How Rehabilitation Care Is Designed, and What Cochrane Found When It Mapped the Evidence

The Question

Healthcare policy and systems research (HPSR) is the discipline that asks how care is organized, financed, governed, and implemented at the system level. The Cochrane Effective Practice and Organisation of Care (EPOC) Group is the methodological home for this work within Cochrane. EPOC's taxonomy divides HPSR into four pillars: delivery arrangements (how care is organized), financial arrangements (how care is paid for), governance arrangements (who decides what care happens), and implementation strategies (how new approaches enter practice).

In 2026, Cochrane Rehabilitation and the WHO Rehabilitation 2030 Programme co-produced four overviews, one for each EPOC pillar, focused on rehabilitation specifically. The Del Furia paper covers the first pillar: delivery arrangements. Sibling papers on financial, governance, and implementation are published or forthcoming.

The clinical motivation is obvious. Rehabilitation is one of the most variable healthcare services in the world. Who provides it (PT, OT, physiatrist, chiropractor, exercise physiologist, primary care) varies by country, region, and payer. Where it is provided (hospital, outpatient clinic, home, telehealth, community center) varies likewise. How it is coordinated (single-provider vs multidisciplinary, integrated vs fragmented) varies further still. Recommending better systems requires evidence on what different arrangements produce.

What They Found

The team searched Epistemonikos and EPOC databases for systematic reviews of randomized and non-randomized intervention studies evaluating delivery-arrangement effects in rehabilitation. They used AMSTAR 2 to appraise the included reviews. The numbers are sobering.

25 reviews met inclusion criteria. 3 of those overlapped, leaving 22 unique. 17 of the 22 had LOW or CRITICALLY LOW confidence on AMSTAR 2. That is the methodological floor: a meta-analysis can only be as trustworthy as the systematic reviews it pools, and three-quarters of the reviews available were not strong enough to anchor confident conclusions.

Five reviews contributed to the synthesis. Three of those reached moderate to high confidence on outcomes of interest. Their findings:

Advanced practice physiotherapy (APP) for adults with spinal pain. One review pooled 2 studies and 225 participants. APP models compared with usual care produced little to no difference on the Pain Disability Index or EuroQol-5D after the intervention (SMD 0.05, 95% CI -0.32 to 0.42, low certainty). In other words: changing who delivers care, from a generalist to an advanced-practice provider, did not move the outcomes patients care about.

Telerehabilitation for older adults. One review pooled 3 studies and 179 participants comparing telerehabilitation with usual care. After 7 to 20 weeks, little or no effect on quality of life (SMD -0.09, low certainty). Mobility, strength, and balance estimates were directionally favorable to telerehab but very low certainty with confidence intervals crossing zero.

Telerehabilitation for stroke survivors. One review pooled multiple studies (up to 1,145 participants for some outcomes). Telerehabilitation vs usual care: probably little or no effect on independence in ADL, self-reported quality of life, or depression at 6 to 24 weeks (moderate certainty). Compared with in-person rehabilitation, telerehabilitation produced little to no difference on ADL, balance, or upper limb function.

The pattern is consistent: where the evidence is strong enough to draw conclusions, the effect of these specific delivery arrangements on patient outcomes is small. Telehealth, where it has been studied, looks roughly comparable to in-person care. Advanced-practice models of care delivery look roughly comparable to usual care.

The Bigger Finding Is in the Gaps

The clinical findings above are a small part of the story. The structural finding is more important.

EPOC defines five subcategories of delivery arrangements: who provides care, what care is provided, where care is provided, how care is coordinated, and how care is delivered with information and communication technology. The Del Furia overview was able to find reliable evidence in only TWO of the five subcategories. Most of that evidence was concentrated on a single topic (telerehabilitation, an ICT subcategory). The other three subcategories (what care is provided, where, how it is coordinated) were essentially blank.

Equity-related outcomes were absent from the included reviews. Quality of care, adverse events, and the team's preferred system-level outcomes were rarely reported. The reviews focused on patient-level effects (does the intervention help an individual?) rather than system-level effects (does the arrangement help a population, an underserved community, a healthcare budget?).

This is the structural critique. The evidence base for how rehabilitation should be organized has been built around what is measurable in conventional RCTs (does telehealth produce the same disability score as in-person?), not around what would help a policymaker design a better system (which arrangements produce equitable access? which produce affordability? which produce quality variation?). The questions that an EPOC overview is designed to answer have not been studied at the scale or with the rigor that those questions deserve.

Why the Evidence Got Built the Way It Did

The pattern visible in the Del Furia paper is consistent with a structural reality I have written about elsewhere. Research follows funding. Funding follows what is measurable, scalable, and answerable in a 3-to-5-year grant cycle. Telehealth-vs-in-person is answerable in that timeframe. "Does our healthcare system produce equitable rehabilitation access?" is not. Quality-of-care variation across arrangements requires multi-year observational research, large administrative datasets, and political will to act on the findings. RCT-graded evidence cannot be expected to fill this gap because the question is not RCT-shaped.

The Del Furia team explicitly notes this. Their concluding recommendation: "Future Cochrane overviews in HPSR should consider a broader range of study designs, such as observational, qualitative, and mixed-design evidence, to better capture evidence on delivery arrangements in rehabilitation."

That recommendation is correct. It also requires a methodological shift that the field has resisted for decades. Observational evidence is harder to pool. Qualitative evidence is harder to summarize. The Cochrane infrastructure is built around RCT meta-analysis. Asking it to absorb mixed-design evidence is asking for a different kind of synthesis. It is the right ask. It will take time.

What This Means for Chiropractic Practice and Patients

Three implications I would flag, with varying confidence levels.

First, the absence of strong delivery-arrangement evidence does NOT mean delivery arrangements do not matter. It means they have been studied less rigorously than the clinical effectiveness questions covered in other commentaries this issue (the Cochrane SMT review and the PACBACK trial). The field has invested heavily in "does this treatment work" and lightly in "does this care arrangement work." The investment imbalance has consequences for what is known and for what gets reformed.

Second, the cash-pay walk-in chiropractic model that Vanguard Chiropractic operates is not represented in any of the included reviews. The Del Furia paper considers arrangements within insurance-billing rehabilitation systems. Cash-pay direct-care models exist outside that frame. Clinics like mine do not appear in the literature because the literature has not studied this arrangement. That is a research gap, not a policy verdict.

Third, telehealth has been studied more than other delivery arrangements largely because telehealth standardizes well. Standardization makes RCTs feasible. Other arrangements (community-based care, integrative-medicine clinics, primary-care-embedded chiropractic, FQHC integration models) standardize less well, so they get studied less. The result is that telehealth dominates the evidence base for delivery arrangements not because it is the most clinically interesting, but because it is the easiest to study under conventional methods.

Connecting to the Declaration

This commentary is the first health-policy article in The Forward Review. It connects directly to Article I (On Design) of the Declaration on Healthcare System Incentives. The Declaration argues that healthcare structure is shaped by incentive design rather than by deliberate planning, and that the system optimizes for what is reimbursable, measurable, and scalable rather than for what is most useful to patients or populations. The Del Furia paper is, in effect, the empirical receipts for that argument as it applies to rehabilitation. The evidence base that exists is the evidence base that grant funders, RCT methodology, and Cochrane publication norms produced. The evidence base that does NOT exist (equity, quality of care, system-level outcomes, four of five EPOC subcategories) is the evidence base that funders, methods, and norms did not produce.

This is not corruption. The Declaration is careful about this point. It is structural gravity. Evidence accumulates where the inputs accumulate. The Cochrane and WHO teams writing this paper are working in good faith to map what exists. What exists has been shaped by decades of incentive design that prioritized treatment effectiveness over arrangement effectiveness.