REF-716 · Relph, Greaves, Armstrong, Prior, Spencer, Griffiths, Dey & Langley, 2022 · Cochrane Database of Systematic Reviews, 2022 Aug 22;8 · Paper · PubMed
Cochrane review, 12 trials and 11,240 runners: no shoe type reliably reduces running injuries
Including cushioned versus minimalist — and prescribing shoes by foot posture does not work
Key Findings
12 trials, 11,240 participants, 6 to 26 weeks, across North America, Europe, Australia and South Africa.
| Comparison | Studies / participants | Injury result |
|---|---|---|
| Neutral/cushioned vs minimalist | 5 / 766 | little or no difference — RR 0.77 (95% CI 0.59-1.01) |
| Soft vs hard midsole | 2 / 1095 | little or no difference — RR 0.82 (0.61-1.10) |
| Shoes prescribed by static foot posture vs not (military recruits) | 3 / 7203 | probably little or no difference — Rate Ratio 1.03 (0.94-1.13), moderate certainty |
| Motion control vs neutral/cushioned | 2 / 421 | uncertain, very low certainty — RR 0.92 (0.30-2.81) |
| Stability vs neutral/cushioned | 1 / 57 | uncertain, very low certainty — RR 0.49 (0.18-1.31) |
| Motion control vs stability | 1 / 56 | uncertain, very low certainty — RR 3.47 (1.43-8.40), i.e. more injuries in motion-control shoes in one small trial |
Authors' conclusion: most evidence demonstrates no reduction in lower-limb running injuries in adults when comparing different types of running shoes, and there is no evidence that prescribing footwear by foot type reduces injuries. No trial compared running shoes with non-running shoes.
Certainty was low to very low for most comparisons; only the foot-posture-prescription comparison reached moderate certainty — and that is the one that found no benefit most confidently. All three of those trials were in military populations, and the authors note the finding may differ in recreational runners.
Practical Application
Across 12 trials and 11,240 adult runners and military recruits, no running-shoe type reliably reduced lower-limb injuries: neutral/cushioned vs minimalist RR 0.77 (95% CI 0.59-1.01, low certainty), soft vs hard midsole RR 0.82 (0.61-1.10), and shoes prescribed by static foot posture Rate Ratio 1.03 (0.94-1.13, moderate certainty, military recruits). Do not sell a shoe category as injury prevention, in either direction.
Read the minimalist point estimate honestly: RR 0.77 nominally favours neutral/cushioned over minimalist, but the confidence interval crosses 1 and the certainty is low. The defensible statement is "no reliable difference", not "minimalist is worse" — and certainly not "minimalist is better".
Foot-type prescription is the clearest negative. Prescribing by static foot posture is the one comparison with moderate certainty, and it found no effect across 7203 military recruits; the authors caution that recreational runners may differ. Inference, not tested: a shop recommendation based on foot type should not be presented as injury prevention.
Limitations
Certainty is low to very low for most comparisons: runners could not be blinded to their shoe, injury definitions and footwear characteristics varied, and there were too few studies for most comparisons. "No reliable difference" here partly reflects the weakness of the underlying trials, not proof of equivalence.
Only 5 trials and 766 participants informed the cushioned-versus-minimalist comparison — thin for the question most runners actually ask.
Trials ran 6 to 26 weeks. A shoe effect that only appears over years of running would not be visible.
Injury definitions and reporting vary between trials; pooled relative risks hide differences in what counted as an injury.
Says nothing about performance, economy or comfort — only injury.
How this entry was checked and changed
Content correction 2026-09-27
Lead rewritten as the review's actual numbers; added the motion-control vs stability comparison (RR 3.47, very low certainty) and the RRs for motion-control and stability vs neutral.
Foot-posture result now states it comes from military recruits and may differ in recreational runners (authors' caveat).
"The levers that do work are load progression (REF-619)" -> those entries report limited or non-significant evidence; wording softened and labelled.
Removed "largest single footwear RCT" for REF-621 (the foot-posture trials were larger); certainty drivers restated as in the abstract.
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.