Guided strength work has the better record
Medium confidence
Across sports, injury-prevention exercise cut injuries by about a third when a coach or physio supervised it, and showed no benefit when done alone. In runners, supervised programs have worked and one self-directed 10-minute program did not, but one online program done alone also cut injuries. The picture favors guidance without proving it.
Why it works
Supervision likely helps through better adherence, progression and technique. Valentin found adherence tracked with fewer injuries, and Toresdahl's self-directed group averaged only 2 sessions a week, but no study has isolated which part of supervision matters.
What it means in practice
Where possible, point runners to guided, progressive strength work such as a coach, physio or class. A structured program with clear week-to-week progression is a reasonable alternative, and one runner trial found an online program cut injuries. Do not claim that app-delivered programs match in-person supervision, because the supervision evidence comes from in-person programs.
The evidence
-
Across 44 RCTs, exercise-based injury prevention programs cut injury risk overall (RR 0.71), with female-only (RR 0.73) and male-only (RR 0.65) cohorts benefiting about equally. Supervision is what separated the results: supervised programs worked (RR 0.67, a 33% reduction), while unsupervised ones did not (RR 1.04). Age up to early middle age and intervention duration showed no significant association with effectiveness, but adherence tracked inversely with injury rate (β = -0.014, p = 0.004). This pools across sports, so the running-specific inference stays limited.
-
Pooled across runner-specific RCTs, exercise-based prevention programs do not appear to reduce the risk or rate of running-related injuries. Two trials did show a significant protective effect, both with high adherence; one was Taddei 2020, a foot-core protocol whose control group had a 2.42-fold higher injury rate, and supervised programs may help partly through higher compliance. Only four of the included trials used strength-based exercise, all of relatively low volume and intensity, which may partly explain the null result. Higher volumes and intensities of strength training have shown consistently favorable results for overuse injuries in other sports (Lauersen 2018), so Wu and colleagues call for further research on strength training for runners.
n=1904
-
Over the 12-month trial, control-group runners were 2.42 times more likely to sustain a running-related injury than runners who trained their foot and ankle muscles (95% CI 1.98-3.62, p=0.035). Time to injury tracked with both Foot Posture Index (r=0.41, p=0.031) and foot strength gain (r=0.45, p=0.044). The authors report the risk reduction showed at 4 to 8 months of training, so treat it as a sustained programme rather than a short block. This is a single single-blind RCT of 118 recreational runners, so the finding still awaits replication.
n=118
-
The physiotherapist-supervised hip and core programme lowered lower-extremity injury incidence versus control (HR 0.66). Average weekly prevalence of overuse injuries was 39% lower and substantial overuse injuries 52% lower. The ankle and foot programme did not reduce overuse injuries and had more acute injuries than control.
n=325
-
A 10-minute home strength program done three times a week did not lower injury risk in first-time NYC Marathon runners. Overuse injuries that kept runners from finishing the marathon hit 7.1% of the strength group and 7.3% of the observation group (RR 0.97, 95% CI 0.57-1.63, p=0.90), and finishing times matched at 5h 1min versus 4h 58min (p=0.35). Overuse drove most of the harm, with 52 of 64 major injuries overuse and 20 of those bone stress injuries, against a background rate of 8.9% major and 48.5% minor injury. The authors suggest the 10-minute program may not have produced the strength gains or neuromuscular control needed (a longer or more frequent program likely would), that it was unsupervised with self-reported compliance averaging 2.0 sessions a week, and that the observation group was not a true control.
n=720
Why we call confidence medium
Medium: the supervision split comes from a cross-sport meta-analysis, and runner evidence is mixed. Valentin's 2023 meta-analysis of 44 trials across sports found supervised injury-prevention exercise cut injuries by about a third (RR 0.67) while unsupervised programs showed no benefit (RR 1.04). In runners, Wu's 2024 meta-analysis (9 trials, mostly low quality) found no overall effect of prevention exercise, with a post hoc supervised-only analysis showing lower risk. Supervised runner programs worked in Taddei 2020 (118 recreational runners, foot-core training) and Leppänen 2024 (325 novice runners, hip and core program, HR 0.66), and a self-directed 10-minute program did not in Toresdahl 2020 (720 first-time marathoners). But Naderi 2025 (264 recreational runners) found an unsupervised online program cut injuries by about half (IRR 0.53).
Where it applies
Adult runners, especially anyone new to structured strength training. The supervision comparison comes mostly from other sports.
Does not apply to: experienced strength-trainers who already train with progressive loading.
Plans that respect this
Plans that scored well on the rubric measures informed by this claim.
- 10-Week Run a Faster 10k (3 days)
- 10-Week Run a Faster 10k (4 days)
- 10-Week Run a Faster 10k (5 days)
- 10-Week Run a Faster 10k (6 days)
- 10-Week Run a Faster Half Marathon (3 days)
- 10-Week Run a Faster Half Marathon (4 days)
- 10-Week Run a Faster Half Marathon (5 days)
- 10-Week Run Your First 10k (3 days)
- 10-Week Run Your First 10k (4 days)
- 10-Week Run Your First Half Marathon (3 days)
- 10-Week Run Your First Half Marathon (4 days)
- 10-Week Sub-2 Half Marathon (3 days)
Last reviewed Oct 14, 2026. See how we score.