A study published online on September 8, 2026 in the British Journal of Sports Medicine, tracking about 65,000 UK adults, found the lowest risk of early death at 7,500 to 10,000 steps a day. A July 2025 meta-analysis in The Lancet Public Health found the benefit curve flattening at roughly 5,000 to 7,000 steps for several outcomes. For readers, 10,000 is not a pass mark: much of the measured gain comes earlier, and the counter you own is good enough to track your trend, if not to match a study's cutoff.
Key takeaways
- In the September 2026 study, 7,500 to 10,000 steps a day had the lowest risk of death of the four step bands compared.
- In the 2025 meta-analysis, 7,000 steps a day was associated with a 47% lower risk of death from any cause than 2,000 steps a day.
- A 2022 pooled analysis put the plateau at about 6,000 to 8,000 steps for adults 60 and over, and 8,000 to 10,000 for younger adults.
- The federal activity guidelines are written in minutes and say 10,000 steps is "not a guideline per se."
- These are observational studies: they show association, not proof that steps cause the lower risk.
What it means for you
You are mostly sitting (under about 5,000 steps)
For the outcomes whose curves level off, most of the measured drop in risk came before 5,000 to 7,000 steps, which is your range. The worked example converts 2,000 extra steps into minutes.
You already average 7,000 to 8,000
You are near where several curves flatten, and the new study's lowest-risk band starts at 7,500. In the 2022 analysis the most active quarter (median about 10,900) had the lowest risk, so more was not linked to higher risk.
You are 60 or older
Your plateau in the 2022 analysis was lower, at 6,000 to 8,000 steps. The guidelines add balance and strength work (see the FAQ).
You use a phone, or you use a wrist tracker
Neither is exact, and the direction of error depends on device and setting, so comparing your own weeks matters more than matching a study's number.
You have a health condition
A clinician can set a goal that fits your heart, joints or medications. The guidelines say adults with chronic conditions who cannot meet the main targets should be active according to their abilities and avoid inactivity.
Educational information only, not medical advice. Check with a clinician before changing your activity level if you have a heart, lung, joint or metabolic condition, or if walking causes pain or dizziness.
Where 10,000 steps came from
In Harvard Health's account, epidemiologist I-Min Lee traced the number to 1965, when a Japanese company made a device called Manpo-kei, which translates to "10,000 steps meter." Lee called the name a marketing tool.
The Physical Activity Guidelines for Americans, 2nd edition, take the same line: walking 10,000 steps a day "is not a guideline per se, but a way people may choose to meet the key guidelines."
What the large studies found
The 2025 review and dose-response meta-analysis (Ding et al., The Lancet Public Health, July 2025) pooled prospective studies of device-measured steps: 57 studies from 35 cohorts, 31 of them in the meta-analysis. Its estimates for 7,000 versus 2,000 steps a day:
| Outcome | Lower risk at 7,000 vs 2,000 steps | Certainty (GRADE) |
|---|---|---|
| All-cause mortality | 47% (HR 0.53) | Moderate |
| Cardiovascular disease incidence | 25% (HR 0.75) | Moderate |
| Cardiovascular disease mortality | 47% (HR 0.53) | Low |
| Cancer mortality | 37% (HR 0.63) | Moderate |
| Cancer incidence | 6%, not statistically significant | Low |
| Type 2 diabetes | 14% (HR 0.86) | Moderate |
| Dementia | 38% (HR 0.62) | Moderate |
| Depressive symptoms | 22% (HR 0.78) | Moderate |
| Falls | 28% (HR 0.72) | Very low |
HR is the hazard ratio, the risk in one group relative to the comparison group; 0.53 means 47% lower. The authors called 10,000 still viable for more active people and 7,000 perhaps more achievable for some, and listed few studies per outcome, no age-specific analysis and residual confounding as limits.
The 2022 pooled analysis of 15 cohorts (Paluch et al., The Lancet Public Health) covered 47,471 adults and 3,013 deaths over a median 7.1 years. Against the lowest quarter (median 3,553 steps), hazard ratios were 0.60 at 5,801 steps, 0.55 at 7,842 and 0.47 at 10,901. Risk fell until about 6,000 to 8,000 steps at 60 and over, and 8,000 to 10,000 under 60.
The September 2026 study (Wei et al.) used UK Biobank wrist-tracker data on 64,743 adults, average age 61.5, of whom 1,697 died. Per the BMJ Group summary, trackers were worn for seven days in 2013 to 2015 and deaths followed to November 2022. Of four bands (under 5,000, 5,000 to 7,500, 7,500 to 10,000, over 10,000), 7,500 to 10,000 had the lowest risk of death. It also linked a faster stepping pace to lower risk below 7,500 steps. Participants were middle-aged and older UK volunteers, so the bands describe that group, not a trial-tested target.
2,000 extra steps in real time
Converting steps needs two assumptions: a walking rate (100 steps per minute, a rough marker of moderate walking discussed below, or a slower 80) and a step length (2,000 to 2,500 steps per mile, varying with height and pace). Both are rough.
Illustrative example (arithmetic from the stated assumptions, not a measurement of any person):
| Extra steps | Minutes at 100 steps/min | Minutes at 80 steps/min | Miles at 2,000-2,500 steps/mile |
|---|---|---|---|
| 2,000 | 20 | 25 | 0.8 to 1.0 |
| 5,000 (2,000 to 7,000) | 50 | 62.5 | 2.0 to 2.5 |
Result: about 20 to 25 minutes and roughly a mile for the first 2,000 steps, and 50 to about 63 minutes for the climb to 7,000. Steps around the house count too, so these minutes are the cost if every step came in one session, not a required routine.
How accurate phones and wrists are
A 2020 systematic review (Fuller et al.) covered 158 publications on nine wearable brands; in the lab, Fitbit, Apple Watch and Samsung appeared to count steps accurately. The full text judged accuracy against limits of plus or minus 3% in the lab and 10% in free-living settings. In the lab, 45% of comparisons fell within 3%, and devices tended to undercount (median error -2%). In a smaller set of everyday-life comparisons, 42% fell within 10%, and the median error was a slight overcount (+6%). No brand measured calories burned accurately.
A 2015 JAMA research letter (Case et al.) had 14 healthy young adults walk 500 and 1,500 steps on a treadmill at 3.0 mph against an observer's tally. Apps on phones carried in a pants pocket were off by -6.7% to 6.2%, wrist devices by -22.7% to -1.5%, and a waist-worn pedometer and clip-on trackers by -0.3% to 1.0%. The sample was small; the devices date from 2014.
So wrist devices undercounted on that treadmill, while everyday-life results leaned toward a slight overcount. With everyday errors often beyond 10%, treat a study's cutoff loosely and compare your weeks with your own earlier weeks on the same device.
Steps versus the 150-minute goal
The federal guidelines call for adults to get 150 to 300 minutes a week of moderate-intensity aerobic activity, such as brisk walking, or 75 to 150 minutes of vigorous activity, or a mix. For step counters, they suggest a time goal first: count the steps in your own 10-minute brisk walk, then add that many per 10 minutes of goal to your usual total. In their example, a person at 5,000 steps can add 500 a day each week; less fit people add fewer.
A 2009 treadmill study (Marshall et al.) with 97 adults put moderate-intensity walking at about 100 or more steps per minute and suggested 3,000 steps in 30 minutes on 5 days a week, while cautioning that steps per minute is a poor proxy for intensity.
A daily total says how much you moved, not whether any of it was moderate. At 100 steps per minute, 2,000 extra steps a day add up to 140 minutes a week, just under 150, and only if every step was at that rate.
Frequently asked questions
Does 7,000 steps mean the benefit stops there?
No. For all-cause mortality, cardiovascular disease incidence, dementia and falls, the 2025 meta-analysis found curves that flatten, rather than end, around 5,000 to 7,000 steps. For cardiovascular mortality, cancer, type 2 diabetes and depressive symptoms, risk kept falling in a straight line across the range studied.
Do these studies prove that walking lowers the risk?
No. The September 2026 study is an observational cohort, and the 2025 authors list residual confounding as a limit. People who walk more may differ in health or mobility, and early illness can reduce walking.
Was one week of tracking enough to judge years of risk?
It is a limit of the 2026 study: one week of steps in 2013 to 2015, then about eight years of follow-up on average. Habits can change over that time.
Do the guidelines differ for adults 65 and older?
The adult guidelines apply, plus multicomponent activity including balance training, with effort judged relative to one's own fitness; by the guidelines' talk test, at moderate effort you can talk but not sing. When chronic conditions prevent 150 minutes a week, older adults should be as physically active as their abilities and conditions allow.
Does walking cover the strength part of the guidelines?
No. The guidelines add muscle-strengthening activity of moderate or greater intensity, involving all major muscle groups, on 2 or more days a week. A step counter does not capture this.
A practical way to use your counter
- Check your recent weekly average on the device you already have.
- If you are under about 5,000, aim for about 2,000 above your average, built up gradually: the guidelines' example adds about 500 steps a day each week, less if you are less fit.
- If you are near 7,000 to 8,000, hold it and check separately whether you reach 150 moderate minutes a week.
- Compare weeks on the same device, and ignore the calorie figure.
- Schedule strength work on two days a week, plus balance work if you are 65 or older.
- Take any pain, dizziness or condition question to a clinician before raising your goal.
Sources
- Wei et al., stepping intensity and daily step counts, BJSM, 2026
- BMJ Group summary of Wei et al.
- Ding et al., daily steps and health outcomes, Lancet Public Health, 2025
- Paluch et al., daily steps and mortality in 15 cohorts, Lancet Public Health, 2022
- Harvard Health, 10,000 steps a day or fewer
- Fuller et al., wearables systematic review, JMIR mHealth uHealth, 2020 (full text)
- Case et al., accuracy of phone apps and wearables, JAMA, 2015
- Marshall et al., 3000 steps in 30 minutes, Am J Prev Med, 2009
- Physical Activity Guidelines for Americans, 2nd edition
