When Half Your Patients Hit the MCID — and When Most Do

A trial reports a statistically significant improvement in pain, and the abstract makes it sound like recovery. Then you look at the actual change score and wonder whether your patient would notice the difference. That gap — between statistical significance and clinical meaningfulness — is what the minimal clinically important difference exists to close, and a recent body of literature makes it worth revisiting how you read results.

Significance is not the same as meaning

Nahum (2026), in a narrative review of evidence on musculoskeletal rehabilitation, argues that interpretation using the MCID or responder thresholds is needed to distinguish statistically detectable associations from clinically meaningful change. The same review notes that conventional imaging — ultrasound, CT, MRI — may identify clinically relevant pathology but often shows only modest concordance with pain, disability, and performance-based function. In other words, a statistically significant number on a PROM or a scan does not automatically describe a change the patient would call recovery.

Responder rates vary — and can fade

How many patients actually cross the MCID, and for how long? The numbers vary more than you might expect. In a case series of 28 adults with chronic pain in physiotherapy, Prasad et al. (2026) observed that PSFS scores increased from a median of 2.0 to 8.6, and 24 patients (85.7%) achieved or exceeded the MCID — though the authors caution the findings should be interpreted with caution due to the exploratory nature of the study.

In a larger observational study of 203 consecutive users with musculoskeletal pain, Zunzarren et al. (2026) found the MCID was achieved by 50.2% of patients, with 74.4% showing any improvement; among those with chronic symptoms lasting more than one year, 51.3% achieved the MCID. Patients with higher baseline pain showed greater improvements in that cohort. And in a case series of 44 ankles with osteoarthritis or talar osteochondral lesions, Benegas et al. (2026) observed MCID responder rates of 65.9% at 3 months falling to 36.4% at 6 months, with mean VAS improving from 7.95 at baseline to 5.52 at 3 months and 6.68 at 6 months; 43.2% of ankles underwent an additional intervention.

Across these studies, the share of patients reaching the MCID ranged from about half to 85.7%, and in the one series with two follow-up points, responder rates fell by six months. Because these are within-group observations without comparison groups, they describe what was seen, not what a treatment caused — but the trajectory itself is the clinically useful part. A responder rate measured at one time point can look very different a few months later — in one case series, the MCID rate for ankle pain fell from 65.9% at 3 months to 36.4% at 6 months.

When between-group differences exceed the MCID

Not every improvement clears the bar, and some interventions clear it more decisively than others. In a randomised trial of 42 patients with partial rotator cuff tears, Vrouva et al. (2026) compared TENS plus kinesiotherapy against MENS plus kinesiotherapy: at the 12-month primary endpoint, between-group comparisons favoured MENS for SPADI Pain (mean difference 14.39), SPADI Disability (13.56), NRS pain (2.33) and EQ-5D health status, substantially exceeding established MCID benchmarks. Notably, ultrasound showed post-intervention stabilisation of tear progression in the MENS cohort (0.02 mm secondary depth increase between months 3 and 12) versus progressive expansion in the TENS arm (0.45 mm) — a clinicoradiological paradox the authors set against functional recovery.

MCID and related thresholds are derived within specific patient populations and outcome measures, so they cannot be transferred uncritically between settings. In a retrospective cohort of 80 patients with trigger finger undergoing percutaneous A1 pulley release, Karasu et al. (2026) derived MCID thresholds of ΔVAS ≥ 3, ΔQuickDASH ≥ 19 and ΔMHQ ≥ 16.5, with SCB thresholds of ΔVAS ≥ 4, ΔQuickDASH ≥ 25 and ΔMHQ ≥ 20, and PASS thresholds at 3 months of VAS ≤ 4, QuickDASH ≤ 29 and MHQ ≥ 76. The authors describe these as preliminary estimates that "require prospective external validation" before clinical application — worth remembering before you borrow a threshold from a paper on a different condition.

Reading results with the MCID in hand

For example, one case series reported significant improvements in PSFS scores, with 85.7% of patients achieving or exceeding the minimal clinically important difference. In a retrospective service evaluation of 64 adults referred for MSK physiotherapy, McNamee et al. (2026) observed significant improvements across all outcomes — MSK-HQ (+11.2), NPRS (−2.2), EQ-VAS (+19.9) — and minimal clinically important thresholds were exceeded for MSK-HQ, NPRS and EQ-VAS. The authors note that further comparative and economic evaluations are warranted to establish effectiveness.

Contrast that with a stratified-care pathway study: in a prospective cohort of 136 low-risk adults in primary care, AlMuhana et al. (2026) reported greater improvement at 12 weeks with a STarT MSK-guided pathway than with usual care in NPRS (−2.12), MSK-HQ (+7.08) and ÖMPQ (−27.83), with clinically important pain reduction in 68.3% versus 25.9% of participants. But the authors are explicit: because the pathway combined multiple components, their independent contributions cannot be separated, and the non-randomised design, unblinded participants, self-reported outcomes, and possible residual confounding preclude causal conclusions, supporting further randomised evaluation. A large between-group difference that exceeds an MCID benchmark still deserves that kind of scrutiny before it changes your practice.

What to do with this on Monday

1	When you read a trial, look past the p-value: find the change score and ask whether it exceeds a published MCID for that instrument in that population.
2	Check which threshold is being used — MCID, SCB or PASS — because they answer different questions and carry different values, as the trigger finger data illustrate.
3	A reasonable approach is to look at responder rates alongside group means, as in the digital-intervention cohort where only half of patients achieved the MCID although 74.4% showed some improvement.
4	Ask when the responder rate was measured. Rates at three months and at six months can differ substantially, as the ankle series showed (65.9% versus 36.4%).
5	Treat within-group improvements without a comparison group as observations, not effects — and treat non-randomised between-group findings as hypothesis-generating until validated.

The MCID will not tell you whether an individual patient is improving — that is what your re-assessment is for. What it does is recalibrate how you read the literature: a significant result below the threshold is a statistical event, not a recovery, and a responder rate without a time point attached is half a finding.

When you reassess, PhysioSync's PainSync records VAS and functional scores over time, so you can see each patient's measured change against the threshold you judge clinically meaningful.