01 PERSONAL SITE
/
← RESEARCHER

Interscalar medicine / long-form research note / January 2025

The Economics of Achieving a Functional Outcome

The price of treatment answers only one question: how much was spent. It does not answer what the person became able to do, whether that ability lasted, or which part of the change was caused by care. For rehabilitation, long-term care, assistive technology and many chronic conditions, those omitted questions are often the decision itself.

A useful additional measure is therefore the incremental cost of achieving a prespecified functional outcome. Examples include the cost of enabling one additional person to walk a defined distance, live at home without personal assistance, use an affected hand in daily tasks, return to a chosen social role, or maintain current function for another year. The adjective incremental is essential: an intervention must be compared with a real alternative, not evaluated in isolation.

This measure is intuitive, but it is not a new universal currency of health. It should sit beside—not automatically replace—clinical outcomes, safety, quality-adjusted life-years (QALYs), budget impact and equity analysis.

A July 2026 WHO/World Rehabilitation Alliance policy brief calls for functioning to be recognized as a third dimension of population health alongside mortality and morbidity, on the ground that what measurement systems omit tends to remain invisible in financing and accountability (WHO, Rehabilitation and Functioning in the 2030 Agenda, 2026). This is a policy and advocacy proposition, not empirical proof that one functional indicator is sufficient. Its useful implication is narrower: function should be visible in the decision set, with definitions and limitations open to inspection.

What counts as function?

The World Health Organization’s International Classification of Functioning, Disability and Health (ICF) treats functioning as more than the absence of disease. Its framework includes body functions and structures, activities, participation and environmental factors. A stronger leg is an impairment-level change; crossing a room is an activity; returning to work, school or community life is participation. The same impairment can produce different real-world function depending on assistive products, housing, transport, caregiver support and social barriers. The ICF was endorsed by the World Health Assembly in 2001 as the international standard for describing and measuring health and disability (WHO ICF; WHA 54.21).

Accordingly, a “functional outcome” must be stated as an observable or reportable change in a defined domain and context. At least five forms are economically relevant:

  1. Attainment: the person reaches a state not present at baseline—for example, independent transfer from bed to chair.
  2. Meaningful improvement: a validated score improves by a prespecified amount.
  3. Maintenance: expected deterioration is delayed or avoided, which is a genuine outcome in progressive or relapsing disease.
  4. Time to loss or recovery: the analysis values when function is lost or regained, not only status at a single visit.
  5. Sustained function: the person remains in the target state for a defined duration, such as days alive and at home without personal assistance.

Diagnosis alone cannot specify any of these. The WHO Disability Assessment Schedule 2.0 (WHODAS 2.0), for example, measures cognition, mobility, self-care, getting along, life activities and participation and is explicitly linked to the ICF (WHO manual, ISBN 978-92-4-154759-8). Disease-specific instruments, performance tests and patient-reported measures may be more sensitive in a particular population. PROMIS Physical Function illustrates an item-response-theory approach intended to improve precision and cross-condition standardization (Fries et al., 2014; DOI 10.3899/jrheum.130813; PMID 24241485). No instrument is automatically fit for every use.

Even use of the ICF does not guarantee complete outcome coverage. A 2026 systematic review of post-acute rehabilitation in older people found that body functions and activities/participation were the components most consistently associated with functional independence and health-related quality of life; environmental factors were studied less often, and personal factors are not formally coded in the ICF. Only seven studies with 896 participants met the review criteria, so the finding supports a structured assessment while also exposing a limited evidence base (Brouwer et al., 2026; DOI 10.1007/s41999-026-01406-0; PMID 41563587).

Before costs are calculated, the study needs an outcome contract:

  • the domain and the reason it matters to patients;
  • instrument, version, language, reporter and assessment mode;
  • baseline state and eligibility criteria;
  • threshold for success or meaningful change;
  • assessment time and required duration of maintenance;
  • handling of death, complications, rescue treatment and missing observations;
  • relevant environmental conditions and assistive products;
  • comparator and target population.

This follows the logic of the FDA’s 2025 patient-focused guidance: define the meaningful aspect of health, concept of interest and context of use, then select a fit-for-purpose clinical outcome assessment (FDA Guidance 3, docket FDA-2022-D-1385). The regulatory guidance is not an economic-evaluation standard, but its measurement discipline is directly useful.

The basic calculation

Let I be the intervention and C the comparator. For each group, estimate mean total cost over the same time horizon and the probability of the prespecified functional outcome:

ΔC = mean cost(I) − mean cost(C)

ΔR = probability of functional outcome(I) − probability of functional outcome(C)

Incremental cost per additional functional responder = ΔC / ΔR

If mean cost is expressed per patient and ΔR > 0, this is also ΔC × NNT, where NNT = 1 / ΔR. Uncertainty must be reported for costs, effects and their joint distribution; a single ratio without an interval, cost-effectiveness plane or acceptability analysis creates false precision.

An illustrative—not empirical—example: a program costs $12,000 per participant versus $10,500 for usual care. At 90 days, 55% versus 45% of participants meet a predeclared independence criterion. The incremental cost is $1,500 and the incremental response is 0.10, giving $15,000 per additional person independent at 90 days. This does not establish value by itself. Decision-makers still need to decide what they are willing to give up for that outcome, examine uncertainty, and ask whether independence persists.

The commonly reported program cost / number of successful patients is an average accounting ratio, not an incremental cost-effectiveness ratio. It attributes successes that would also have occurred under usual care to the program and can therefore misrank alternatives.

Better denominators than a one-time “success”

A binary responder endpoint is easy to explain but discards information. Thresholds can move with context, baseline severity and measurement error. A systematic survey found major credibility limitations in many published minimal important difference estimates, including weak or unreported correlations with their anchors and imprecision (Devji et al., 2021; DOI 10.1016/j.jclinepi.2020.11.024; PMID 33321175). Minimal detectable change for an individual is also not the same as a clinically important mean difference between groups. These concepts should not be substituted for one another.

The information loss is measurable. An empirical study of 21,435 randomized trials in Cochrane reviews estimated that, when a continuous outcome was dichotomized, only about 60% of its information was retained on average. The authors recommend “model continuously but interpret dichotomously”—for example, derive the proportion with clinically meaningful improvement from a continuous model rather than throw away the underlying score (van Zwet et al., 2026; DOI 10.1002/sim.70402; PMID 41641823). The study was not rehabilitation-specific, so 60% is not a universal correction factor; it is evidence against making a threshold the sole analytic endpoint.

For that reason, a functional-economic analysis should usually retain the continuous score and add the responder result for interpretation. Other useful denominators include:

  • cost per point of validated functional gain, accompanied by an explanation of what a point means;
  • cost per additional function-day, calculated from the area under the curve of days alive in the target state;
  • cost per year of maintained function in a progressive condition;
  • cost per day at home without personal assistance;
  • cost per additional participant achieving a patient-prioritized goal, if goal attainment is measured consistently and independently;
  • a cost-consequences table that leaves mobility, self-care, participation, adverse events, caregiver time and costs disaggregated.

The function-day approach often has the clearest interpretation. It distinguishes a transient discharge score from a state maintained for most of a year, handles different recovery times, and can expose relapse or readmission. It still requires explicit rules: death cannot simply become missing data, and “alive but below threshold” must not be confused with “unobserved.”

Which costs belong in the numerator?

The answer depends on perspective. At minimum, label the currency and price year, the payer, the costing method, the time horizon and discounting. The Second Panel on Cost-Effectiveness in Health and Medicine recommends reference cases from both the health-care-sector and societal perspectives and an impact inventory showing what falls inside and outside the analysis (Sanders et al., 2016; DOI 10.1001/jama.2016.12195; PMID 27623463).

A functional-outcome analysis may need to include:

  • assessment, clinician and therapist time;
  • inpatient, outpatient, home and community services;
  • devices, software, maintenance, consumables, connectivity and training;
  • medication, procedures and treatment of adverse events;
  • transport and patient time;
  • home modifications and assistive products;
  • formal social care and unpaid caregiver time;
  • readmissions and downstream health-care use;
  • education, employment and productivity effects, when the societal perspective is relevant.

Consistent classification improves comparability but does not remove judgment. WHO’s 2026 technical note for the System of Health Accounts clarifies the boundary of rehabilitative care under health-care function HC.2 and how to disaggregate interventions and providers (WHO, 2026; ISBN 978-92-4-011878-2). ISO 9999:2022 provides a classification and terminology for assistive products intended to optimize functioning and reduce disability (ISO 9999:2022). Its scope explicitly excludes, among other things, medicines, implanted devices, personal assistance and financial support; those resources must therefore remain visible in separate cost categories. National health-account codes and product classifications help define quantities, but they do not replace patient-level microcosting or the chosen analytic perspective.

Cost offsets are not free money. Reduced care hours or shorter stays become real savings only if resources can be redeployed or budgets actually change. Productivity gains should not be used to value non-working people less; they are one sectoral consequence, not a measure of a person’s worth. Caregiver burden and patient time must be reported transparently, especially when a “cheaper” home program transfers work from paid staff to families.

Cost-effectiveness and affordability are separate questions. A program can be cost-effective yet unaffordable at scale, or cost-saving per participant but require a large implementation budget. Budget-impact analysis should therefore accompany, not be inferred from, the cost-per-function result (Sullivan et al., 2014; DOI 10.1016/j.jval.2013.08.2291; PMID 24438712). WHO’s rehabilitation-financing guidance likewise emphasizes revenue raising, pooling, purchasing, access and financial protection, not just unit cost (WHO, 2024; ISBN 978-92-4-008182-6).

There is also a striking gap between the precision of patient-level economic models and the evidence available for system-level financing rules. A 2026 Cochrane overview found only one directly relevant rehabilitation study within the one review of acceptable methodological quality; the resulting evidence on insurance and access was low certainty and restricted to one population and country. It could not identify which financial arrangements best support effective and equitable rehabilitation (Gimigliano et al., 2026; DOI 10.1002/14651858.CD016327; PMID 41700590). Therefore, outcome-based payment, bundled payment or insurance expansion should not be presented as evidence-established consequences of the cost-per-function framework. They are policy hypotheses requiring prospective evaluation.

Relationship to QALYs and value-based care

Porter’s influential formulation defines value as health outcomes achieved per dollar spent (Porter, 2010; DOI 10.1056/NEJMp1011024; PMID 21142528). Cost per functional outcome gives that idea a concrete rehabilitation denominator. It can reveal benefits that a broad utility instrument compresses or misses, and it is often easier for patients and service managers to interpret.

But disease- or task-specific outcomes do not support comparisons between restoring hand use, preventing blindness and extending survival. QALYs supply a common unit by combining longevity and health-related quality of life. NICE’s reference case uses QALYs for cost-utility analysis and asks for a time horizon long enough to capture important differences in costs and outcomes (NICE health technology evaluation manual, economic evaluation). Where impacts are not adequately captured in costs or QALYs, NICE’s HealthTech methods say they should be highlighted and, when possible, quantified as relevant non-QALY outcomes (NICE HealthTech programme manual).

The defensible position is therefore plural:

  • use cost per functional outcome for a concrete clinical or service question;
  • use cost-utility analysis when cross-program allocation requires a common health unit;
  • show a cost-consequences table for outcomes that should not be collapsed;
  • report budget impact for affordability;
  • add an equity analysis rather than assuming average efficiency is equitable.

There is a genuine ethical dispute about QALYs and disability. Some authors argue that certain utility conventions can undervalue life extension for disabled people (Schneider, 2022; DOI 10.1007/s11136-021-03052-4; PMID 34882282); a 2026 theoretical and empirical analysis argues that QALY-based analyses do not appear inherently discriminatory in the scenarios and disease examples it examined (Ma et al.; DOI 10.1016/j.jval.2026.04.005; PMID 42061674). The dispute is not resolved by replacing QALYs with a single independence threshold: that can itself encode ableist assumptions. In the United States, HHS’s Section 504 rule specifically addresses discriminatory use of value-assessment methods in federally funded programs (HHS Office for Civil Rights, Section 504). Legal requirements differ by jurisdiction and require local review.

From events to an interscalar model

Here, interscalar model should be understood as a proposed analytical architecture, not as an established named standard. It joins events at several levels without pretending they are interchangeable:

  • person: baseline state, goals, intervention dose, functional assessments, adverse events and lived context;
  • care pathway: staff time, waiting, missed sessions, transitions, equipment and readmissions;
  • household and community: unpaid care, transport, home environment, participation and access barriers;
  • provider and payer: capacity, unit costs, contracts and budget consequences;
  • population: reach, distribution of outcomes, unmet need and opportunity cost.

Every record needs time, provenance and a common person/pathway identifier. The model should preserve the chain context → intervention → intermediate change → functional state → resource consequence, while marking which links are observed and which are modeled. An event sequence is a data structure, not proof of causality.

This distinction matters because natural recovery, regression to the mean, case selection and baseline severity can make a service look effective even when it is not. Randomized comparisons remain preferable where feasible. When only observational data are available, a target-trial approach forces the analyst to state eligibility, treatment strategies, time zero, follow-up, outcomes and the causal contrast before analysis (Hernán et al., 2025; DOI 10.7326/ANNALS-24-01871; PMID 39961105). Risk adjustment may improve comparability, but it cannot repair unmeasured confounding.

Practical scenarios

1. Post-stroke transition home

Define the outcome as days alive at home during 180 days, plus independence in selected activities at day 90. Count hospital, community rehabilitation, readmissions, equipment, social care and caregiver time. Compare an early-supported-discharge pathway with current practice. The Cochrane review of coordinated early supported discharge found a lower combined risk of death or institutional care and small improvements in extended activities of daily living, while evidence for some other outcomes and longer-term benefits was less clear (Langhorne & Baylan, 2017; DOI 10.1002/14651858.CD000443.pub4; PMID 28703869). This is a reason to model multiple outcomes and uncertainty, not to assume that earlier discharge is automatically better or cheaper.

A 2026 review of Taiwan’s nationwide post-acute stroke program adds a system-scale example: across 33 mostly observational and quasi-experimental studies, the program was associated with better functional scores, lower short-term readmission and lower acute-care costs than traditional care. However, heterogeneity in Barthel Index gains was high, certainty for functional gain was rated low, and improvement was attenuated in older and more severely disabled groups (Huang et al., 2026; DOI 10.1016/j.apmr.2026.03.014; PMID 42150726). The result is promising evidence within one single-payer design, not a transferable price or causal guarantee.

2. Complex inpatient neurorehabilitation

For a person with high dependency, the relevant gain may be fewer hours of assistance rather than crossing an arbitrary “independent” threshold. Link a validated dependency measure to estimated ongoing care needs, then report functional change, care-cost change and the modeled time for rehabilitation cost to be offset. A prospective UK cohort used this approach in hyperacute specialist rehabilitation (Turner-Stokes et al., 2016; DOI 10.1136/bmjopen-2016-012112; PMID 27609852). Because it was not randomized, its estimates are evidence of association and service-level feasibility, not a causal price tag transferable to every system.

3. Maintaining function in progressive disease

Define success as remaining above a patient-relevant functional state for 12 months, not improvement from baseline. Report time to confirmed decline, function-days, adverse effects and caregiver consequences. Use a comparator that represents the actual alternative. Averted decline must be estimated from concurrent evidence; subtracting this year’s score from last year’s score without a control does not reveal what treatment prevented.

4. Telerehabilitation or robotic rehabilitation

Compare equivalent clinical pathways, not a device with “nothing.” Include acquisition, depreciation, maintenance, licensing, cybersecurity, connectivity, technical support, therapist review, replacement and patient space/time. Measure completion, functional outcome, safety and differential access by age, income, language, geography and digital literacy. A systematic review of robotic and virtual-reality rehabilitation found heterogeneous populations, settings, protocols and economic outcomes, with overall moderate methodological quality (Cano-de-la-Cuerda et al., 2024; DOI 10.3390/jcm13061531; PMID 38541755). Lower travel cost alone does not establish cost-effectiveness.

A newer stroke-specific review reinforces that warning. Among 15 telerehabilitation studies of upper-limb training, only two performed cost-effectiveness analyses; cost evidence was mixed, EQ-5D reporting was inconsistent, and studies often omitted maintenance, depreciation and the need for an internet-enabled device. Reduced therapist time could offset technology costs, but longer-term economic value remains uncertain (Syukriyah et al., 2026; DOI 10.1177/02692155261441561; PMID 42068268).

5. Return to work or education

Measure participation in the role chosen by the person, its sustainability and accommodations, alongside symptoms and function. Report productivity and education consequences under a societal perspective but do not make employment the sole success criterion. Retired people, children, unpaid carers and people unable or unwilling to enter paid work must not become invisible in the outcome definition.

Risks and failure modes

  1. Threshold gaming: providers can concentrate effort on people just below the cut-off. Safeguard by retaining continuous change and auditing score distributions.
  2. Cream-skimming: payment per success may discourage admission of people with severe or complex disability. Stratify by baseline risk, publish outcomes by severity and protect access independently of predicted gain.
  3. Ceiling and floor effects: an instrument may be unable to register change at the extremes. Test measurement properties in the intended population.
  4. Short horizons: a discharge gain can disappear after relapse, while preventive value may emerge years later. Use a horizon that captures meaningful differences and show shorter operational horizons separately.
  5. Missing-not-at-random data: people with deterioration or access barriers may be least likely to complete follow-up. Predefine collection and sensitivity analyses; never silently treat missing as success or failure.
  6. Double counting: caregiver time, social-care savings, utility gains and monetized productivity can overlap. Maintain an impact inventory and make aggregation rules explicit.
  7. Ignoring harms and mortality: better function among survivors can coexist with worse survival or complications. Report safety and survival separately and in composites where justified.
  8. False precision and transferability: prices, care patterns and willingness to pay vary by country and year. Publish quantities as well as unit prices and rerun scenarios locally.
  9. Surveillance and bias: sensor-based function can exclude people without technology, mistake activity for valued participation, or introduce privacy and algorithmic bias.
  10. Patent-as-evidence error: a granted patent establishes neither clinical validity nor economic value.

Four patents illustrate the technical landscape without validating any product: a rehabilitation-condition assessment and management system using outcome measures (WO2019067554A1); a system that relates a person’s functional ability to task-specific complexity (US11540757B2); a secure telerehabilitation system that stores performance results (WO2007019446A2); and a PCT publication claiming ICF-linked assessment, prognostication and tracking of functional independence after neurological injury (WO2024110974A1). The latter record is reported as ceased by Google Patents, whose status notice is not a legal opinion. Patent claims, legal status, freedom to operate, measurement validity and effectiveness are separate questions.

A minimum reporting set

A publishable estimate of the economics of functional outcomes should disclose:

  • decision problem, population, setting, intervention and comparator;
  • outcome contract and patient involvement in choosing it;
  • study design, causal estimand and analysis population;
  • perspective, currency/price year, time horizon and discount rate;
  • quantities and unit costs, including implementation and downstream care;
  • baseline and follow-up function, both continuous and responder results where possible;
  • deaths, harms, missing data and protocol deviations;
  • incremental costs, incremental outcomes and joint uncertainty;
  • subgroup and distributional results, including severity and access;
  • QALYs, cost-consequences and budget impact when relevant;
  • assumptions, model validation, conflicts of interest and access to the analytic model.

CHEERS 2022 provides the current reporting framework for economic evaluations and emphasizes the decision context, perspectives, stakeholder engagement, distributional effects, uncertainty and model sharing (Husereau et al.; DOI 10.1136/bmj-2021-067975; PMID 35017145). A systematic review of rehabilitation economic evaluations found that methods remained frequently under-reported even after the earlier CHEERS checklist became available (Flemming et al., 2022; DOI 10.1080/09638288.2020.1830441; PMID 33126829). Transparency is therefore not an appendix to the metric; it is what makes the metric interpretable.

Conclusion

The central economic question is not “How much did rehabilitation cost?” but “Compared with what, how much additional, patient-relevant function was achieved, for how long, for whom, at whose expense, and with what uncertainty and harm?”

Cost per functional outcome can make rehabilitation value visible. It can connect a clinical change to care hours, family burden, participation and budgets. Its power comes from specificity, and its danger comes from the same source: a narrow or poorly chosen function can exclude what matters, reward selection and conceal mortality, inequality or transferred costs.

The most defensible interscalar analysis therefore does not collapse medicine and economics into one seductive number. It links them through traceable events and presents a portfolio: functional change and duration, safety, QALYs where appropriate, consequences across sectors, affordability and equity. A decision can then be efficient without confusing efficiency with the value of a human life.

References / Список литературы

All sources below are new relative to the current four-sentence article. The bibliography is organized by source type; links point to the primary publication, official standard or guidance, or patent record rather than to a secondary summary. / Все перечисленные источники добавлены к исходной краткой статье. Список организован по типам источников; ссылки ведут на первичную публикацию, официальный стандарт или руководство либо патентный документ, а не на вторичный пересказ.

Official classifications, standards and guidance

  1. World Health Organization. International Classification of Functioning, Disability and Health (ICF); endorsed in resolution WHA 54.21 (2001). Official page.
  2. Üstün TB et al., eds. Measuring Health and Disability: Manual for WHO Disability Assessment Schedule (WHODAS 2.0). WHO; 2010. ISBN 978-92-4-154759-8. Official publication.
  3. World Health Organization. Rehabilitation in health financing: opportunities on the way to universal health coverage. 2024. ISBN 978-92-4-008182-6. Official publication.
  4. NICE. Technology appraisal and highly specialised technologies guidance: economic evaluation. Current online methods manual. Official guidance.
  5. NICE. HealthTech programme manual: methods for guidance. Official guidance.
  6. US Food and Drug Administration. Patient-Focused Drug Development: Selecting, Developing, or Modifying Fit-for-Purpose Clinical Outcome Assessments. Final Guidance 3; October 2025; docket FDA-2022-D-1385. Official guidance.
  7. US Department of Health and Human Services, Office for Civil Rights. Section 504 of the Rehabilitation Act: 2024 Final Rule. Official page.
  8. World Health Organization. Tracking rehabilitative care expenditure under the System of Health Accounts 2011. 2026. ISBN 978-92-4-011878-2. Official publication.
  9. World Health Organization / World Rehabilitation Alliance. Rehabilitation and Functioning in the 2030 Agenda. Global advocacy brief; 8 July 2026. Official publication.
  10. International Organization for Standardization. ISO 9999:2022 — Assistive products: Classification and terminology. 7th ed.; May 2022. Official standard record.

Peer-reviewed methods and thematic publications

  1. Husereau D et al. CHEERS 2022 statement: updated reporting guidance for health economic evaluations. BMJ. 2022;376:e067975. DOI 10.1136/bmj-2021-067975; PMID 35017145.
  2. Sanders GD et al. Recommendations for Conduct, Methodological Practices, and Reporting of Cost-effectiveness Analyses: Second Panel on Cost-Effectiveness in Health and Medicine. JAMA. 2016;316:1093–1103. DOI 10.1001/jama.2016.12195; PMID 27623463.
  3. Sullivan SD et al. Budget impact analysis—principles of good practice. Value Health. 2014;17:5–14. DOI 10.1016/j.jval.2013.08.2291; PMID 24438712.
  4. Porter ME. What Is Value in Health Care? N Engl J Med. 2010;363:2477–2481. DOI 10.1056/NEJMp1011024; PMID 21142528.
  5. Fries JF et al. Item response theory, computerized adaptive testing, and PROMIS: assessment of physical function. J Rheumatol. 2014;41:153–158. DOI 10.3899/jrheum.130813; PMID 24241485.
  6. Devji T et al. Minimal important difference estimates for patient-reported outcomes: a systematic survey. J Clin Epidemiol. 2021;133:61–71. DOI 10.1016/j.jclinepi.2020.11.024; PMID 33321175.
  7. Hernán MA et al. The Target Trial Framework for Causal Inference From Observational Data: Why and When Is It Helpful? Ann Intern Med. 2025;178:402–407. DOI 10.7326/ANNALS-24-01871; PMID 39961105.
  8. Flemming J et al. Quality of reporting of economic evaluations in rehabilitation research: a systematic review. Disabil Rehabil. 2022;44:2233–2240. DOI 10.1080/09638288.2020.1830441; PMID 33126829.
  9. Schneider P. The QALY is ableist: on the unethical implications of health states worse than dead. Qual Life Res. 2022;31:1545–1552. DOI 10.1007/s11136-021-03052-4; PMID 34882282.
  10. Ma Y et al. Do Quality-Adjusted-Life-Year-Based Cost-Effectiveness Analyses Discriminate Against Disabled Patients? A Theoretical and Empirical Analysis. Value Health. 2026, online ahead of print. DOI 10.1016/j.jval.2026.04.005; PMID 42061674.
  11. van Zwet EW, Harrell FE Jr, Senn S. An Empirical Assessment of the Cost of Dichotomization of the Outcome of Clinical Trials. Stat Med. 2026;45(3–5):e70402. DOI 10.1002/sim.70402; PMID 41641823.
  12. Brouwer VHEW et al. Rehabilitation outcomes of older persons within the context of the International Classification of Functioning, Disability and Health (ICF): a systematic review. Eur Geriatr Med. 2026;17(1):261–272. DOI 10.1007/s41999-026-01406-0; PMID 41563587.
  13. Gimigliano F et al. Financial arrangements for rehabilitation services in health systems: an overview of systematic reviews. Cochrane Database Syst Rev. 2026;2:CD016327. DOI 10.1002/14651858.CD016327; PMID 41700590.

Applied rehabilitation evidence

  1. Langhorne P, Baylan S; Early Supported Discharge Trialists. Early supported discharge services for people with acute stroke. Cochrane Database Syst Rev. 2017;7:CD000443. DOI 10.1002/14651858.CD000443.pub4; PMID 28703869.
  2. Turner-Stokes L et al. Cost-efficiency of specialist hyperacute in-patient rehabilitation services for medically unstable patients with complex rehabilitation needs: a prospective cohort analysis. BMJ Open. 2016;6:e012112. DOI 10.1136/bmjopen-2016-012112; PMID 27609852.
  3. Hurley MV et al. Economic evaluation of a rehabilitation program integrating exercise, self-management, and active coping strategies for chronic knee pain. Arthritis Rheum. 2007;57:1220–1229. DOI 10.1002/art.23011; PMID 17907207.
  4. Cano-de-la-Cuerda R et al. Economic Cost of Rehabilitation with Robotic and Virtual Reality Systems in People with Neurological Disorders: A Systematic Review. J Clin Med. 2024;13:1531. DOI 10.3390/jcm13061531; PMID 38541755.
  5. Meyer MJ et al. Rehabilitation impact indices and their independent predictors: a systematic review. BMJ Open. 2013;3:e003483. DOI 10.1136/bmjopen-2013-003483; PMID 24068767.
  6. Syukriyah D et al. Costs and cost effectiveness of the use of telerehabilitation training for upper limb function in people after stroke: a systematic review. Clin Rehabil. 2026, online ahead of print. DOI 10.1177/02692155261441561; PMID 42068268.
  7. Huang C-C et al. The Impact of Taiwan’s Nationwide Postacute Care Program on Functional Recovery, Health Care Utilization, and Cost-Effectiveness for Patients With Stroke: A Systematic Review and Meta-Analysis. Arch Phys Med Rehabil. 2026, online ahead of print. DOI 10.1016/j.apmr.2026.03.014; PMID 42150726.

Patents (technology landscape; not clinical evidence)

  1. Rehabilitation Institute of Chicago. Assessment and management system for rehabilitative conditions and related methods. WO2019067554A1. Priority 2017-09-27; publication 2019-04-04.
  2. Koninklijke Philips NV. Assessing the functional ability of a person to perform a task. US11540757B2. US grant published 2023-01-03.
  3. NeuroTone Inc. Secure telerehabilitation system and method. WO2007019446A2. Publication 2007-02-15.
  4. PRS Neurosciences & Mechatronics Research Institute PNMRI Pvt Ltd. A system and methodology for assessment, prognostication and tracking recovery of damaged “functional” brain circuits for predicting functional independence. WO2024110974A1. Publication 2024-05-30; the linked aggregator currently reports the PCT record as ceased and disclaims legal analysis.