Reach out to me on LinkedIn for a copy of the research report
U.S. Market and Policy Analysis
When Acute Care Cannot Be Reached
Mobile integrated health as a value-based response to the aging constraint in the United States. A critical review of Wang et al. (2026), a causal-constraint analysis, and a 2026 to 2050 market outlook.
- Adults age 65+
- 2026 to 2050 outlook
- Census, CDC, CMS, HRSA
- Evidence current through August 10, 2026
- 16 sections, 6 interactive models
Central conclusion
Mobile integrated health is primarily an infrastructure solution to a crucial challenge in location-based care. It brings clinically appropriate urgent care closer to older adults when issues such as travel difficulties, frailty, cognitive impairment, caregiver shortages, workforce gaps, and payment restrictions prevent traditional access.
The strategic question is not whether to buy a mobile unit. It is whether the organization can assemble a reliable clinical operating system, and whether the party that creates the value is the party that captures it.
Adults age 65+
82.1 M by 2050
Up from 57.8 million in 2022, a 42 percent increase. Census projections, main series.
Adults age 85+
+168% by 2050
From 6.5 million to 17.4 million. This group grows about four times faster than the age-65+ population.
Annual ED visits, age 65+
46.8 M by 2050
From 32.9 million observed in 2022, holding the 2022 visit rate constant. A demographic exposure floor, not a forecast.
Macro support ratio
31.3 → 12.2
Adults age 18-64 per adult age 85+. A 61 percent decline in the macro support base by 2050.
What the report finds
Six findings, in the order the argument builds.
1. Demographic pressure is already measurable
The United States had 61.2 million adults age 65 and older in 2024. Census projections rise from 57.8 million in 2022 to 82.1 million in 2050, while the age-85+ population rises from 6.5 million to 17.4 million.
2. A conservative utilization floor is large
Holding the 2022 older-adult ED visit rate constant yields an illustrative increase from 32.9 million annual visits in 2022 to 46.8 million in 2050, a 42 percent increase. This is a demographic exposure scenario, not an official forecast or an estimate of avoidable visits.
3. The constraint should accelerate, not merely scale
Adults age 75 and older already had 76 ED visits per 100 people in 2022, compared with 46 per 100 for adults ages 65-74. Because the oldest-old share is rising, a constant age-65+ rate is conservative by construction.
4. The anchor article is operationally useful but not causally dispositive
Wang et al. identify plausible value mechanisms in a Finnish MIH pilot, including a paramedic and geriatric nurse dyad, point-of-care diagnostics, integrated health and social care, and standardized pathways. Yet a provider-only qualitative design cannot establish reductions in ED use, hospitalizations, costs, or patient-valued outcomes.
5. U.S. evidence is promising but heterogeneous
A 2023 meta-analysis found a pooled ED-visit risk ratio of 0.56, but heterogeneity exceeded 90 percent and all included studies were observational. A 2026 quasi-experimental study in frail older adults reported lower 30-day rehospitalization and ED use, while a rigorous 2021 propensity-matched study found no benefit on most short-term outcomes.
6. Commercial adoption will follow risk ownership
The highest-probability purchasers are integrated delivery systems, Medicare Advantage plans, ACOs, provider-sponsored plans, Medicaid managed care organizations, hospital-at-home programs, and rural systems with transformation funding. Traditional transport-linked ambulance payment remains the main scaling barrier.
Decision implication
U.S. organizations should adopt MIH selectively, not generically. The investable unit is a defined cohort in a defined geography with a closed-loop clinical pathway and a payer or provider that captures downstream savings.
Expansion should be conditional on prespecified safety, utilization, experience, equity, and total-cost outcomes.
Author-derived view
Scale and composition are not the same problem
Figure 1 presents the two populations on separate axes, which makes each trend legible but hides their relative speed. Indexing both to 2022 puts them on one scale. The divergence does not begin immediately: through 2025 the two series move together, and the oldest-old curve only separates after 2030.
Growth of the age-65+ and age-85+ populations, indexed to 2022 = 100
Same Census values as Figure 1, rebased so relative speed is directly comparable.
Values: U.S. Census Bureau (2023), 2023 National Population Projections, Table 2, as printed in Figure 1 of the report. Indexing is author-derived. Projections are of population, not of health service use.
Executive bottom line
By 2030 to 2040, MIH is likely to shift from an optional innovation to a practical capacity strategy for selected older populations.
Whether it becomes equitable national infrastructure or fragmented local access will depend more on payment and integration than on clinical technology.
How to use this dashboard
Six of the sixteen sections are working models rather than exhibits. Each one reproduces a published figure or table before it lets you move anything, so you can see what the report actually establishes and where your own assumptions begin.
ED exposure
Reproduces Appendix Table A1 exactly, then prices the compositional shift the report describes but does not quantify.
Claim discipline
Matches a claim you want to make against the study designs you actually hold, applying the report’s own causal-discipline rule.
Unit economics
Uses the published $550 and $1,400 visit costs to find the substitution rate a contract would require, and flags when value capture sits elsewhere.
Transferability
Scores your setting against the seven Finnish features in Table 3, with the two least transferable acting as binding constraints.
Investable unit
Five conjunctive conditions. It reports no score when a condition fails, because a partial pass is not a smaller version of the same thing.
Executive scorecard
Configures the nine Table 5 domains, with clinical safety acting as the gate the report specifies.
Section 3
What a causal constraint is
A causal constraint is a condition that limits whether a needed process can occur, even when demand exists. For older adults, the binding condition is often not the absence of a clinically appropriate intervention. It is the inability to make that intervention accessible at the right time and place.
Figure 4. Conceptual causal-constraint model and the point at which MIH intervenes. Framework synthesized from Wang et al. (2026), U.S. MIH evidence, Census projections, CDC utilization data, and CMS payment policy. It is a conceptual causal model, not an estimated structural model.
The conceptual pressure function
Acute-care pressure at time t equals population scale multiplied by clinical intensity multiplied by the gap between needed and accessible capacity. This is not a fitted equation. It clarifies why adding population alone understates risk: population, intensity, and the access gap can move in the same direction. MIH acts primarily on the access gap, by bringing assessment, diagnostics, treatment, and coordination into the home.
Six mechanisms intensify together
The report is explicit that these are not independent additive pressures. They compound, which is why the demographic exposure model in the next section is described as a floor rather than a forecast.
Scale
More older adults generate more acute episodes even if per-person utilization remains unchanged.
57.8M to 82.1MAdults age 65+, 2022 to 2050
Composition
The fastest growth occurs among adults aged 85+, the group with greater frailty, cognitive burden, and emergency use.
11.2% to 21.2%Age-85+ share of the 65+ population
Clinical complexity
Multiple conditions and polypharmacy make apparently minor symptoms harder to triage by telephone and harder to manage without diagnostics.
76 vs 46ED visits per 100, age 75+ vs 65-74
Support compression
The projected number of adults aged 18-64 per adult aged 85+ falls sharply. This is not a caregiver forecast, but it signals a smaller macro support base for transportation, monitoring, and unpaid care.
31.3 to 12.2Adults 18-64 per adult 85+
Capacity lag
HRSA projects a shortage of 141,160 physicians by 2038, including 70,610 primary care physicians and 1,570 geriatricians. Nonmetro areas are projected to experience a 39 percent shortage of primary-care physicians.
141,160Projected physician shortage by 2038
Financing mismatch
Traditional Medicare ambulance payment historically rewards transport to covered facilities rather than treatment in place. The ET3 demonstration created alternatives but delivered very low intervention volume and ended in 2023.
ET3 ended 2023Transport-linked payment remains the default
Where MIH acts
MIH relaxes the location constraint through field assessment, point-of-care diagnostics, physician oversight, and coordinated follow-up. It does not act on scale, and it acts on composition and complexity only indirectly, by making a home-based response clinically adequate for a frailer patient.
It will not replace emergency departments, primary care, home health, or hospital-at-home. Its value lies in connecting them when their boundaries fail.
What the model does not claim
- It does not estimate how many visits are avoidable or MIH-eligible.
- It does not treat provider perceptions, associations, or scenario projections as causal effects.
- It treats escalation to the ED as a safety outcome when home care is inappropriate, not as a program failure.
- It does not model morbidity change, technological change, prevention, hospital capacity, payer behavior, or substitution.
Section 3
The United States is entering a composition shift, not only a scale shift
The age-65+ population is projected to rise from 57.8 million in 2022 to 82.1 million in 2050. Over the same interval, the age-85+ population rises from 6.5 million to 17.4 million. The oldest-old share of the older population therefore rises from 11.2 percent to 21.2 percent.
Figure 1. Both scale and a faster shift toward the oldest-old population drive U.S. population aging. Source: U.S. Census Bureau (2023), Table 2.
Both populations on a single shared axis
Figure 1 uses two panels with different scales. Plotting both against one axis shows how small the oldest-old population still is in absolute terms, which is why its growth rate matters more than its headcount.
Values as printed in Figure 1. Shared axis is author-derived. Millions of people.
The oldest-old share of the older population
This ratio is stated in the text but not plotted in the report. It is the cleanest single statement of the composition shift.
Author-derived from the Figure 1 values: age-85+ population divided by age-65+ population. Note the dip through 2025, when the 65+ denominator grows faster than the 85+ numerator.
Why composition changes the service question
Service need is nonlinear across older age. In 2022, adults age 75 and older had 76 ED visits per 100 people, compared with 46 per 100 among adults ages 65-74.
What drives the higher rate
The report attributes it to accumulated multimorbidity, frailty, sensory and cognitive impairment, medication complexity, falls, and reduced physiologic reserve. It also means that a constant age-65+ utilization rate is conservative when the oldest-old share is rising, which is the basis for calling the projection a floor.
| Year | Age 65+ | Age 85+ | 85+ share | Index 65+ | Index 85+ |
|---|---|---|---|---|---|
| 2022 | 57.8 | 6.5 | 11.2% | 100.0 | 100.0 |
| 2025 | 63.3 | 7.0 | 11.1% | 109.5 | 107.7 |
| 2030 | 71.2 | 8.6 | 12.1% | 123.2 | 132.3 |
| 2035 | 75.8 | 11.2 | 14.8% | 131.1 | 172.3 |
| 2040 | 78.3 | 13.7 | 17.5% | 135.5 | 210.8 |
| 2045 | 79.8 | 15.9 | 19.9% | 138.1 | 244.6 |
| 2050 | 82.1 | 17.4 | 21.2% | 142.0 | 267.7 |
Population columns are as printed in Figure 1. Share and index columns are author-derived from those values and are shown to one decimal.
A detail worth naming
The acceleration is not immediate. Between 2022 and 2025 the age-85+ population actually grows slightly more slowly than the age-65+ population, so the oldest-old share dips from 11.2 percent to 11.1 percent before rising. The separation begins after 2030 and is steepest between 2030 and 2040, which is precisely the window the report identifies as the adoption inflection point. Planning that treats the pressure as already accelerating will misread the near term; planning that waits for it will be late.
Section 3 and Appendix A
A demographic floor under older-adult emergency demand
The report’s exposure model is deliberately austere. Projected ED exposure equals the projected age-65+ population multiplied by the 2022 observed age-65+ ED visit rate. No utilization trend, morbidity change, policy effect, or MIH effect is imposed.
Figure 2. Illustrative older-adult ED exposure under constant 2022 utilization. Sources: CDC/NCHS, 2022 NHAMCS (32.9M visits by adults age 65+); U.S. Census Bureau, 2023 projections. This is not an official utilization forecast and does not estimate avoidable visits.
Interpretation rule, stated in the report
The numerical forecast is a demographic floor. Because adults age 75+ use emergency departments more intensively than adults age 65-74, the compositional shift toward age 85+ creates a credible upward bias relative to a constant-rate model. Its magnitude cannot be estimated from the anchor article.
Interactive model, author-derived
Pricing the compositional shift the report leaves unquantified
The report says the upward bias exists but cannot be sized from the anchor article. It can, however, be bounded from the report’s own two published age-specific rates, once the 2022 composition is calibrated to reproduce the report’s own overall rate. That calibration is exact and requires no outside data.
The calibration
The report prints two age-specific 2022 rates (0.46 visits per person-year for ages 65-74, and 0.76 for age 75+) and one overall rate (32.90 million visits divided by 57.795 million adults, or 0.5693). Only one 75+ share reconciles them:
0.3642 x 0.76 + 0.6358 x 0.46 = 0.5693
So a 36.4 percent oldest-quartile share is embedded in the published model. Raising only that share, while holding both age-specific rates fixed, isolates the compositional effect from every other change.
What the slider does and does not represent
It sets the 75+ share of the 65+ population in 2050 and interpolates linearly back to the calibrated 2022 value. It is an exploration control, not a published Census projection: the report does not print the 75+ share by year.
For scale, the published age-85+ share of the 65+ population rises 10.0 points, from 11.2 percent to 21.2 percent. A rise of broadly similar order in the 75+ share is the natural exploration range. The default position reproduces Table A1 exactly, so any departure from the published numbers is one you made deliberately.
Composition-adjusted ED exposure model
Reproduces Appendix Table A1 at the default setting, then prices the compositional shift.
blended rate 0.5693 visits per person-year
Calibrated 2022 value is 36.4 percent. Both age-specific rates are held at their published 2022 values throughout.2050 published floor
46.75M
Constant 2022 rate, as in Table A1
2050 composition-adjusted
46.75M
Same age-specific rates, shifted mix
Difference
+0.00M(+0.0%)
Visits above the published floor
Versus 2022 observed
+42.1%
Change from 32.9M in 2022
The slider is at the calibrated 2022 composition, so the model reduces exactly to the published Table A1 scenario.
Published floor with the composition-adjusted overlay
Bars are the published constant-rate values. The dashed rules are the composition-adjusted levels at your current setting.
Bars: Appendix Table A1. Overlay: author-derived, holding the published age-specific rates fixed and varying only the age mix.
| Year | Age-65+ population (M) | 75+ share | Blended rate | Floor, ED visits (M) | Adjusted (M) | Difference |
|---|
The floor column reproduces Table A1. Note a small rounding artifact in the published table: it prints the rate as 0.569 but computes the visits column from the unrounded quotient 0.569253. Using the printed 0.569 instead would give 40.50, 44.55, and 46.73 rather than 40.52, 44.57, and 46.75. This dashboard uses the unrounded quotient so that it reproduces the published visit column exactly.
The rate difference that drives the adjustment
2022 emergency department visit rates by age group.
CDC/NCHS, 2022, as stated in the report body text. Age 75+ used emergency departments at 1.65 times the rate of adults ages 65-74.
What the model deliberately excludes
No attempt is made to estimate the share of visits that is avoidable or MIH-eligible. The exposure figure is the size of the demand surface, not the size of the addressable market.
The report also excludes morbidity change, technological change, prevention, hospital capacity, payer behavior, and substitution. Each of those could move the number in either direction.
How to read the adjustment honestly
The adjustment is arithmetic, not epidemiology. It assumes the two published age-specific rates stay fixed for another 28 years, which no one should believe literally. Its value is in showing that the direction of the bias is knowable and its rough size is small relative to the scale effect: at a 10 point compositional shift, roughly 5 percent on top of a 42 percent increase. The composition argument is real, but scale remains the larger term.
Section 3
The support base compresses as the oldest-old population expands
The projected number of adults aged 18-64 per adult aged 85+ falls from 31.3 in 2022 to 12.2 in 2050, a 61 percent decline. The report is careful about what this is: a macro support ratio, not a measure of actual caregiver availability and not a caregiver forecast.
Figure 3. A macro indicator of support compression as the age-85+ population grows. Source: derived from U.S. Census Bureau (2023), Table 2. Numerator: age 18-64; denominator: age 85+.
The compression is a numerator problem, not a denominator problem
Both series indexed to 2022 = 100. The support ratio does not fall because the working-age population shrinks. It falls because the oldest-old population nearly triples.
Author-derived indexing of the Figure 1 and Figure 3 values. Multiplying the published ratio by the published age-85+ population implies a working-age population that is broadly flat across the period, rising from about 203 million to about 212 million, which is consistent with the ratio being driven almost entirely by its denominator.
Support ratio, 2022
31.3
Adults age 18-64 per adult age 85+
Support ratio, 2050
12.2
A 61 percent decline in the macro support base
The report reads this as a signal about transportation, monitoring, and unpaid care rather than as a caregiver estimate. That distinction matters operationally: MIH substitutes for the transport and monitoring functions of an absent support base, not for the relationship.
Capacity lag on the professional side
HRSA projects a shortage of 141,160 physicians by 2038. Half of that gap is in primary care. The geriatrician gap is numerically small but proportionally severe, and nonmetro areas are projected to experience a 39 percent shortage of primary-care physicians.
Projected U.S. physician shortage by 2038
HRSA health workforce projections, as cited in the report.
HRSA (2025), health workforce projections. Primary care is a subset of the all-physician figure, and geriatricians are a subset again; the bars are nested categories, not additive components.
Why the geriatrician number is the one to watch
At 1,570, the projected geriatrician shortage is about 1.1 percent of the total physician gap in absolute terms. It matters disproportionately because the age-85+ population, the group whose care most benefits from geriatric assessment, is the fastest-growing segment in the projection. The report’s response is not to wait for more geriatricians but to extend geriatric judgment through the paramedic and nurse dyad, tele-geriatrics, and standardized pathways.
Timing
The HRSA projection reaches its 2038 horizon inside the same decade the report identifies as the demand acceleration window. Between 2030 and 2040 the age-85+ population rises from 8.6 million to 13.7 million, a 60 percent increase, while the workforce gap matures. That coincidence, rather than either trend alone, is what makes 2030 to 2040 the strategic decision point.
The strategic choice this creates
Health systems will face a choice: transport more complex older adults into already-constrained facilities, or expand the clinical radius of their existing workforce through mobile and remote teams.
The second option is not cheaper by default. It is a different way of spending the same scarce clinical hours, and whether it pays depends on who captures the downstream saving.
Section 2
Critical review of Wang et al. (2026)
Wang, Pekanoja, Ahokangas, and Jansson examine value-based health care in a one-year MIH pilot in North Ostrobothnia, Finland. The service operated from 10 AM to 10 PM in an urban area and paired a paramedic with a geriatric nurse, equipped for point-of-care testing, venous blood collection, selected treatment initiation, physician consultation, and referral to hospital-at-home. The empirical study used semistructured interviews with 21 of 23 eligible frontline professionals, including nine nurses and 12 paramedics.
What the article contributes
Its strongest contribution is a detailed implementation mechanism. Value is not located in the vehicle or the paramedic alone. It is produced by reconfiguring personnel, information, diagnostics, social care, and follow-up around a home-based episode.
- The paramedic contributes acute assessment and speed.
- The geriatric nurse contributes functional, social, and longitudinal knowledge.
- Remote physicians enable diagnostic and therapeutic reach.
- Shared records and social-care networks reduce handoff loss.
This mechanism map shifts the strategic question from whether to buy a mobile unit to whether the organization can assemble a reliable clinical operating system.
Where the article overreaches
The study’s conclusions exceed its design when they state that MIH reduces emergency interventions and hospitalizations. Interview data can establish that professionals believe a mechanism is valuable and feasible. They cannot establish that utilization, total cost, patient outcomes, or equity changed.
- No patient or caregiver was interviewed.
- No comparison group was used.
- No registry outcome, safety endpoint, resource-use measure, or cost denominator was reported.
The theoretical use of value creation, delivery, and capture is generative but not fully aligned with the canonical value-based definition of patient outcomes achieved per unit of cost across a care cycle.
The distinction that matters most in the United States
In the article, value capture often refers to the perceived value to patients or providers, inferred from staff testimony. That is closer to stakeholder-perceived benefit than to measured value capture. The distinction matters here because the party that creates clinical value may not be the party receiving payment or retaining the avoided cost.
Table 1. Critical appraisal
| Domain | Assessment | Strength | Limitation |
|---|---|---|---|
| Research question | Clear and operationally important | Asks how MIH creates, delivers, and captures value | Mechanism-generating, not effect-estimating |
| Sampling | Near-census of eligible staff: 21 of 23 | Experienced clinicians, with a mean of 16.2 years in health care | Single program; homogeneous provider perspectives |
| Data collection | Semistructured, face-to-face interviews | Rich operational detail and direct quotations | Researcher involvement in program design raises confirmation-bias risk |
| Analysis | Thematic analysis using a value framework | Coherent categories and actionable implementation themes | Supplemental reflexivity detail limits independent appraisal in the main text |
| Outcomes | Perceived patient, provider, and system value | Identifies plausible mechanisms and risks | No patient-reported, clinical, utilization, cost, or safety measurement |
| External validity | Finland’s integrated public system | Strong example of coordinated health and social care | Transferability is limited in fragmented U.S. payer, data, and regulatory environments |
| Causal inference | None was intended by the qualitative design | Appropriate for exploration | Claims of reduced visits, hospitalization, and cost should be hypotheses |
Overall scholarly rating
Strong implementation and theory-building contribution; weak basis for causal, economic, or patient-outcome claims.
The paper should be used to design and test a U.S. model, not to justify a broad scale-up in itself.
Failure modes the article names
These are often absent from promotional accounts, and they are among the most transferable parts of the study:
- Unclear eligibility
- Dementia and alcoholism
- Self-care limitations
- Inadequate home diagnostics
- Excessive workload
- Patient misconceptions
- Ambiguous boundaries among EMS, home health, hospital-at-home, and social services
Section 4
What the U.S. evidence supports
The evidence base has moved beyond anecdotes, but it remains heterogeneous in intervention design, population, comparison method, and outcome definition. The most defensible conclusion is that some MIH configurations reduce acute-care utilization for selected populations, while others do not. The model is not a uniform treatment.
Figure 5. Selected U.S. relative estimates show directionally favorable but nonuniform results. Sources: Lurie et al. (2023); Gingold et al. (2021); O’Connor et al. (2026). Meta-analysis I-squared exceeded 90 percent; all included meta-analysis studies were observational. The figure compares direction, not intervention equivalence.
Why these estimates are displayed together but never pooled
The report is explicit on this point: relative estimates in Figure 5 are shown together to display directional heterogeneity. They are not statistically pooled, because outcomes, populations, designs, and effect measures differ. Two of the five are incidence rate ratios and three are risk ratios; one is itself a pooled estimate carrying more than 90 percent heterogeneity. A combined number across this set would be arithmetically computable and substantively meaningless.
Author-derived view
Precision, which the forest plot shows but does not quantify
Figure 5 prints every point estimate and interval, so rebuilding it would add nothing. What it does not state is how precise each estimate is. On a log scale the natural measure is the ratio of the interval bounds, and it separates these five studies more sharply than their point estimates do.
Log-scale confidence-interval width
Computed as the natural log of the upper bound divided by the lower bound. Larger means less precise. Green estimates favor lower utilization, brass crosses the null, clay favors higher.
Author-derived from the interval bounds printed in Figure 5 and Table 2. This is a precision measure only; it says nothing about bias, which is the larger threat in an observational evidence base.
What the widths reveal
The pooled meta-analytic estimate is the most precise in the set at 0.566, which is exactly what pooling twelve studies is supposed to buy. But its precision is illusory for planning purposes: with heterogeneity above 90 percent, the interval describes the average of very different programs rather than the plausible range for any one of them.
The quasi-experimental estimates most relevant to a frail post-discharge cohort are the least precise, at 1.086 and 1.088. The single estimate that crosses the null, Gingold’s readmission result, is among the tighter ones, which makes its null harder to dismiss as underpowering.
The reading that follows
Precision and transportability point in opposite directions here. The tightest interval comes from the least transportable source, and the most relevant design carries the widest interval. That is the structural reason the report calls MIH a platform with variable effectiveness rather than an intervention with an effect size.
Table 2. Selected MIH evidence relevant to U.S. older adults
| Study | Design | Population | Outcome | Finding | Inference |
|---|---|---|---|---|---|
| Louras et al., 2023 | Systematic review; 15 studies | Older adults | EMS calls, ED transport, later use, safety, satisfaction, cost | Favorable immediate-use and experience signals; downstream outcomes mixed | Heterogeneous, mostly nonrandomized evidence |
| Lurie et al., 2023 | Exploratory meta-analysis: 12 observational studies | Mixed MIH populations | ED visits | Pooled RR 0.56 (0.42-0.74) | I-squared above 90 percent; magnitude not transportable as a single expected effect |
| Gingold et al., 2021 | Propensity-matched observational; n=464 enrolled | Adults after medical discharge | 30-day readmission, ED, charges | Readmission IRR 1.19 (0.89-1.60); no benefit on most outcomes | Strong warning against assuming every MIH program works |
| O’Connor et al., 2023 | Prospective feasibility; 153 visits | Community-dwelling older adults; mean age 81 | 72-hour ED use, hospitalization, safety | 10.4 percent ED within 72 hours; 5 unforeseen visits | Supports triage feasibility; lacks a comparator |
| O’Connor et al., 2026 | Prospective quasi-experiment; n=297 | Frail adults age 65+ after discharge | 30-day rehospitalization and ED use | RR 0.45 for readmission; RR 0.58 for ED | Promising; residual selection and single-center limits |
| O’Connor et al., 2026 | Single-system cost simulation | Acute MIH encounters | Program delivery cost | $550 basic; $1,400 advanced median visit cost | Useful unit-cost inputs; not a cost-effectiveness trial |
The essential null result
In 464 enrolled patients compared with propensity-matched controls, Gingold et al. found MIH was not associated with lower 30-day readmission, ED use, charges, or most secondary outcomes. Observation stays were higher at 30 days. This demonstrates why pre-post program reports can overstate benefit when regression to the mean and selection are uncontrolled.
The feasibility result
O’Connor et al. (2023) evaluated 153 acute MIH visits among community-dwelling older adults with a mean age of 81. Sixteen ED visits followed within 72 hours, of which the MIH team recommended 11; only 5 were unforeseen. This supports feasibility and a safety-oriented escalation pathway, not comparative effectiveness.
The most relevant recent result
Among 297 frail older adults after discharge, PACED participants had lower 30-day rehospitalization (12.6 versus 21.5 percent; adjusted RR 0.45) and lower 30-day ED use without admission (11.5 versus 18.7 percent; adjusted RR 0.58). The design remains nonrandomized; comparator eligibility, geography, affiliation, and refusal can leave residual confounding.
Cost evidence
A 2026 single-system cost analysis estimated median delivery costs of $550 for a basic MIH call and $1,400 for an advanced call. These are operational inputs rather than proof of savings.
Evidence verdict
MIH should be considered a platform with variable effectiveness, not a single intervention with a universal effect size.
The program’s target population, clinical authority, diagnostics, follow-up, and payment contract are part of the treatment. A $550 visit is not valuable merely because it is cheaper than an ED encounter; it must safely replace or prevent a higher-cost episode, or create patient value the payer is willing to purchase.
Interactive model
Claim discipline: what your evidence actually licenses
The report sets an explicit causal-discipline rule in its methods: provider perceptions, associations, and scenario projections are not described as causal effects. That rule is the basis of its critique of the anchor article, and it applies equally to any U.S. program making a case internally. This tool matches the claim you want to make against the study designs you actually hold.
Claim discipline checker
Select the claim, then switch off any design you do not hold. The default state is the evidence base cited in this report.
Verdict
Supported
Association level
The rule the tool applies
Each claim sits at a level, and each level requires a design capable of establishing it. The mapping is not a scoring judgment; it follows directly from what each design can and cannot observe.
| Claim level | Claim | Design that licenses it | Status in this evidence base |
|---|---|---|---|
| Mechanism | MIH is a plausible way to reach older adults at home | Qualitative interviews with frontline clinicians | Available Wang et al. (2026) |
| Feasibility | MIH is feasible and safe in a population like ours | Prospective single-arm feasibility, or a quasi-experiment | Available O’Connor et al. (2023, 2026) |
| Association | MIH is associated with lower acute utilization | Observational comparison of any strength | Available four studies, in both directions |
| Causal | MIH causes lower acute utilization in our cohort | Randomization or an equivalent causal design | Provisional only nonrandomized comparison exists |
| Economic | MIH lowers total cost of care | A causal utilization design plus a full cost denominator | Not supported cost inputs exist, causal design does not |
| Patient outcome | MIH improves patient-valued outcomes | Patient-reported or caregiver-reported measurement | Not supported no cited study measured one |
The last row is the sharpest finding in the report’s appraisal. No cited study interviewed a patient or a caregiver. A program that wants to claim patient value will have to generate that evidence itself, not import it.
Why this matters commercially, not only academically
An overreaching claim is not a scholarly discourtesy. It is a contracting risk.
A program that promises a payer reduced total cost of care on the strength of an association study has priced a result its design cannot deliver. When the null arrives, as it did for Gingold et al., the contract, not the evidence, is what fails.
Escalation is not a failure
The report treats escalation to the emergency department as a safety outcome when home care is inappropriate. In the feasibility study, the MIH team recommended 11 of the 16 ED visits that occurred within 72 hours. Those are correct decisions, not leakage.
Any measurement plan that counts all post-visit ED use against the program will penalize good triage and reward unsafe conservatism. That is why the scorecard separates unforeseen ED use from ED use overall.
Sections 4 and 5
Unit economics and the substitution a contract would require
The report supplies two published cost inputs and one explicit warning. The inputs are median delivery costs of $550 for a basic call and $1,400 for an advanced call. The warning is that a $550 visit is not valuable merely because it is cheaper than an ED encounter: it must safely replace or prevent a higher-cost episode, or create patient value the payer is willing to purchase.
What this model does and does not take from the report
The visit costs are published. The value of a safely avoided episode is not in the report, and the report explicitly declines to estimate the avoidable share of visits. That figure is therefore a user assumption throughout this tool, and every output that depends on it inherits that status. The model’s purpose is to find what a contract would have to be true for, not to assert that it is.
Substitution break-even model
Published visit costs, user-supplied episode value and substitution rate.
Verdict
Clears break-even
Blended visit cost
$805
Weighted by your call mix
Cost per enrolled patient
$1,610
Visits times blended cost times overhead
Total program cost
$805,000
Across 500 patients at 2.0 visits each
Required substitution rate
32.2%
To cover program cost at your episode value
Gross savings at your rate
$875,000
Substitution rate times episode value
Net position
$70,000
Savings less program cost
Break-even episode value
$4,600
Episode value at which your rate exactly pays
Required substitution rate against avoided-episode value
The curve is your cost per enrolled patient divided by the episode value. Above the dashed rule the program would need to avoid more than one episode per patient, which no substitution rate can deliver.
Published inputs: $550 and $1,400 median visit costs, O’Connor et al. (2026). All other inputs are user assumptions.
What the blended cost is made of
Contribution of each call type to the blended visit cost.
Because an advanced call costs 2.55 times a basic call, the mix moves the blended figure faster than the volume does. Shifting the advanced share from 30 to 50 percent raises the blended cost by more than 20 percent.
Section 5
The decisive U.S. difference: who captures the value
In the Finnish case, public funding and integrated governance make it plausible that savings in one part of the system benefit the same regional authority. In the United States, a municipal EMS service may create value for a health plan or hospital without receiving payment. A hospital may avoid readmission penalties but lose fee-for-service volume. A payer may save on an ED visit while the delivery system bears the field cost.
Value capture alignment
The break-even arithmetic above is only decision-relevant when these two are the same organization.
Alignment: Aligned
Why risk-bearing organizations are the credible early purchasers
In 2026, 511 Medicare Shared Savings Program ACOs served 12.6 million Traditional Medicare beneficiaries. CMS also extended Acute Hospital Care at Home through September 30, 2030, and launched a $50 billion Rural Health Transformation Program for fiscal years 2026 to 2030.
These programs do not automatically cover MIH, but they expand the institutional settings in which home-based acute care can be financially viable. The value-based business model requires contractual alignment, not only clinical integration.
Section 5
Transferability from Finland to the United States
The Finnish pilot and the U.S. market share a common clinical problem but not a common institutional substrate. Finland’s centralized, tax-funded regional structure places emergency, social, home, and rescue services within a coordinated public architecture. U.S. MIH must often bridge independent EMS agencies, hospitals, physician groups, plans, post-acute providers, home health, social services, and multiple electronic health records.
| Finnish feature | Transferability | Why it matters | U.S. adaptation |
|---|---|---|---|
| Paramedic plus geriatric expertise | High | Acute assessment combined with function, cognition, and social context | Use a geriatric nurse, advanced practice clinician, care manager, or tele-geriatrics, depending on the cohort |
| Point-of-care diagnostics | High | Reduces diagnostic uncertainty at home | State scope, CLIA, pharmacy, medical direction, quality control, and supply chain must be explicit |
| Integrated health and social records | Low to moderate | Finland’s integrated platform is a major enabling condition | Use HIE, FHIR interfaces, consent, shared documentation, and closed-loop referral; expect gaps |
| Unified public funding | Low | The Finnish system can internalize broad social and clinical value | A U.S. model needs a named purchaser and attribution logic |
| Standardized care pathway | High | Eligibility and escalation reduce safety risk | Standardize locally while complying with state EMS and professional scope rules |
| Fixed service availability | Moderate | The pilot operated 10 AM to 10 PM | Coverage window should match the avoidable-demand pattern; overnight capacity may be expensive |
| Patient self-management at home | Conditional | Home treatment assumes the capacity for monitoring and follow-up | Screen cognition, function, caregiver availability, housing, language, and health literacy |
The pattern in the ratings
Every feature the report rates highly transferable is clinical. Both features it rates least transferable are institutional. Nothing about the Finnish clinical model resists import; what resists import is the substrate that made it financially and informationally coherent. That is why a strong clinical score cannot compensate for a weak institutional one, and why this instrument treats the two low-transferability features as binding constraints rather than as two items among seven.
Interactive model
Score your own setting
Rate each of the seven Finnish features against what your organization can actually operate today. The composite is a simple unweighted mean, deliberately, because the report supplies no weights and inventing them would import a judgment it does not make.
Finland to United States transferability scorer
Seven features, four levels each. Two of them act as binding constraints.
Paramedic plus geriatric expertise
High transferabilityAcute assessment combined with function, cognition, and social context
U.S. adaptation: Use a geriatric nurse, advanced practice clinician, care manager, or tele-geriatrics, depending on the cohort
Point-of-care diagnostics
High transferabilityReduces diagnostic uncertainty at home
U.S. adaptation: State scope, CLIA, pharmacy, medical direction, quality control, and supply chain must be explicit
Integrated health and social records
Low to moderate transferabilityFinland's integrated platform is a major enabling condition
U.S. adaptation: Use HIE, FHIR interfaces, consent, shared documentation, and closed-loop referral; expect gaps
Unified public funding
Low transferabilityThe Finnish system can internalize broad social and clinical value
U.S. adaptation: A U.S. model needs a named purchaser and attribution logic
Standardized care pathway
High transferabilityEligibility and escalation reduce safety risk
U.S. adaptation: Standardize locally while complying with state EMS and professional scope rules
Fixed service availability
Moderate transferabilityThe pilot operated 10 AM to 10 PM
U.S. adaptation: Coverage window should match the avoidable-demand pattern; overnight capacity may be expensive
Patient self-management at home
Conditional transferabilityHome treatment assumes the capacity for monitoring and follow-up
U.S. adaptation: Screen cognition, function, caregiver availability, housing, language, and health literacy
Composite readiness
Absent
0.0 of 100
Binding constraint rule
Readiness by feature
Bar colour shows the report’s transferability rating: green high, pale green moderate or conditional, brass low.
How to read the bands
The four bands are neutral quartiles of the 0 to 100 range, not validated thresholds. The report proposes no transferability index and offers no cutoffs, so these are an arithmetic convenience for triage and should be treated as such. What is grounded in the report is the capping rule: it states that payment and data integration determine whether demand converts into a scalable market, so the two institutional features are allowed to hold the band down regardless of clinical strength.
Section 6
U.S. market outlook, 2026 to 2050
The addressable need for MIH will grow faster than the age-65+ population, because the age-85+ segment grows faster, uses emergency care more intensively, and is more likely to face functional, cognitive, transportation, and caregiver constraints. Market conversion will be slower than need, because reimbursement, scope of practice, workforce, and data integration remain local and fragmented.
Figure 6. Base-case U.S. MIH adoption horizon. Forecast basis: U.S. Census aging projections, CMS ACO and AHCAH policy, Rural Health Transformation funding, ET3 experience, and peer-reviewed MIH outcomes. This is a qualitative adoption forecast, not a revenue forecast.
2026 to 2030
Selective, contract-driven adoption
Near-term growth will concentrate in integrated health systems, Medicare Advantage, ACOs, hospital-at-home programs, Medicaid managed care, and state-funded rural initiatives. Programs will target high-risk post-discharge patients, frequent 911 or ED users, homebound adults, chronic disease exacerbations, falls, medication reconciliation, and urgent diagnostic needs. Broad community coverage without a payer contract will remain difficult to sustain.
2030 to 2040
The acceleration window
This decade is the most likely inflection point. The age-85+ population rises from 8.6 million in 2030 to 13.7 million in 2040, a 60 percent increase, while the HRSA workforce projection reaches its 2038 horizon. Health systems will face a strategic choice: transport more complex older adults into already-constrained facilities, or expand the clinical radius of their workforce through mobile and remote teams.
2040 to 2050
Normalization or geographic inequality
Two outcomes are plausible. In the normalization scenario, MIH becomes a standard component of age-friendly regional care, integrated with hospital-at-home, primary care, home health, pharmacy, and social services. In the inequality scenario, financially aligned urban systems and selected states develop mature networks while rural, fragmented, or fee-for-service markets remain episodic and grant-dependent. Payment reform and interoperability determine which path dominates.
Table 4. Purchaser segments and value-capture logic
| Purchaser or sponsor | Problem purchased | Value capture | Principal risk |
|---|---|---|---|
| 1. Integrated delivery systems | Readmission, ED congestion, bed capacity, patient experience | Avoided acute use, faster discharge, home-based continuity | Single-system bias; volume displacement |
| 2. Medicare Advantage and provider-sponsored plans | Total cost of care, stars, retention, high-risk members | Capitated savings and member experience | Vendor fragmentation, attribution, and encounter capture |
| 3. ACOs and value-based physician groups | Shared savings, high-needs beneficiaries, primary-care capacity | Acute access without new brick-and-mortar sites | Savings timing; beneficiary leakage |
| 4. Medicaid managed care | Access, avoidable ED use, behavioral and social needs | Flexible care management and community partnerships | State variability; rate adequacy |
| 5. Hospital-at-home programs | Field capability, diagnostics, escalation, logistics | Shared mobile workforce and command center | Different eligibility and hospital-level requirements |
| 6. Rural systems and states | Distance, closures, workforce scarcity | Mobile access and regional resource sharing | Travel productivity, sparse volume, broadband, and workforce |
| 7. Municipal EMS agencies | 911 demand, unit availability, nontransport care | Operational relief and public value | Value often accrues to another payer or provider |
Purchaser segments positioned by capture ability and constraint exposure
Numbers correspond to the table rows above. Positions are an author-derived reading of the Table 4 descriptions, not measured values.
Author-derived positioning. The upper right quadrant is where need and incentive coincide; the upper left is where need is high but no mechanism exists to fund it.
The structural problem the plot makes visible
Municipal EMS agencies sit at the top left. They face the constraint most directly, encounter it first through 911, and have the least ability to capture what they save. Rural systems sit nearby for a different reason: high exposure, but thin volume and travel economics that undercut capture. The two segments with the strongest need are the two least able to fund a response on their own.
What follows for a vendor or an operator
Selling to the segment that feels the pain hardest is the intuitive move and the wrong one. The report’s logic points instead to the segments in the lower right, where capture is strong even if exposure is more moderate, and to public funding instruments such as the Rural Health Transformation Program where the upper-left segments are concerned.
Forecast confidence, stated explicitly
The report grades its own market claims rather than presenting them at uniform confidence. The last row is the one most often omitted from commercial analyses.
High confidence
Demand for home-accessible acute assessment will rise as the age-85+ population grows.
Moderate confidence
The fastest adoption will occur in risk-bearing organizations and hospital-at-home-adjacent systems.
Moderate confidence
2030 to 2040 will be the strongest demand acceleration period.
Low confidence
A uniform national payment pathway will emerge before 2030.
Not estimated
National revenue, program count, or percentage of older adults served. Available evidence does not support a defensible dollar total addressable market without strong assumptions about eligibility, visit frequency, price, and substitution.
Primary adoption determinant
Whether the contracting entity can capture downstream emergency, inpatient, and post-acute savings.
Not clinical capability, not technology, and not demonstrated need. All three are necessary; none of them is what decides whether a program is funded a second year.
Section 7
A U.S. operating model built for evaluability
The appropriate executive response is a staged operating model that makes eligibility, clinical authority, escalation, value capture, and evaluation explicit before scale. The design below is compatible with both the Finnish mechanism and the strongest U.S. evidence.
Figure 7. A U.S. MIH operating model designed for clinical reliability and causal evaluation. Framework synthesized from Wang et al. (2026), O’Connor et al. (2023, 2024, 2026), and CMS payment and quality requirements.
Stage 1
Identify
- High-risk discharge
- Frequent 911 or ED use
- Acute change at home
Stage 2
Triage
- Clinical eligibility
- Frailty and cognition
- Home safety and consent
Stage 3
Treat
- Paramedic or nurse
- Point-of-care tests
- Remote physician
Stage 4
Dispose
- Remain at home
- Alternative destination
- ED escalation
Stage 5
Learn
- 72-hour safety
- 30-day utilization
- Cost and experience
Clinical governance
Protocols, scope, and medical direction. Define diagnostics, medications, documentation, consent, and escalation before launch, not after the first ambiguous case.
Information governance
Electronic health record, health information exchange, consent, and closed-loop handoff. Every encounter should route findings to the accountable primary or specialty team and verify that follow-up is complete.
Financial governance
Contract, attribution, and shared savings. Identify which payer, ACO, hospital, or public authority funds the program and which downstream costs it can capture.
Safety gate
Escalation is a correct outcome when the home is not the appropriate site of care.
This is not a softening of the standard. It is what makes the standard measurable: a program that counts every post-visit emergency department encounter as a failure will optimize toward unsafe conservatism, which is why the scorecard measures unforeseen ED use rather than ED use.
The five executive recommendations
Stated in the report as a sequence. Each one is a precondition for the next, which is why the investable-unit screener in the following section treats them as gates rather than as a checklist to be scored.
1. Choose one cohort and one failure mechanism
Examples include frail adults within 72 hours of discharge, frequent 911 callers whose needs are primary-care treatable, or homebound adults with acute symptom change.
2. Name the economic principle
Identify which payer, ACO, hospital, or public authority funds the program and which downstream costs it can capture.
3. Specify clinical authority
Define medical direction, scope, diagnostics, medications, documentation, consent, and escalation before launch.
4. Build a closed loop
Every encounter should route findings to the accountable primary or specialty team and verify that follow-up is complete.
5. Measure safety before savings
Track unplanned ED use within 24 to 72 hours, mortality, delayed escalation, medication harm, falls, and complaints.
Then, and only then
Use a prospective comparison chosen before expansion, and scale only after heterogeneity analysis has determined which diagnoses, frailty levels, geographies, caregiver contexts, and visit types produce benefit or harm.
Research design for the next generation of evidence
A hybrid effectiveness-implementation design
The clinical study should prespecify primary safety and utilization outcomes, measure patient and caregiver value, collect all-payer claims or health information exchange encounters, and use intention-to-treat logic when feasible.
The implementation study should measure reach, adoption, fidelity, staffing burden, referral loss, and context.
Economic evaluation should report the incremental cost per eligible patient and per safely avoided acute episode, with sensitivity to travel radius, staffing mix, and substitution rate.
When randomization is not feasible
Randomization is preferable when operationally feasible. When it is not, the report names four alternatives: stepped-wedge rollout, regression discontinuity around eligibility thresholds, difference-in-differences across comparable geographies, or an instrumental variable based on dispatch availability.
One construction to avoid. Comparator groups should not rely solely on decliners or on people outside the service area, because the same factors that produced the decline or the location may also affect the outcome.
Interactive model
Is this an investable unit?
The report’s decision implication is stated conjunctively: the investable unit is a defined cohort in a defined geography with a closed-loop clinical pathway and a payer or provider that captures downstream savings, with expansion conditional on prespecified outcomes. This tool honours that structure. It reports no composite score, because a partial pass is not a smaller version of the same thing.
Investable unit screener
Five conditions to pilot, three further conditions to scale. All conjunctive.
Conditions to define an investable pilot
One cohort and one failure mechanism named
Frail adults within 72 hours of discharge, frequent 911 callers whose needs are primary-care treatable, or homebound adults with acute symptom change. Not "older adults" in general.
One geography with a stated travel radius and coverage window
The Finnish pilot ran 10 AM to 10 PM in an urban area. Coverage should match the avoidable-demand pattern; travel time can erase unit economics.
A closed-loop clinical pathway with named clinical authority
Medical direction, scope, diagnostics, medications, documentation, consent, and escalation defined before launch, with verified follow-up on every encounter.
A named purchaser that captures the downstream saving
Identify which payer, ACO, hospital, or public authority funds the program and which downstream costs it can capture. Value creation without value capture does not scale.
Prespecified safety, utilization, experience, equity, and total-cost outcomes
Expansion should be conditional on outcomes chosen before the program starts, not selected afterwards from what moved.
Further conditions before expansion
Safety measured before savings
Unplanned ED use within 24 to 72 hours, mortality, delayed escalation, medication harm, falls, and complaints tracked first.
A prospective comparison chosen before expansion
Geography, propensity, stepped implementation, regression discontinuity, or randomization, selected in advance.
Heterogeneity analysed before scaling
Which diagnoses, frailty levels, geographies, caregiver contexts, and visit types produce benefit or harm.
Screening result
Not yet an investable unit
0 of 5 pilot conditions met
Why gates rather than a score
What a score would conceal
A weighted composite would let a program with an excellent cohort definition, a clean geography, a rigorous pathway, and prespecified outcomes score 80 out of 100 with no purchaser identified. That reads as nearly ready. It is not nearly ready. It is a clinical demonstration with no second year.
The same is true in reverse. A well-funded contract attached to an undefined cohort produces a program that cannot be evaluated, which is precisely the position the report says the anchor article leaves the field in.
The separation between piloting and scaling
The five pilot conditions and the three expansion conditions are deliberately kept apart. A program can be entirely defensible as a pilot while remaining unfit to scale, and treating the two decisions as one is how heterogeneity gets ignored.
The evidence base supports this separation directly. Programs that looked strong in single-arm feasibility work did not all survive controlled comparison; the propensity-matched null is the clearest instance.
The correct strategic posture
Disciplined expansion: target narrowly, integrate deeply, measure causally, and scale only when safety and total value are demonstrated.
The strongest U.S. market will emerge where risk-bearing organizations can capture downstream value and where programs are designed as measurable care pathways rather than as mobile units.
Section 7, Table 5
The minimum executive scorecard
Nine domains, each with a defined key performance indicator, a direction of travel, and a guardrail. The guardrails are the most useful column and the most frequently dropped: each one names a specific way the metric can be gamed or misread.
| Domain | KPI | Definition | Direction | Guardrail |
|---|---|---|---|---|
| Clinical safety | Unforeseen ED use within 72 hours | ED visit not recommended by MIH | Lower without delayed necessary care | Primary safety gate |
| Acute utilization | 30-day ED and hospitalization | All-payer encounters per eligible person | Lower versus comparable baseline | Capture out-of-network use |
| Transitions | 30-day unplanned rehospitalization | Risk-adjusted readmission after index discharge | Lower in the defined cohort | Align with the discharge pathway |
| Patient value | Experience and goal concordance | Validated PREM plus stay-at-home preference | Improvement without coercion | Include hearing, cognition, and language |
| Caregiver value | Burden and confidence | Brief caregiver-reported measure | Reduced burden or improved confidence | Caregiver absence is also data |
| Operational | Response time and visit productivity | Referral-to-arrival; visits per staffed hour | Stable within geographic target | Travel time can erase unit economics |
| Economic | Incremental total cost of care | Program cost plus acute and post-acute spend | Net savings or justified value | Do not compare the visit price alone |
| Equity | Reach and outcomes by subgroup | Rurality, race and ethnicity, dual status, language, disability | No avoidable access or outcome gap | Do not let digital eligibility exclude need |
| Workforce | Staff burden and retention | Overtime, missed breaks, turnover intent, and safety events | Sustainable workload | Task-shifting can shift burnout |
Interactive model
Configure your measurement position
Mark where each domain actually stands. Defined means the metric has an agreed definition; instrumented means data is being captured; reporting means it reaches a decision-making forum on a schedule.
Executive scorecard configurator
Clinical safety operates as a gate, per the report’s own sequence.
Clinical safety
Safety gateUnforeseen ED use within 72 hours
ED visit not recommended by MIH
Direction: Lower without delayed necessary care Guardrail: Primary safety gate
Acute utilization
30-day ED and hospitalization
All-payer encounters per eligible person
Direction: Lower versus comparable baseline Guardrail: Capture out-of-network use
Transitions
30-day unplanned rehospitalization
Risk-adjusted readmission after index discharge
Direction: Lower in the defined cohort Guardrail: Align with the discharge pathway
Patient value
Experience and goal concordance
Validated PREM plus stay-at-home preference
Direction: Improvement without coercion Guardrail: Include hearing, cognition, and language
Caregiver value
Burden and confidence
Brief caregiver-reported measure
Direction: Reduced burden or improved confidence Guardrail: Caregiver absence is also data
Operational
Response time and visit productivity
Referral-to-arrival; visits per staffed hour
Direction: Stable within geographic target Guardrail: Travel time can erase unit economics
Economic
Incremental total cost of care
Program cost plus acute and post-acute spend
Direction: Net savings or justified value Guardrail: Do not compare the visit price alone
Equity
Reach and outcomes by subgroup
Rurality, race and ethnicity, dual status, language, disability
Direction: No avoidable access or outcome gap Guardrail: Do not let digital eligibility exclude need
Workforce
Staff burden and retention
Overtime, missed breaks, turnover intent, and safety events
Direction: Sustainable workload Guardrail: Task-shifting can shift burnout
Measurement maturity
Not defined
0.0 of 100
Safety gate: Closed
Maturity by domain
The brass bar is the gating domain.
On the bands
As with the transferability scorer, the four bands are neutral quartiles rather than validated thresholds; the report proposes no maturity index. What is taken directly from the report is the gate: it lists clinical safety as the primary safety gate and instructs that safety be measured before savings. Until unforeseen ED use within 72 hours is instrumented, a high composite would describe reporting capability the program cannot yet act on safely.
The two domains most often missing
Caregiver value. No cited study measured it, and the guardrail is unusually sharp: caregiver absence is itself data. A program serving a cohort with no available caregiver is doing something structurally different from one serving a supported cohort, and a scorecard that only records burden among caregivers who exist will miss that entirely.
Equity. The guardrail warns against letting digital eligibility exclude need. Screening criteria that assume broadband, a smartphone, or a responsive caregiver will systematically exclude the patients whose access constraint is most binding, which inverts the purpose of the program.
Why economic sits last, not first
The economic guardrail is the one most directly aimed at the failure mode this report keeps returning to: do not compare the visit price alone. Incremental total cost of care means program cost plus acute and post-acute spend, not the difference between a field visit and an emergency department bill.
A scorecard that reports the economic domain before the safety and utilization domains are instrumented is reporting a number whose denominator it does not yet have.
Section 1 and Appendix B
Methods and evidentiary boundaries
This report uses a targeted critical review of the anchor article, peer-reviewed MIH systematic reviews and controlled observational studies, U.S. Census population projections, CDC emergency department utilization data, and current CMS and HRSA policy sources. The analysis is narrative and scenario-based, not a systematic review, a causal meta-analysis, or a revenue forecast.
Scope and sources
- Scope. United States market, adults aged 65 and older, with special attention to the age-85+ population and community-dwelling frail adults.
- Evidence date. Sources available through August 10, 2026.
- Primary quantitative sources. U.S. Census Bureau 2023 National Population Projections and 2024 population estimates; CDC/NCHS 2022 NHAMCS; CMS program and payment materials; HRSA workforce projections.
Model and causal discipline
- Model. Projected ED exposure equals projected age-65+ population multiplied by the 2022 observed age-65+ ED visit rate. No utilization trend, morbidity change, policy effect, or MIH effect is imposed.
- Causal discipline. Provider perceptions, associations, and scenario projections are not described as causal effects. Escalation to the ED is treated as a safety outcome when home care is inappropriate.
Reporting and reproducibility notes, Appendix B
- Observed values, government projections, modeled scenarios, and study estimates are visually and verbally distinguished.
- All figures were generated from cited source values; no synthetic patient-level data were used.
- Relative estimates in Figure 5 are displayed together to show directional heterogeneity. They are not statistically pooled, because outcomes, populations, designs, and effect measures differ.
- The report uses ASCII hyphens and avoids causal language where the study design does not support it.
Verification ledger
Every quantitative claim in the report was independently recomputed from the published inputs before this dashboard was built. All of them reproduce. Two rounding artifacts are disclosed here rather than silently corrected.
| Claim as published | Published | Recomputed | Status |
|---|---|---|---|
| Age-65+ growth, 2022 to 2050 | +42% | +42.1% | Reproduces |
| Age-85+ growth, 2022 to 2050 | +168% | +167.7% | Reproduces |
| Oldest-old share, 2022 | 11.2% | 11.25% | Reproduces |
| Oldest-old share, 2050 | 21.2% | 21.19% | Reproduces |
| Age-85+ growth, 2030 to 2040 | 60% | 59.3% | Reproduces |
| Age-85+ grows about four times faster than age-65+ | 4x | 3.98x | Reproduces |
| Implied 2022 ED visit rate, age 65+ | 0.569 | 0.569253 | Rounding disclosed |
| Modeled ED visits, 2030 | 40.52M | 40.521M | Reproduces |
| Modeled ED visits, 2040 | 44.57M | 44.569M | Reproduces |
| Modeled ED visits, 2050 | 46.75M | 46.753M | Reproduces |
| Percent change in ED exposure, 2050 | +42.1% | +42.11% | Reproduces |
| Support ratio decline, 2022 to 2050 | 61% | 61.02% | Reproduces |
| ED visits within 72 hours, feasibility study | 10.4% | 10.46% | Reproduces (16 of 153) |
| Primary care share of the projected physician shortage | not stated | 50.0% | Derived (70,610 of 141,160) |
Rounding artifact 1: the visit rate
Appendix Table A1 prints the rate as 0.569 visits per person-year, but its ED visits column is computed from the unrounded quotient of 32.90 divided by 57.795, which is 0.569253. Recomputing the column from the printed 0.569 would give 40.50, 44.55, and 46.73 rather than the published 40.52, 44.57, and 46.75. This dashboard uses the unrounded quotient so it reproduces the published column exactly. The difference is immaterial to any conclusion but is disclosed because the two are visibly different at two decimals.
Rounding artifact 2: the fourfold claim
The statement that the age-85+ population grows about four times faster than the age-65+ population recomputes to 3.98 using the report’s own figures (167.7 percent divided by 42.1 percent). The word “about” carries this correctly. It is noted here only so that a reader recomputing the ratio does not conclude the figures disagree.
Limitations of the report
On the evidence
This is a targeted critical review, not a registered systematic review. It emphasizes studies most directly relevant to older adults, acute and transitional MIH, and current U.S. market conditions. Publication bias and rapidly evolving state policy may affect the evidence base.
The demographic exposure model assumes a constant 2022 older-adult ED visit rate and does not model age subgroups, morbidity, technological change, prevention, hospital capacity, payer behavior, or substitution.
On the market outlook
The market outlook is qualitative because national program counts, standardized prices, eligible population definitions, and durable reimbursement pathways are not sufficiently consistent to support a reliable national revenue forecast. Cost inputs from a single 2026 program are not generalized as national prices.
The report does not evaluate state-by-state scope of practice, licensure, Medicaid benefits, or commercial plan coverage. Those are necessary for an investment decision in a named market.
What this dashboard adds, and how it is labelled
Everything drawn from the report is presented as published. Four things are author-derived and are marked as such wherever they appear.
Composition calibration
The 36.4 percent 75+ share that reconciles the report’s two age-specific rates with its overall rate, used to isolate the compositional effect the report describes but does not size.
Interval precision
Log-scale confidence-interval widths computed from the bounds printed in Figure 5, showing that the most precise estimate is the least transportable.
Indexed growth views
Rebasing the Figure 1 and Figure 3 series to 2022 = 100, which reveals that the support-ratio decline is driven almost entirely by its denominator.
Gating rules
The binding-constraint rule in the transferability scorer and the safety gate in the scorecard follow the report’s stated logic; the numeric band cutoffs are neutral quartiles and are not validated thresholds.
Appendix A. Projection calculations
| Year | Age-65+ population (M) | Visits per person-year | ED visits (M) | Status |
|---|---|---|---|---|
| 2022 | 57.795 | 0.569 | 32.90 | Observed |
| 2030 | 71.183 | 0.569 | 40.52 | Modeled (+23.2%) |
| 2040 | 78.294 | 0.569 | 44.57 | Modeled (+35.5%) |
| 2050 | 82.130 | 0.569 | 46.75 | Modeled (+42.1%) |
Calculation: 32.9 million ED visits divided by 57.795 million adults age 65+ in 2022 equals 0.569 visits per person-year. That constant rate is multiplied by the projected population in each year. No attempt is made to estimate the share of visits that is avoidable or MIH-eligible.
References
Sources cited in the report
Twenty-two sources, filterable by the role each plays in the argument. The anchor article is the study under critical review; the evidence group is the U.S. outcomes literature that tests its claims.
22 sources
Reading the evidence group together
Six of the twenty-two sources carry the outcomes argument, and they do not agree. Lurie and Louras summarize heterogeneous observational work; Gingold supplies the null; the two 2026 O’Connor papers supply the most relevant recent comparison and the cost inputs; the 2023 O’Connor paper supplies feasibility.
The report’s position is that this disagreement is informative rather than unresolved. It is what a platform with variable effectiveness looks like when studied across different configurations.
Where the policy sources matter
The four CMS sources are not background. The ET3 final evaluation documents a payment alternative that existed, delivered very low intervention volume, and ended in 2023, which is the single most important cautionary datum for anyone assuming a national pathway will appear.
The ACO participation, Acute Hospital Care at Home extension, and Rural Health Transformation Program sources define the institutional settings in which home-based acute care can currently be financially viable.
When Acute Care Cannot Be Reached: Mobile Integrated Health as a Value-Based Response to the Aging Constraint in the United States.
Prepared by Kelly Emrick, DHSc, PhD, MBA, BSRT(ARRT)R. August 2026. Evidence current through August 10, 2026.
All published figures, tables, and values are drawn from the report and its cited sources. Author-derived views, calibrations, and index bands are labelled as such at the point of use. The demographic exposure model is a scenario, not an official forecast, and does not estimate avoidable or MIH-eligible visits.