How many steps a day: the research behind the number

Steps are the easiest activity measure there is: nearly every phone and watch counts them. That has produced a large body of research linking device-measured steps to health outcomes. Two meta-analyses in The Lancet Public Health are the best summary.

The two big analyses

Paluch 2022: the age split

The Steps for Health Collaborative pooled raw data from 15 cohorts, followed for a median of 7.1 years. Compared with the lowest quarter of walkers (a median of 3,553 steps a day), the highest quarter (10,901 steps) had a 53 percent lower risk of death (HR 0.47), and the two middle quarters (5,801 and 7,842 steps) had 40 and 45 percent lower risk. Risk fell steadily with more steps, then flattened. The point where it flattened depended on age: about 6,000 to 8,000 steps a day for adults 60 and older, about 8,000 to 10,000 for younger adults. The authors open by noting that the 10,000 a day target has little evidence behind it. They also looked at pace. After accounting for total steps, a faster peak 30-minute and peak 60-minute cadence was still linked to lower mortality (HR 0.67 for each, top vs bottom quarter), but minutes per day spent walking at 40 or 100 steps a minute or faster were not. Total steps did most of the work; a brisk stretch each day may add a little.

Ding 2025: beyond mortality

The 2025 review widened the outcomes to heart disease, cancer, diabetes, dementia, depression, physical function and falls. Compared with 2,000 steps a day, 7,000 steps a day was associated with:

• 47% lower all-cause mortality (HR 0.53)
• 25% lower cardiovascular disease incidence (HR 0.75) and 47% lower cardiovascular mortality (HR 0.53)
• 37% lower cancer mortality (HR 0.63); the 6% lower cancer incidence was not statistically significant
• 14% lower type 2 diabetes (HR 0.86)
• 38% lower dementia (HR 0.62)
• 22% lower depressive symptoms (HR 0.78)

For death, cardiovascular disease, dementia and falls, the curve bent at around 5,000 to 7,000 steps a day, with smaller gains after that. For cardiovascular and cancer death, diabetes and depressive symptoms the relationship was closer to a straight line, so more kept helping.

Read these carefully

Both analyses are observational. People who walk more are, on average, healthier to start with, and some walk less because they are already ill. The studies adjust for this, but not perfectly, so the percentages are associations, not guaranteed effects. Some of the confidence intervals are wide, and a few outcomes rest on two or three studies. The direction, though, is consistent: going from very few steps to a moderate number is where most of the benefit sits. That includes falls, which matter most for older adults: in the 2025 analysis, the risk of falls also fell as steps rose, with the curve bending in the same 5,000 to 7,000 range.

What to aim for

• If you are under 5,000 a day, the gain from adding 2,000 to 3,000 is the largest on the curve. Start there.
• A target of about 7,000 is a reasonable, evidence-based goal for most adults. Over 60, 6,000 to 8,000 is where the Paluch curve flattened; under 60, 8,000 to 10,000.
• More is fine. Nothing in either analysis suggests harm from more steps, and some outcomes kept improving.
• Steps are not the whole guideline. The US guidelines ask for 150 to 300 minutes of moderate activity plus two days of strength work. Brisk walking counts towards the first; it does not cover the second. See strength training and cardio and Zone 2.

Practical targets by starting point

Step counts from phones and watches differ by device and by where the device is worn, so compare your own numbers over time rather than against someone else's.

Easy ways to add steps

A 10-minute walk after meals, walking calls, getting off a stop early, stairs instead of lifts. For most people 1,000 steps is roughly 10 minutes of walking.

Logging it

Steps are only useful as a trend. Peptide AI pulls daily steps from Apple Health, Health Connect, Garmin, Oura and WHOOP through wearable integrations and puts them on the same timeline as the training log, sleep and weight. A rolling weekly average is more honest than any single day.