Step Count Targets

Daily step counts as a dose of physical activity and the thresholds at which health outcomes change. The question exists because wearables made steps the most widely tracked activity metric, and because the dominant target — 10,000 a day — came from marketing, not evidence.

What the 2025 meta-analysis found

Ding et al., The Lancet Public Health 2025 (57 studies, 35 cohorts, device-measured steps, prospective; 31 studies pooled). Compared with 2,000 steps/day, 7,000 steps/day was associated with:

Outcome	HR (95% CI)
All-cause mortality	0.53 (0.46–0.60)
CVD incidence	0.75 (0.67–0.85)
CVD mortality	0.53 (0.37–0.77)
Cancer mortality	0.63 (0.55–0.72)
Cancer incidence	0.94 (0.87–1.01) — n.s.
Type 2 diabetes	0.86 (0.74–0.99)
Dementia	0.62 (0.53–0.73)
Depressive symptoms	0.78 (0.73–0.83)
Falls	0.72 (0.65–0.81)

The relationships are non-linear with inflection points around 5,000–7,000 steps/day for mortality, CVD incidence, dementia and falls. Benefit continues past 7,000 but flattens; the authors frame 10,000 as fine for the already-active and 7,000 as "more realistic for some".

Caveats attached to the number

• Several outcomes rest on two to four studies; cancer and CVD-mortality estimates are heterogeneous.
• No age stratification — the inflection point may differ for older adults.
• Observational throughout; residual confounding acknowledged by the authors.
• Cadence/speed is not in the model. The Economist lede raises "at what speed"; nothing paged yet answers it.

Related

• Minimum Effective Dose of Exercise — the minutes-based evidence shows the same saturation shape.
• Sedentary Behaviour — steps are the standard offset for sitting time.
• Running Injury — the same wearable-cohort methodology applied to runners, with the opposite question (when does more hurt?).