clinician-approved urgency and escalation rules.
- Accountable leadership role
- Radiology medical director with receiving services.
Overdue cases must not disappear through reclassification, and a weekly dashboard is not a response mechanism for high-priority work.
Design choices
- Age matters, but age alone is an inadequate triage rule. Prioritizing only urgent work without an aging safeguard can leave routine patients waiting indefinitely.
- Distinguish three information states: urgency at ordering, urgency inferred from clinical context before interpretation, and urgency identified after a finding is recognized.
- Give each priority category a locally approved expected interval, an escalation threshold, an accountable recipient, and a backup recipient.
- Specify whether time runs continuously or in business hours, and how weekends, downtime, external examinations, and incomplete technical information are handled.
- Preserve original urgency and due-date values, with the reason for any revision.
Hypothetical example
A patient’s examination enters a routine queue, and the referring service later reports a material change in symptoms. Staff document the update and route it through the established clinical escalation pathway. The registry captures the revised clinical decision and the resulting assignment. The patient does not stay in the original queue because that queue is administratively convenient.
All examples are hypothetical illustrations.
Success measures
- Share of cases overdue within each clinically defined category
- Oldest-case age
- Review of extreme delays
- Aging review of lower-priority work approaching the local threshold
Failure conditions and tradeoffs
The tradeoff is responsiveness against instability. If every service can label its work urgent without review, prioritization loses meaning. If reclassification is too rigid, changing clinical needs are missed. Medical leadership audits both escalation appropriateness and persistent low-priority waiting, and protects staff who raise credible concerns.
Self-check: three controls
Every priority category has an approved interval, an escalation threshold, an accountable recipient, and a backup recipient.
Changes to urgency or due dates keep the original values and the reason for revision.
An aging safeguard reviews lower-priority work that approaches or exceeds the local threshold.
Component 3 of 8
Match effective capacity to the actual work
- Required operating product
- Coverage model by work type, shift, and availability.
- Accountable leadership role
- Practice leader and health-system operations executive.
Recovery is credible only when the proposed capacity is both deliverable and sustainable.
Design choices
- Begin with observed arrivals by modality, subspecialty, care setting, day, and shift. A total study count can mislead when the mix of complex work changes.
- Maintain a rolling coverage forecast that separates contracted hours, scheduled hours, credentialed availability, expected leave, and realized productive capacity.
- Identify coverage that cannot readily be substituted, such as specific subspecialty or procedural responsibilities.
- Keep recruiting plans in a separate future-capacity scenario until the clinician can actually contribute.
- Protect time for consultation, critical-result communication, quality work, and recovery from sustained workload. Fatigue research (Stec et al., 2018) supports including fatigue and diagnostic quality in any plan that relies on longer shifts.
Hypothetical example
A service receives 500 comparable work units per operating day and can reliably complete 480. Adding enough capacity to reach 500 stabilizes the queue but does not eliminate the accumulated backlog. Recovery needs a defined surplus above incoming demand, a clinically appropriate reduction in avoidable demand, or both, sustained long enough to clear the identified backlog.
All examples are hypothetical illustrations.
Success measures
- Arrivals and completions
- Effective coverage and uncovered shifts
- Aged cases
- Quality signals
- Workflow friction: unavailable prior images, repeated clarification requests, fragmented worklists, interruptions
Failure conditions and tradeoffs
The failure condition is a plan that assumes everyone will work faster indefinitely. When output gains depend on removing consultation time or postponing quality review, the organization may be borrowing against another part of its clinical service.
Self-check: three controls
The coverage forecast separates contracted, scheduled, credentialed, and realized productive capacity.
Promised recruitment is held in a future scenario and is not counted as current capacity.
The plan protects time for consultation, critical-result communication, quality work, and recovery.
Component 4 of 8
Make contract oversight clinically meaningful
- Required operating product
- Service specification, early-warning indicators, and continuity plan.
- Accountable leadership role
- Executive sponsor with medical and contracting leaders.
The test is whether the arrangement creates timely visibility, workable incentives, and enforceable operational responsibility.
Design choices
- Define the operating relationship, not only price and aggregate performance: scope, coverage, escalation, quality review, data access, service continuity, and work outside routine assumptions.
- Build a shared service specification: which examinations are included, when responsibility begins, how unavailable comparisons are handled, who communicates urgent findings, how preliminary and final discrepancies are resolved, and how unresolved cases transfer between providers.
- Use measurable event definitions so both parties calculate performance from the same population and timestamps.
- Keep information requests proportionate. A hospital does not need every internal business detail to require credible evidence that promised coverage can be delivered.
Early-warning indicators of service resilience
- Vacancy duration
- Credentialed coverage
- Departures
- Dependence on a small number of clinicians
- Unfilled shifts
- Feasibility of replacement plans
Scenario tests
- An unexpected departure
- A credentialing delay
- A system outage
- Failure of a major outside provider
These are proposed governance features, not a legal assessment of any existing contract.
Hypothetical example
In a multi-vendor recovery, a second group accepts a set of examinations. The handoff establishes a unique assignment, access to prior studies, clinical contact routes, and responsibility for finalization and discrepancies. If two vendors read the same examination while an older examination remains unassigned, additional spending has not become proportional clinical capacity.
All examples are hypothetical illustrations.
Success measures
- Service intervals
- Oldest-case exposure
- Quality
- Communication completion
- Continuity indicators
- Who can authorize temporary coverage, what triggers a recovery plan, and how its effectiveness is verified
Failure conditions and tradeoffs
Financial remedies alone are insufficient if the underlying problem is a shortage of qualified capacity. A collegial relationship should not substitute for documented escalation when service deteriorates. Employment, independent contracting, and hybrid arrangements each require these controls. The ownership form does not establish reliability.
Self-check: three controls
Both parties calculate performance from the same population, event definitions, and timestamps.
The health system receives early-warning indicators such as vacancy duration, departures, and unfilled shifts.
The contract names who can authorize temporary coverage and what triggers a recovery plan.
Component 5 of 8
Support patients across the clinical handoff
- Required operating product
- Navigation workflow and documented transfer of responsibility.
- Accountable leadership role
- Ambulatory or diagnostic-safety leader.
The patient should not be required to reconcile conflicting institutional instructions alone.
Design choices
- Identify who will explain the result, who will arrange any required care, how the patient can get assistance, and what happens when contact or access fails.
- Keep role boundaries explicit. A navigator coordinates; clinical interpretation and recommendations stay with qualified clinicians.
- Sustain the work with manageable caseloads, language access, reliable contact information, and backup coverage.
The minimum communication record: seven distinct states
- Attempted contact
- Successful contact
- Clinician explanation when required
- Patient understanding assessed
- Accepted referral
- Scheduled action
- Completed action
“Message left” is not receipt. A referral order is not an appointment.

Hypothetical example
A patient receives an actionable result but cannot schedule the recommended appointment because the receiving clinic requires additional documentation. The navigator identifies the missing item, coordinates its completion, and keeps ownership of the unresolved administrative barrier until the receiving service accepts the handoff. The responsible clinician reviews any delay that changes the clinical plan.
All examples are hypothetical illustrations.
Success measures
- Completed or clinically justified dispositions, not call volume alone
- Review of repeated unsuccessful contacts
- Distinct dispositions: documented patient refusal, unreachable status, accepted transfer
Failure conditions and tradeoffs
Documented refusal, unreachable status, and accepted transfer are distinct dispositions. None should be used casually to improve closure rates.
Self-check: three controls
The communication record distinguishes attempted contact, successful contact, scheduled action, and completed action.
A navigator keeps ownership of an unresolved barrier until the receiving service accepts the handoff.
Refusal, unreachable status, and accepted transfer are recorded as distinct dispositions and reviewed.
Component 6 of 8
Measure the whole service and protect the denominator
- Required operating product
- Balanced dashboard with quality and equity safeguards.
- Accountable leadership role
- Quality leader and analytical lead.
No single composite score should conceal a serious deterioration in one domain.
Design choices
- Answer four questions: how much work is arriving, how long patients are waiting, whether the next action occurs, and what costs or harms accompany improvement.
- Publish a metric dictionary before setting targets. Every measure needs a population, numerator, denominator, observation window, timestamp source, exclusions, owner, and review interval.
- Display completed-case turnaround and open-case aging separately. A case still open at period end must not vanish from exposure measures.
- Report newly discovered cases, exclusions, changed urgency labels, cancellations, and external transfers as a reconciliation, so a falling backlog cannot be manufactured by changing who counts.
- Use a concise operational view for daily work and a broader governance view for monthly review.
Scorecard domains
- Throughput
- Open-case exposure
- Communication
- Downstream action
- Quality
- Workforce
- Equity
Hypothetical example
A department improves its median turnaround while its oldest cases become older. The dashboard should call this a mixed result. New routine work may be flowing while difficult cases remain stranded. A review examines those cases, their barriers, and whether the operating rules encourage staff to select easier work.
All examples are hypothetical illustrations.
Success measures
- Separate flow and exposure displays
- Denominator reconciliation each period
- Monthly sampling of apparently closed cases against the documented evidence
Failure conditions and tradeoffs
The tradeoff is information burden. An extensive scorecard can overwhelm frontline teams. Remove measures that consume effort without changing a decision, and keep the safeguards against selective or misleading reporting.
Self-check: three controls
A metric dictionary was published before performance targets were set.
Completed-case turnaround and open-case aging are displayed separately.
Apparently closed cases are sampled monthly against the documented evidence.
Component 7 of 8
Recover the backlog without displacing today’s patients
- Required operating product
- Time-limited recovery plan with protected prospective service.
- Accountable leadership role
- Incident lead and finance partner.
Temporary relief becomes durable only when the organization has corrected the mechanism that produced the backlog.
Design choices
- Define the starting population, clinical review, assigned capacity, and a completion test.
- Separate the legacy cohort from new arrivals while keeping one shared clinical prioritization process.
- Decide which constraint actually limits progress. More readers help only if interpretation is the bottleneck. Missing priors, technical holds, missing recipients, or unavailable appointments need an exception team with authority over noninterpretive barriers.
- Set a daily plan: capacity surplus, expected reduction in aged work, protected level of prospective service, and safety limits that trigger reassessment.
- Update the projected recovery date when the evidence changes. Holding an optimistic deadline after its assumptions fail is not effective leadership.
Hypothetical example
For a legacy queue of 1,000 comparable studies, a net surplus of 50 completions per operating day clears the queue in 20 operating days under constant conditions. That statement is arithmetic, not a prediction. Complexity, absences, data defects, rework, and urgent arrivals can change the result.
All examples are hypothetical illustrations.
Success measures
- Actual arrivals and completions against assumptions
- Legacy cases with a supported disposition
- Prospective service performance during recovery
- Quality signals
- Rebound after temporary capacity is withdrawn
Failure conditions and tradeoffs
The main failure mode is declaring victory when the reading list looks smaller. Recovery ends only after a sustained verification period, with transfers, refusals, and revised clinical plans visible as separate outcomes.
Self-check: three controls
The legacy cohort is tracked separately from new arrivals under one shared prioritization process.
The binding constraint was identified before reading capacity was purchased.
Recovery will be declared complete only after a sustained verification period.
Component 8 of 8
Make accountability durable and learning routine
- Required operating product
- Board review, independent sampling, and learning cycle.
- Accountable leadership role
- Executive sponsor and board quality committee.
The purpose is a service that can explain its decisions and improve them, not a dashboard that remains green regardless of patient experience.
Design choices
- Run a governance cadence: a daily operational review during instability, a weekly cross-service review of unresolved barriers, and a monthly executive assessment of performance, resources, and emerging risk.
- Give the executive sponsor a concise account of unresolved decisions, not a colored dashboard. Each issue carries a decision owner, a due date, the resources requested, and a record of what happened after escalation.
- Commission an independent sample review of closed and overdue episodes, including cases classified as transferred, declined, unreachable, or no further action indicated.
- Route serious events through the established patient-safety processes.
Questions for the board quality committee
- Is oldest-case exposure declining?
- Do closure classifications withstand sampling?
- Do affected patient groups experience comparable completion?
- Can frontline staff raise concerns without having to prove that harm has already occurred?
Hypothetical example
The backlog returns several weeks after temporary coverage ends. The response is to compare the revised arrival pattern, realized capacity, and unresolved workflow defects with the original recovery assumptions. A recurrence prompts learning about the system. It is not automatically attributed to poor individual effort or treated as a reason to suppress unfavorable measures.
All examples are hypothetical illustrations.
Success measures
- Oldest-case exposure trend
- Closure audit agreement from independent sampling
- Stratified completion across patient groups
- Unresolved decisions past their due date
Failure conditions and tradeoffs
The tradeoff is useful oversight against punitive surveillance. A punitive response invites defensive documentation and reclassification. An informal response lets persistent failures continue. Effective accountability distinguishes system constraints, reasonable clinical judgment, preventable process breakdowns, and repeated failure to act on known responsibilities.
Self-check: three controls
Every unresolved issue has a decision owner, a due date, requested resources, and a record of what followed escalation.
The board quality committee reviews oldest-case exposure and sampled closure classifications.
Independent sampling includes transferred, declined, unreachable, and no-further-action dispositions.
Readiness profile
The profile shows each component separately. It reports no composite score, because the model’s own rule is that no single number should conceal a serious deterioration in one domain. The bands are a planning aid for local discussion and have not been validated.
- Not rated
- Not rated
- Not rated
- Not rated
- Not rated
- Not rated
- Not rated
- Not rated
Controls rated
0 of 24
Components with all three controls evidenced
0 of 8
Rate the three controls under each component’s Self-check to build the profile.
Governance architecture
Job titles should be adapted locally, while preserving a single named accountable person for each operating product.
| Component | Required operating product | Accountable leadership role |
|---|---|---|
| Auditable episode registry with owners and exception states. | Imaging operations leader with informatics lead. | |
| Clinician-approved urgency and escalation rules. | Radiology medical director with receiving services. | |
| Coverage model by work type, shift, and availability. | Practice leader and health-system operations executive. | |
| Service specification, early-warning indicators, and continuity plan. | Executive sponsor with medical and contracting leaders. | |
| Navigation workflow and documented transfer of responsibility. | Ambulatory or diagnostic-safety leader. | |
| Balanced dashboard with quality and equity safeguards. | Quality leader and analytical lead. | |
| Time-limited recovery plan with protected prospective service. | Incident lead and finance partner. | |
| Board review, independent sampling, and learning cycle. | Executive sponsor and board quality committee. |
Clinical boundaries of a shared obligation
Executive sponsor
Accountable for the operating system, staffing, resources, and unresolved cross-department barriers.
Medical director
Retains authority over clinical prioritization and interpretation quality.
Ordering and receiving clinicians
Remain responsible for clinical decisions within their scope.
Navigation and analytical staff
Make unresolved work visible and support follow-through. They do not determine clinical urgency or substitute for clinical judgment.
The ownership test
The framework is useful only when ownership includes the authority, information, and resources needed to act.
If a navigator finds an overdue specialist appointment but cannot reach anyone with scheduling authority, the design is incomplete. If a hospital can measure a contractor’s missed target but cannot obtain staffing forecasts or activate alternate coverage, contract oversight is incomplete.
Author’s assessment
A department can finish its scans and still leave care unfinished
An imaging examination promises that someone will interpret the findings, communicate their significance, and help the patient reach an appropriate next step. Organizational reporting frequently ends much earlier. This investigation examines that gap and the leadership arrangements that keep it difficult to see.
277
Claims-based reporting lag in 2024, indexed to 2014 = 100
Drake et al. (2026)
8.5%
Annual radiologist practice turnover in 2022, up from 5.3% in 2013
Parikh et al. (2026)
92 of 1,196
Critical imaging alerts lacking timely follow-up within four weeks, one VA outpatient setting
Singh et al. (2009)
5 weeks
Nonemergency result waits described by Main Line Health patients
Brubaker (2026), patient accounts
Three conclusions the evidence supports
First
National claims-based research identifies worsening reporting lag, although its date-based measurement cannot describe every patient’s actual waiting time.
Second
Physician availability is shaped by retention, coverage, workload, and the distribution of expertise, not simply national headcount.
Third
Interpreting an examination and completing the resulting clinical work are different processes. Research supports tracking both, with more limited evidence about long-term patient outcomes.
The central leadership test
Four things the organization must be able to show, at the level of individual cases as well as aggregate performance. Mark the ones you could evidence today.
0 of 4 evidenced. A convincing answer requires all four.
More interpretation capacity may be necessary. It will not, by itself, establish that clinical responsibility transferred or that recommended care occurred.
Recommended operating principle
Closure must be clinically justified and auditable. Administrative disappearance from a worklist is not sufficient evidence of completed care.
The report recommends a staged 90-day implementation with explicit clinical oversight, a measurement dictionary, an illustrative recovery budget, and an evaluation plan that protects diagnostic quality and workforce sustainability.
Scope and reading guide
How the investigation was assembled
The inquiry began with a current radiology leadership story and followed four evidence streams:
- Institutional statements and contemporaneous journalism
- Peer-reviewed research on reporting turnaround and workforce dynamics
- Research on communication and follow-up
- Professional guidance on communicating imaging findings
Searches combined radiology, turnaround time, backlog, workforce, turnover, incidental findings, diagnostic completion, and closed-loop communication. This was a focused critical review, not a registered systematic review or a meta-analysis. Documented observations are kept apart from managerial inference and proposed design. Unproven causal explanations remain hypotheses.
Boundaries of the conclusions
The scope is diagnostic imaging interpretation and subsequent action in U.S. health systems, with emphasis on outpatient and nonemergency work. Emergency findings require established urgent escalation pathways. Breast imaging, interventional procedures, and other services with distinct requirements need service-specific integration. No single turnaround deadline is presented as appropriate for every examination.
The most important missing evidence is local: complete worklist histories, timestamp reliability, staffing schedules, contract obligations, patient contact records, and downstream outcomes. That limitation drives the model’s first requirement: build an auditable account of unfinished work.
Report sequence
- Establish what is documented and where inference begins. Report pp. 4 to 8.
- Define the pathway and separate four types of backlog. Report p. 9.
- Translate evidence into ownership, workflow, and decision rules. Report pp. 10 to 18.
- Specify operational definitions and a feasible adoption sequence. Report pp. 19 to 21.
- Test capacity, costs, effectiveness, and unintended consequences. Report pp. 22 to 24.
- Identify evidence still needed and provide a verified source trail. Report pp. 25 to 27.
Evidence status. Public-source investigation and critical synthesis. Research and reporting reviewed through November 15, 2026. No original interviews, patient-record audit, or institution-specific outcome analysis was performed. Editorial illustrations depict representative situations, not documented patients or events.
The current case
A visible delay and an incomplete public record
On November 13, 2026, The Philadelphia Inquirer reported that Main Line Health patients described nonemergency imaging-result waits reaching five weeks. These are attributed statements and patient accounts, not an independently audited turnaround distribution.
Over two years
CEO Ed Jimenez attributed the pressure to a roughly 20% decline in the contracted group’s radiologist count alongside approximately 20% imaging-volume growth.
April
Executives reportedly became aware of staffing shortages.
Response
The system selected three additional interpretation companies.
November 13, 2026
Public reporting of waits reaching five weeks (Brubaker, 2026). The system’s imaging webpage acknowledged that results could take longer because of a national diagnostic-radiologist shortage (Main Line Health, n.d.).
After Thanksgiving
Improvement expected. It remained prospective at this report’s November 15 evidence cutoff.
What the two attributed percentages imply together
Fewer readers and more studies compound. Move the sliders to see the implied change in studies per radiologist. This is arithmetic on attributed statements and assumes work is spread evenly; it is not a measured workload.
Implied studies per radiologist
+50.0%
1.20 times the volume divided by 0.80 times the readers equals 1.50 times the studies per radiologist.
Author-derived illustration from Brubaker (2026). Not a finding of the report.
The leadership issue
A contracted clinical service remains part of the patient’s experience of the health system. If recruitment difficulties, departures, or coverage gaps accumulate outside the hospital’s direct employment structure, leaders still need a way to determine whether promised service remains achievable. The relevant question is what information reached decision-makers, when it arrived, and which operational decisions followed.
Three competing explanations that can coexist
Explanation 1
Absolute shortage
Too little appropriately qualified interpretation capacity.
Explanation 2
Mismatch
Available capacity does not match the timing or subspecialty composition of demand.
Explanation 3
Coordination failure
Incomplete routing, missing comparisons, unresolved technical problems, or fragmented responsibility reduce the output of an already constrained workforce.
Public information reviewed here cannot estimate the relative contributions of the three. The case does not establish the prevalence of patient harm or the comparative performance of other institutions.
What a credible case review would require
| Question | Evidence needed |
|---|---|
| When did deterioration become detectable? | Daily age distributions, missed service targets, coverage rosters, and escalation logs predating public complaints. |
| Which patients experienced the longest delays? | Examination-level timestamps, urgency, setting, modality, referral source, and clinically reviewed outcomes. |
| Did extra coverage resolve the old queue? | Matched before-and-after cohorts, vendor assignments, duplicate-read controls, and reopened cases. |
| Did accountability transfer successfully? | Contract obligations, named clinical recipients, acknowledgment records, and documented next actions. |
The fair test of corrective work
Announcing additional vendors or staffing is an intervention, not proof of recovery.
A system should receive credit for sustained improvement when the evidence supports it. The test is whether the oldest clinically relevant work diminishes while new demand, interpretation quality, communication, and downstream access remain under control.
The national signal
Reporting lag is increasing, but measurement matters
Christensen et al. (2026) examined nearly 2.6 million matched Medicare imaging studies from 2014 through 2023, measuring the difference between technical-component and professional-component service dates.
2014 = 100; 2024 = 277, calculated from the reported rounded 177% increase. Source: Drake et al. (2026), as summarized by the Harvey L. Neiman Health Policy Institute (2026). A calendar-date claims proxy, not elapsed hours or patient follow-up time. Endpoints only; no intervening annual values are shown.
View the published figure

0.091 to 0.193
Mean lag in calendar days, 2014 to 2023
Christensen et al. (2026)
+177%
Increase in mean lag, 2014 to 2024, across 2.9 million office and hospital outpatient studies
Drake et al. (2026)
+27%
Increase from 2023 to 2024 alone
Neiman HPI (2026)
0
What a same-day study contributes to the measure, however many hours elapsed
Date-based design
Do not convert this measure into a clock-time service benchmark or treat it as a direct measure of patient-notification delay.
Flow measures and exposure measures answer different questions
Select a view to see what it counts and what it can hide.
Flow measures
Describe the cases completed in a period, such as completed-case turnaround. They look only at work that reached a final report.
Blind spot: survivorship. The most delayed cases may be absent because they have not yet reached final interpretation.
Exposure measures
Describe patients still waiting at the measurement cutoff, such as the daily open-case age distribution: median, 90th percentile, maximum, and counts by age band.
Blind spot: exposure alone does not show how quickly finished work moved through the service.
Report both
A daily open-case age distribution should accompany completed-case turnaround statistics. A mean can improve while a smaller group of patients continues to wait excessively. A change in case mix, data capture, or the fraction of studies finalized on a later calendar date can also shift the mean without the same change in every clinical pathway.
Strategic implication
The national trend is a case for better local surveillance, not a license to extrapolate a national mean into a specific hospital’s tail of overdue cases.
Workforce pressure and local capacity
National headcount does not equal usable coverage
Supply projections do not support a simple claim that the national workforce must shrink. Christensen, Parikh, et al. (2025) projected growth through 2055 under alternative residency assumptions, and the companion utilization study projected demographic demand growth that varies by modality and assumptions (Christensen, Drake, et al., 2025). Neither can establish whether a particular community will have enough overnight coverage, procedural expertise, or subspecialty capacity.
Annual practice turnover was 5.3% in 2013 and 8.5% in 2022 among 39,439 radiologists represented in the study. Source: Parikh et al. (2026). Turnover means changing practices, not necessarily leaving radiology; the observational design does not establish why an individual departed.
View the published figure

The practical unit of planning
Reliably available clinical time, matched to work the clinician can appropriately perform. A nominal full-time position includes much more than routine interpretation. Select each activity to see why counting every position as interchangeable productive time conceals a service deficit before a vacancy even occurs.
0 of 7 activities acknowledged. Routine interpretation appears to fill the whole position.
What the workforce evidence does and does not say
Parikh et al., 2026
Workload and turnover
A nonlinear association between estimated workload and turnover. No universal safe productivity threshold. The value for leaders is to challenge the assumption that additional volume can always be assigned without affecting retention.
Crummy, 2026
How workforce data are collected
The redesigned ACR/RBMA workforce survey shows why survey participation and respondent composition matter when comparing releases. Distinguish a useful professional signal from a representative estimate of all radiologists.
Local evaluation
What to examine at home
Staffing stability, coverage intensity, autonomy, schedule fairness, and the support available for difficult work.
Capacity rule
Distinguish promised recruitment from credentialed, scheduled availability.
Additional vendors become usable capacity only when routing, comparisons, communication, and discrepancy responsibilities are functioning.
The unfinished work after interpretation
Acknowledgment and clinical completion are different outcomes
A message can be sent successfully while the recommended appointment, imaging examination, or clinical decision remains incomplete. A management system should preserve these intermediate states instead of compressing them into a single “closed” status.
What each study contributes
Select a study to see what it found and how far the finding can be carried.
92 of 1,196 critical imaging alerts lacked timely follow-up within four weeks
A single Veterans Affairs outpatient setting. That is 7.7% of the selected alert cohort (author calculation from the reported counts).
Limit: the denominator was a selected alert cohort, not all imaging examinations. The study is historical and cannot establish current national prevalence. It shows why delivery and acknowledgment are not proof that clinical work followed.
Substantial variation in settings, definitions, and failure rates
A systematic review of ambulatory test-result follow-up. It supports attention to system design.
Limit: the highest reported rate is not a universal estimate. The same caution applies when organizations compare dashboards that define an open case differently.
A structured incidental-finding workflow with navigation across emergency departments
An example of explicit follow-up infrastructure.
Limit: a largely descriptive institutional design.
A tracking program in which some findings needed escalation beyond routine workflow
The radiology actionable findings tracking program shows that routine workflow does not close every loop.
Limit: a single-program description; definitions of resolution differ across studies.
Prolonged institutional experience with an electronic-record protocol and nursing follow-up
Nine years of outcomes support the feasibility of organized tracking.
Limit: conclusions about comparative effectiveness, mortality, or generalizability remain limited.
Quality measures focused on completion of recommended follow-up
A measure set for noncritical actionable incidental findings. The distinction between sending a message and completing the recommendation is the consequential one.
Limit: measure design does not establish an intervention effect size.
The patient’s view of the handoff
Steitz et al. (2023) surveyed portal users across four academic centers and found a strong preference for immediate access to results. Respondent selection limits generalization, and access preferences do not show that every result is understood.
The leadership implication is to support timely access with a clear route to interpretation and next-step assistance. A portal can be one communication channel without becoming the institution’s only completion strategy.
The ACR communication practice parameter emphasizes timely communication appropriate to the clinical situation and documentation of nonroutine communication (ACR, n.d.). It supports a defined escalation process. It does not justify imposing one operational deadline on every type of examination.
A recurring boundary problem
Patients move across the boundary between radiology and the treating clinician. Organizational information often does not.
Radiology may own interpretation and communication of an unexpected finding, while another clinician owns diagnosis and treatment. A workable model must make the transfer visible, accepted, and recoverable when it fails.
From completed scan to completed care
Four backlogs require four operating responses
Diagnostic completion requires observable transfers of responsibility. Each transition records who acted, when the event occurred, and what evidence supports the new state. Select a transition, then select a backlog to see where it stalls.
The examination exists and is technically complete. The record carries the acquisition timestamp, modality, location, and urgency at ordering.
A qualified radiologist has finalized the report. Preliminary and final report timestamps are both retained, so unsigned drafts stay visible.
An appropriate recipient has confirmed receipt of a finding that requires it. A portal release or an attempted call does not establish receipt.
The responsible clinician has recorded what should happen next, and a due date for it. An accepted transfer establishes a new owner.
The recommended step occurred, or a clinically justified alternative did. A scheduling order is not the action.
The closing state is supported by source evidence and can withstand sampling. Each alternative disposition is recorded distinctly.
Clinical urgency and escalation apply throughout the pathway. Original conceptual diagram. The pathway is a proposed management representation, not a validated clinical pathway. An accepted transfer can establish a new owner; it does not by itself prove that the recommended action was completed.
View the published figure

Backlog mix: which remedy does your queue actually need?
Select a backlog to light the transition where it stalls. Replace the placeholder counts with your own open counts to see which operating response dominates. Cases can appear in more than one view and must not be counted as separate patients without explicit rules.
Acquired examinations awaiting interpretation.
- Operating response
- Reading coverage.
- Stalls before transition
- 2
- What staff can honestly tell the patient
- “Your examination has been completed and is waiting to be interpreted by a radiologist.”
A preliminary interpretation or draft exists, but finalization is still required.
- Operating response
- Finalization workflow.
- Stalls before transition
- 2
- What staff can honestly tell the patient
- “A preliminary reading exists. The final report still needs to be signed.”
Results that require confirmed receipt by an appropriate recipient.
- Operating response
- Recipient verification and escalation.
- Stalls before transition
- 3
- What staff can honestly tell the patient
- “The report is final. We are confirming that your clinician has received and reviewed it.”
Recommended or agreed clinical steps that have not occurred by the applicable due date.
- Operating response
- Navigation, appointment access, clinical review, or resolution of patient barriers.
- Stalls before transition
- 5
- What staff can honestly tell the patient
- “Your clinician has reviewed the result. A recommended next step has not been completed yet, and here is who is working on it.”
Share that more reading capacity can address
40.0%
Default counts are placeholders for demonstration, not data from the report or any institution.
Two states in every case record
Interpretation state
Unread, unsigned, or final, with who acted and when.
Action state
Receipt, decision, and the recommended step, each with an owner, a due date, and supporting evidence.
Dispositions that must be recorded distinctly
Completed
The recommended action occurred.
No further action indicated
Clinical review concluded that nothing more is needed.
Informed decline
The patient declined after an informed discussion.
Transferred
Care has moved to another team that accepted it.
The last three should not be described as completed treatment. Assigning all four backlogs to the interpreting radiologist creates responsibility without authority over the downstream resources required. Saying only that a case is “in process” hides the bottleneck from the people who could help resolve it.
Measurement dictionary
Eleven measures, with their owners and safeguards
Proposed local implementation specifications, not externally validated performance standards. Filter by family, then check your own queue arithmetic with the reconciliation tool.
Showing 11 of 11 measures
Flow and exposure
Open interpretation inventory
Unique acquired examinations without required final interpretation at the daily cutoff. Show unsigned and unread states separately.
Owner and use: Imaging operations; daily capacity and exception review.
Flow and exposure
Open-case age
Cutoff timestamp minus acquisition timestamp for each open examination. Report median, 90th percentile, maximum, and counts by locally approved age band.
Owner and use: Analytical lead; daily exposure surveillance.
Flow and exposure
Completed-case turnaround
Final-report timestamp minus acquisition timestamp for examinations finalized in the reporting period. Report by urgency, modality, setting, and shift.
Owner and use: Medical director; weekly review with open-case aging.
Flow and exposure
Overdue interpretation share
Open examinations past the approved category-specific deadline divided by all open examinations in that category at cutoff.
Owner and use: Operations and medical director; daily escalation.
Flow and exposure
Net queue change
New eligible entries plus reopenings minus valid exits during the same interval. Reconcile to opening and closing inventory.
Owner and use: Data steward; daily integrity check.
Flow and exposure
Unassigned episode share
Open episodes without a named next-action owner divided by all open episodes. Keep missing and invalid recipients visible.
Owner and use: Diagnostic-safety leader; daily ownership review.
Follow-through and balancing
Required communication completed on time
Findings with documented receipt by the designated recipient within the approved interval, divided by findings requiring that communication and due in the period.
Interpretation safeguard: A portal release or attempted call does not automatically establish receipt.
Follow-through and balancing
Recommended action completed on time
Actions documented as performed by the clinical due date, divided by recommended actions due in the period.
Interpretation safeguard: Report clinically justified alternatives, informed declines, transfers, and missing outcomes separately.
Follow-through and balancing
Closure audit agreement
Sampled closures supported by source evidence, divided by sampled closures reviewed.
Interpretation safeguard: Include difficult disposition categories; describe sample design and uncertainty.
Follow-through and balancing
Reopened episode share
Episodes reopened within a specified follow-up window, divided by closures with a complete observation window.
Interpretation safeguard: A reopening can indicate new information or a false closure; review the reason.
Follow-through and balancing
Quality and workforce balance
Report discrepancies requiring clinical action, repeat work, additional hours, unfilled shifts, and turnover with appropriate denominators.
Interpretation safeguard: Interpret trends with case mix and detection changes; do not combine unrelated outcomes into one score.
The clinical due date must be recorded before evaluation or retained with an auditable revision history.
Daily reconciliation check (net queue change)
Opening inventory plus new eligible entries plus reopenings, minus valid exits, should equal the closing inventory. A gap means cases entered or left the count without a recorded reason.
Expected closing inventory
950
Unexplained difference
0
The count reconciles.
Default values are illustrative.
Clock and target rules
- Synchronize time zones and document the handling of daylight-saving changes.
- Distinguish event time from data-import time.
- Treat missing or implausible timestamps as a data-quality exception.
- Report the amount of missing data beside the metric. Do not silently exclude it.
- Approve targets through clinical and operational governance after baseline measurement.
- A target requires a population and a purpose.
- Do not apply one numeric interval to urgent findings, routine interpretation, and months-long surveillance recommendations alike.
- Version threshold changes, and recalculate or annotate results when comparability is affected.
Equitable completion
Compare completion and waiting experience across relevant groups when data quality and sample sizes permit.
- Preferred language
- Age
- Geography
- Insurance category
- Disability-related assistance needs
- Portal access
Safeguards
- A missing demographic field is not a valid comparison category for concluding that no disparity exists.
- Small cells require privacy protection and cautious interpretation.
- Patient and staff interviews can guide hypotheses with appropriate consent. They are not representative prevalence estimates.
Review mechanisms before attributing a gap to patient behavior
- Contact methods
- Referral acceptance
- Travel requirements
- Appointment availability
- Financial barriers
Stratified review is proposed as a management safeguard, not as evidence that the named health system has any particular disparity.
A staged implementation plan
Ninety days to establish a testable operating system
Four phases, each with an accountable owner and an evidence gate. Tick the evidence you hold. A phase opens for expansion only when every item in its gate is met. Timing is an illustrative planning structure, not a guarantee of recovery or a clinical deadline.
0 of 4 gates met. Begin with Days 1 to 14: establish control.
Days 1 to 14
Establish control
Work and accountable owner. Executive names sponsor; medical director approves escalation; operations reconciles baseline; data steward validates feeds.
Evidence required before expansion
Gate not yet met
Days 15 to 30
Pilot one pathway
Work and accountable owner. The service leader tests the registry, navigation, recipient verification, and daily exception review in a defined cohort.
Evidence required before expansion
Gate not yet met
Days 31 to 60
Recover and evaluate
Work and accountable owner. Incident lead activates justified capacity surplus; finance tracks incremental cost; quality monitors diagnostic and workforce balance.
Evidence required before expansion
Gate not yet met
Days 61 to 90
Stabilize and transfer
Work and accountable owner. Executive reviews outcomes; permanent owners assume routines; temporary capacity is tapered only when justified.
Evidence required before expansion
Gate not yet met
The stop rule
Pause expansion while the defect is corrected if any of these appear:
- Data feeds are unreliable.
- Clinical escalation cannot be executed.
- Staff are closing cases without adequate evidence.
This is a control on implementation quality, not a reason to postpone urgent clinical care. During correction, the organization needs a workable temporary process for affected patients.
The implementation team
Participation should be tied to actual decision rights, not ceremonial membership. The team needs protected time for analysis and clinical work.
- Executive sponsor
- Radiology medical director
- Operations lead
- Informatics representative
- Quality professional
- Navigation lead
- Referring-service representative
- Scheduling lead
- Finance partner
Patient representatives can improve communication design and identify assumptions that professionals may overlook.
Six training cases
A policy is not operational until the people expected to use it can perform the workflow. Select a case; staff should demonstrate all three steps for it.
A missing recipient
- Document the event. Record that no valid recipient exists for the finding and keep the missing recipient visible in the registry.
- Assign responsibility. Name a next-action owner so the episode does not sit unassigned.
- Escalate uncertainty. Route to the backup recipient or qualified clinical review under the approved escalation rules.
The day 90 decision
Is the operating system reliable enough to sustain and expand?
Judge success from evidence of functioning responsibility, not from completed training or an installed dashboard. A longer stabilization period may be necessary when staffing, referral capacity, or data integration remains unresolved.
The arithmetic of backlog recovery
Matching demand prevents growth but does not clear old work
A stock-and-flow model of the open queue: B(t+1) = max[0, B(t) + arrivals − completions]. Set your own inputs, or load one of the report’s three scenarios.
Backlog recovery simulator
Report inputs: opening backlog of 1,000 comparable studies and 500 arrivals per operating day, with daily completions of 480, 500, or 550.
Sensitivity analysis
Net daily change
−50
Backlog at day 20
0
Operating days to clear
20
View the published figure

Assumes constant conditions, no reopenings or cancellations, and sufficient qualified capacity for the mix. These are illustrative inputs, not observed institutional data. Do not present a single deterministic date as certain when its inputs are unstable.
What the three report scenarios show
480 completions per day
Growth
The queue grows by 20 studies per operating day and reaches 1,400 after 20 operating days.
500 completions per day
Stabilization
The queue stays at 1,000. Demand is matched and nothing old is cleared.
550 completions per day
Recovery
A daily surplus of 50 clears the modeled legacy inventory in 20 operating days.
Expand the model locally
- Variable arrivals
- Available clinical time
- Case complexity
- Technical holds
- Rework
- Downstream actions generated by interpreted examinations
What this calculation cannot do
The stock-and-flow calculation does not estimate patient harm.
Clinical risk depends on the finding, indication, patient circumstances, and appropriate next action. The calculation shows whether the operating plan can reduce unfinished work. Clinical review determines how to prioritize that work.
Recovery economics
Fund the full work and avoid invented savings
The calculator opens on the report’s hypothetical planning example (Table 7). Every rate and hour is an arbitrary assumption for demonstrating the calculation, not a market benchmark, vendor quotation, recommended compensation, or a budget for any named organization.
Recovery budget calculator
Replace each input with your own figures.
| Input | Calculation | Modeled cost |
|---|---|---|
| Additional interpretation capacity | 50 studies/day × 20 days × $18 | $18,000 |
| Navigation support | 80 hours × $45/hour | $3,600 |
| Analytical and reconciliation support | 40 hours × $60/hour | $2,400 |
| Planning subtotal | Sum of the three inputs | $24,000 |
| Contingency | 10% of planning subtotal | $2,400 |
| Total illustrative budget | Subtotal plus contingency | $26,400 |
Cost per additional examination interpreted
$26.40
Cost per supported episode
Not entered
Share spent on interpretation
The per-examination figure does not establish cost per completed patient episode. One examination can produce multiple follow-up actions, and a patient may have several examinations.
Four rules for an honest business case
Revenue
Clearing already acquired examinations should not automatically be credited with new imaging revenue.
Benefits
Less rework, fewer repeated calls, fewer avoidable cancellations, and better use of clinical appointments each require local measurement. Avoided adverse events and liability costs should not be inserted as guaranteed savings.
Cost types
Separate one-time setup (interfaces, registry configuration, training, initial reconciliation) from recurring service costs (navigation, coverage, maintenance, quality review).
Sensitivity
Vary additional capacity, incremental unit cost, navigation workload, and the share of cases requiring downstream intervention.
The most important economic question
Are resources being applied to the actual constraint?
Paying for more interpretation while unresolved scheduling barriers dominate the final stage may produce little improvement in diagnostic completion. A program can be justified by a credible patient-safety obligation even when a precise positive financial return cannot be demonstrated. A recovery funded only through temporary overtime may improve the visible queue while leaving the future service under-resourced.
Evaluating effectiveness
A smaller queue is a result that still needs explanation
Prespecify the primary outcome, eligible population, observation period, and comparison strategy. A suitable initial primary outcome might be the share of clinically defined actions completed by their due date, accompanied by open-case exposure and interpretation turnaround.
Claim checker: what can your program say it achieved?
A program should claim success only for what it measured. Select the scope of your pilot.
- Interpretation turnaround improved
- Open interpretation inventory and open-case age fell
- Required communication was completed on time
- Recommended actions were completed by their due date
- Closures are supported by sampled source evidence
- Patient outcomes or mortality improved (needs a stronger design and a clearly measured endpoint)
A program focused only on reading capacity should not claim success for downstream clinical completion that it did not measure.
The count moved. What else could explain it?
Stable definitions and a traceable reconciliation of entries and exits are what separate a real change from an artifact.
A falling count can reflect
- Real completion of care
- Deletion
- Recoding
- External transfer
- Relaxed deadlines
- Changed exclusions
Test it by sampling closed cases against source records.
A rising count can reflect
- Real deterioration
- Better detection: a more complete registry finding previously invisible cases
- Reopenings after false closures
An apparent deterioration can represent better detection. Explain initial registry growth as newly discovered work when that is what it is.
Design the evaluation
Comparison strategy
Begin with a sufficiently stable baseline and a prospective sequence of weekly observations. Where feasible, a comparable service or staggered rollout helps separate the intervention from broader changes in staffing, demand, or scheduling. An interrupted time-series approach may be useful with enough observations and clear intervention dates; it requires attention to secular trends, seasonality, autocorrelation, and concurrent changes. A simple before-and-after average should be described as such.
Case mix
Examine modality, urgency, care setting, site, and relevant patient characteristics. The intervention may appear effective if the later cohort contains easier work or excludes difficult referrals. Prespecify and report clinically justified exclusions. Time-to-completion analyses should retain patients who remain unresolved at the observation cutoff.
Outcome verification
Include sampled source-record review. Test whether “completed,” “transferred,” “declined,” and “no further action” classifications match the underlying evidence. Patient-reported understanding and ability to obtain help can complement process measures, with response rates and sampling limits disclosed. Complaints are useful signals but not a complete or representative outcome dataset.
Balancing outcomes
Track diagnostic discrepancies requiring action, avoidable repeat work, additional hours, staffing stability, and downstream appointment pressure. Faster reporting achieved by shifting unsustainable work onto clinicians or navigators may not be durable. More detected discrepancies after better surveillance is not automatically worse diagnostic performance.
Revise and publish
Revise the model if the evaluation finds ineffective components, excessive documentation burden, or unintended access barriers. Report both success and failure, distinguish implementation fidelity from patient benefit, and disclose concurrent interventions.
Two different kinds of evidence
Evidence that a program can be implemented is valuable. Evidence that it improves clinically meaningful outcomes requires a stronger design and a clearly measured endpoint.
Evidence appraisal
What each evidence stream can support
An author synthesis, not a formal certainty-of-evidence grading exercise. The cited studies address related but distinct populations and outcomes.
Showing 8 of 8 evidence rows
| Evidence | Contribution | Principal limitation |
|---|---|---|
| Current caseMain Line Health reporting and notice, 2026 | A current example of publicly acknowledged results delays and corrective plans. | No examination-level audit or verified distribution of harm was available. |
| National signalChristensen et al., 2026; Drake et al., 2026 | A national longitudinal signal using matched Medicare claims. | Date-based proxy; cannot measure all clinical handoffs or explain local causes. |
| WorkforceSupply and utilization projections, 2025 | Scenarios for long-term workforce and demand planning. | Dependent on assumptions; national totals obscure local coverage and case mix. |
| WorkforceParikh et al., 2026 | Practice turnover and its association with estimated workload. | Observational claims-linked analysis; turnover is not retirement or causation. |
| Follow-upSingh et al., 2009; Callen et al., 2012 | Evidence that abnormal-result follow-up can remain incomplete. | Older and heterogeneous settings; unsuitable for current national failure rates. |
| Follow-upFu et al., 2023; Whalen et al., 2025; Shah et al., 2026 | Operational examples of structured tracking and navigation. | Institutional interventions; definitions of resolution and outcome ascertainment differ. |
| GuidanceKadom et al., 2022; ACR, n.d. | Measurement concepts and professional communication guidance. | Guidance and measure design do not establish an intervention effect size. |
| Patient and fatigueSteitz et al., 2023; Stec et al., 2018 | Patient communication preferences and evidence about fatigue. | Differences in selection, setting, and design constrain generalization. |
A defensible interpretation
Where the evidence is strongest
The existence of measurable process gaps, and the feasibility of organizing work around those gaps.
Where it is less definitive
Which combination of staffing, incentives, navigation, or contract design produces the largest patient benefit in every setting.
That difference should influence investment decisions. A health system can justify improving visibility and responsibility without claiming that its chosen implementation has already been proven to reduce mortality. Reported performance should also separate process reliability from clinical effectiveness.
Author-derived checks on the published figures
These are simple calculations from numbers printed in the report. They are labeled as derived and are not additional findings.
| Published figures | Derived value | Reading |
|---|---|---|
| Mean lag 0.091 to 0.193 calendar days, 2014 to 2023 | +112.1% | The lag slightly more than doubled before the 2024 update. |
| +177% from 2014 to 2024; +27% from 2023 to 2024 | About 0.25 calendar days in 2024 by either route | 0.091 × 2.77 = 0.252 and 0.193 × 1.27 = 0.245. The two reported percentages agree within rounding, although the studies use different cohorts. |
| Turnover 5.3% to 8.5%, 2013 to 2022 | +3.2 points, or +60.4% relative | State which convention you are using when quoting the change. |
| 92 of 1,196 critical imaging alerts | 7.7% | A share of a selected alert cohort, not of all imaging. |
| $26,400 over 1,000 additional examinations | $26.40 each | Reproduces the report’s per-examination figure. |
What leadership must be able to answer
Questions for a follow-on institutional investigation
Six questions and the documents that would answer them. Record the evidence status for your own organization. The tally separates information that is unavailable from information that was requested but withheld, as the report asks a follow-on investigation to do.
Obtained
0
Pending
0
Unavailable
0
Withheld
0
Not requested
6
Question 1
Did leaders have an early warning?
Documents or observations to seek: Historical worklist extracts, escalation messages, staffing forecasts, and committee records.
Question 2
Was the response proportionate to the deficit?
Documents or observations to seek: Arrival and capacity assumptions, recovery scenarios, authorized resources, and plan changes.
Question 3
Did the contract preserve clinical continuity?
Documents or observations to seek: Coverage obligations, recipient responsibilities, discrepancy procedures, and contingency arrangements.
Question 4
Were patients told what to expect?
Documents or observations to seek: Patient-facing notices, call scripts, contact attempts, complaints, and patient interviews.
Question 5
Did apparent closure represent actual resolution?
Documents or observations to seek: Sampled episode histories, accepted transfers, action records, and reasons for alternative dispositions.
Question 6
Did recovery create another problem?
Documents or observations to seek: Quality trends, additional hours, retention indicators, downstream access, and stratified completion data.
Table 8. Proposed reporting agenda. These questions identify evidence to obtain; they are not allegations that a named institution failed each test.
Who a subsequent investigation should hear from
- Hospital leadership
- The contracted medical group
- Referring clinicians
- Frontline staff
- Affected patients
Questions should be specific enough to permit a substantive answer. The present report did not conduct those interviews or seek a new organizational response.
Source trail
References
Eighteen sources, reviewed through November 15, 2026. Links open in a new tab.
Showing 18 of 18 references
- Professional guidance
American College of Radiology. (n.d.). ACR practice parameter for communication of diagnostic imaging findings. Retrieved November 15, 2026. https://gravitas.acr.org/PPTS/GetDocumentView?docId=74
- Journalism and institutional
Brubaker, H. (2026, November 13). Main Line Health expects imaging delays to end after Thanksgiving, as a radiologist shortage forces a rethink. The Philadelphia Inquirer. https://www.inquirer.com/health/main-line-health-x-ray-mri-delays-radiology-20261113.html
- Peer-reviewed research
Callen, J. L., Westbrook, J. I., Georgiou, A., & Li, J. (2012). Failure to follow-up test results for ambulatory patients: A systematic review. Journal of General Internal Medicine, 27(10), 1334-1348. https://doi.org/10.1007/s11606-011-1949-5
- Peer-reviewed research
Christensen, E. W., Drake, A. R., Parikh, J. R., Rubin, E. M., & Rula, E. Y. (2025). Projected US imaging utilization, 2025 to 2055. Journal of the American College of Radiology, 22(2), 151-158. https://doi.org/10.1016/j.jacr.2024.10.017
- Peer-reviewed research
Christensen, E. W., Drake, A. R., Rula, E. Y., Yuan, C. X., Wald, C., Johnson, M. H., & Nicola, G. N. (2026). National turnaround time trends for Medicare fee-for-service beneficiaries, 2014-2023. Journal of the American College of Radiology, 23(7), 1244-1252. https://doi.org/10.1016/j.jacr.2026.02.038
- Peer-reviewed research
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. https://doi.org/10.1016/j.jacr.2024.10.019
- Professional guidance
Crummy, T. A. (2026, November 9). The 2026 ACR/RBMA workforce survey: Radiology’s workforce reality. ACR Bulletin. https://www.acr.org/Clinical-Resources/Publications-and-Research/ACR-Bulletin/2026/2026-acr-rbma-workforce-survey
- Peer-reviewed research
Drake, A. R., Yuan, C. X., Wald, C., Nicola, G. N., Johnson, M. H., Rula, E. Y., & Christensen, E. W. (2026). National turnaround time trends for Medicare fee-for-service beneficiaries: Updated through 2024. Journal of the American College of Radiology. Advance online publication. https://doi.org/10.1016/j.jacr.2026.06.015
- Peer-reviewed research
Fu, T., Berlin, S., Gupta, A., Plecha, D., Sunshine, J., & Sommer, J. (2023). Implementing a streamlined radiology workflow to close the loop on incidental imaging findings in the emergency department. Journal of Digital Imaging, 36(3), 776-786. https://doi.org/10.1007/s10278-022-00773-x
- Journalism and institutional
Harvey L. Neiman Health Policy Institute. (2026, August 4). Imaging interpretation turnaround time continued to increase in 2024, by 27%. https://www.neimanhpi.org/press-releases/imaging-interpretation-turnaround-time-continued-to-increase-in-2024-by-27/
- Peer-reviewed research
Kadom, N., Venkatesh, A. K., Shugarman, S. A., Burleson, J. H., Moore, C. L., & Seidenwurm, D. (2022). Novel quality measure set: Closing the completion loop on radiology follow-up recommendations for noncritical actionable incidental findings. Journal of the American College of Radiology, 19(7), 881-890. https://doi.org/10.1016/j.jacr.2022.03.017
- Journalism and institutional
Main Line Health. (n.d.). Imaging and diagnostic radiology. Retrieved November 15, 2026. https://www.mainlinehealth.org/specialties/imaging-and-diagnostic-radiology
- Peer-reviewed research
Parikh, J. R., Drake, A. R., Rula, E. Y., Golding, E., & Christensen, E. W. (2026). Radiologist turnover in the United States. Journal of the American College of Radiology, 23(6), 1058-1066. https://doi.org/10.1016/j.jacr.2026.01.009
- Peer-reviewed research
Shah, R. M., Dehadrai, A., Keshwani, A., Schmitz, B., & Carr, J. C. (2026). Closing the loop: 9-year outcomes of an electronic medical record-based protocol for reporting incidental imaging findings. Journal of the American College of Radiology, 23(1), 45-51. https://doi.org/10.1016/j.jacr.2025.08.047
- Peer-reviewed research
Singh, H., Thomas, E. J., Mani, S., Sittig, D., Arora, H., Espadas, D., Khan, M. M., & Petersen, L. A. (2009). Timely follow-up of abnormal diagnostic imaging test results in an outpatient setting: Are electronic medical records achieving their potential? Archives of Internal Medicine, 169(17), 1578-1586. https://doi.org/10.1001/archinternmed.2009.263
- Peer-reviewed research
Stec, N., Arje, D., Moody, A. R., Krupinski, E. A., & Tyrrell, P. N. (2018). A systematic review of fatigue in radiology: Is it a problem? American Journal of Roentgenology, 210(4), 799-806. https://doi.org/10.2214/AJR.17.18613
- Peer-reviewed research
Steitz, B. D., Turer, R. W., Lin, C.-T., MacDonald, S., Salmi, L., Wright, A., Lehmann, C. U., Langford, K., McDonald, S. A., Reese, T. J., Sternberg, P., Chen, Q., Rosenbloom, S. T., & DesRoches, C. M. (2023). Perspectives of patients about immediate access to test results through an online patient portal. JAMA Network Open, 6(3), e233572. https://doi.org/10.1001/jamanetworkopen.2023.3572
- Peer-reviewed research
Whalen, S., Trivedi, S., Herren, J., Fuguitt, K., & Bui, J. T. (2025). Improving communication of unexpected findings: The radiology actional findings tracking (RAFT) program. Current Problems in Diagnostic Radiology, 54(4), 418-421. https://doi.org/10.1067/j.cpradiol.2025.01.006