For about two decades, the number attached to walking has been 10,000. It came from a 1960s Japanese pedometer brand name, not from evidence, and the research has been quietly dismantling it ever since. The latest analysis goes further than correcting the number. It suggests the number was the wrong variable.
Researchers at the University of Sydney, Monash University and the Universidad Europea de Madrid took the UK Biobank accelerometer subsample — 103,684 adults who wore an Axivity AX3 on the wrist for seven days between 2013 and 2015 — and asked a question most step studies skip: what happens when you look at how many steps people take and how fast they take them, together?[1] The all-cause mortality analysis covered 64,743 people, mean age 61.5, followed to 30 November 2022, with 1,697 deaths. The results were published in the British Journal of Sports Medicine and picked up by Harvard Health this month.[2]
What "pace" actually means here
This is the part most coverage gets wrong, and it changes the practical advice completely.
The intensity metric is peak 30-minute cadence: the researchers ranked every minute of your day by steps taken, pulled out your top 30 minutes, and averaged them. Those minutes do not have to be consecutive.[1] Two minutes hurrying to a platform, six minutes to the shop, a brisk stretch after lunch — it all counts toward the same pool.
So this is not a study about going for a 30-minute power walk. It is a study about whether, in the moments you happen to be walking, you walk with intent. The reference point was the fifth percentile of cadence, roughly 45 steps per minute, which is close to the 40-steps-per-minute threshold researchers use to distinguish purposeful walking from shuffling around a kitchen.
The numbers
Below are the adjusted all-cause mortality hazard ratios from the study, with the reference group — under 5,000 steps a day at the slowest cadence — set to 1.00. A hazard ratio of 0.70 means roughly 30% lower risk of dying during follow-up than that reference group.
| Steps/min | <5,000 steps | 5,000–7,500 | 7,500–10,000 | >10,000 |
|---|---|---|---|---|
| 50 | 0.96 (0.91–1.00) | 0.72 (0.51–0.99) | — | — |
| 60 | 0.87 (0.76–1.01) | 0.70 (0.58–0.86) | — | — |
| 70 | 0.80 (0.66–0.97) | 0.69 (0.60–0.81) | 0.65 (0.47–0.90) | — |
| 80 | 0.72 (0.54–0.97) | 0.68 (0.58–0.80) | 0.60 (0.50–0.72) | 0.91 (0.61–1.32) |
| 90 | 0.64 (0.32–1.32) | 0.67 (0.56–0.81) | 0.58 (0.48–0.69) | 0.71 (0.59–0.87) |
| ~100 | — | 0.66 (0.54–0.81) | 0.57 (0.47–0.69) | 0.67 (0.56–0.81) |
Read across any row and the trade-off appears. At 60 steps per minute, someone managing 5,000–7,500 steps a day sat at 0.70. At 80 steps per minute, someone under 5,000 steps sat at 0.72. Statistically those are the same place, arrived at by opposite routes — one person walking further, the other walking faster.
That symmetry is the actual finding, and it is more useful than "walk faster." It means the dose can be assembled two ways, which matters enormously if one of the two is closed to you. Someone with knee osteoarthritis who cannot stride out can bank volume instead. Someone with ninety minutes of commute and no spare hour can bank intensity instead.
Where the headline oversells it
The claim travelling furthest this month is that sedentary people can skip the step count entirely and just walk briskly. The table above is where that claim starts to wobble.
Look at the under-5,000 column as cadence rises. At 70 steps per minute the estimate is 0.80 and the interval (0.66–0.97) excludes 1.00 — a real effect. At 80 it is 0.72 (0.54–0.97), still just clearing the line. At 90 the point estimate looks best of all, 0.64, but the interval blows out to 0.32–1.32. That interval contains 1.00. It is compatible with a 68% risk reduction and with a 32% increase.
The reason is simple: very few people who take under 5,000 steps a day also sustain 90 steps per minute for half an hour. The cell is nearly empty, so the estimate is unstable. The authors are candid about this, noting that wide intervals at high cadence "may also reflect data sparsity at the upper tail."[1]
There is a second wobble worth knowing about. In the 7,500–10,000 group the curve is U-shaped: risk bottoms out near 100 steps per minute and then drifts back up. Taken literally that would mean walking too fast is bad for you, which is almost certainly not what is happening. The authors ran the same analysis with a generalised additive model and the curve straightened out, pointing to sparse data rather than real harm at the fast end.[1]
So the honest summary is narrower than the headline: a brisk pace helps across every step band, and the best-evidenced combination is a moderate step count walked briskly — not a tiny step count rescued by speed.
How this fits the wider evidence
None of this arrived from nowhere. A 2025 Lancet Public Health systematic review pooled 31 studies from 24 cohorts and found the all-cause mortality curve bends sharply and then flattens somewhere around 5,000–7,000 steps a day. Compared with 2,000 steps, 7,000 steps was associated with 47% lower all-cause mortality (HR 0.53, 95% CI 0.46–0.60), 25% lower cardiovascular disease incidence, 38% lower dementia risk and 28% fewer falls.[3] An earlier meta-analysis of 15 international cohorts had already found the benefit plateauing well before 10,000 in older adults.[4]
On intensity specifically, the same UK Biobank cohort analysed in 2022 showed that higher stepping intensity was associated with lower cancer and cardiovascular incidence and mortality beyond the contribution of total steps[5] — and, in a companion paper on 78,430 adults, with lower incident dementia.[6] That is relevant if you have been reading about what happens to the brain after 50: cadence keeps showing up as a signal, not just volume.
Why would pace carry independent information? The most plausible explanation is that cadence is a cheap proxy for cardiorespiratory fitness. You cannot walk at 100 steps per minute for half an hour without a certain amount of aerobic capacity and leg function. The study's authors make this point directly, suggesting the steeper benefit seen in low-step walkers may reflect lower baseline fitness and therefore more room to improve.[1] Cadence may be partly measuring VO2 max without the laboratory — and VO2 max remains one of the strongest single predictors of mortality we have.
What to actually do with this
The practical translation is unusually clean, because the metric is forgiving.
You are aiming for roughly 7,500 steps on a normal day, with about half an hour of it — scattered however it falls — at a pace near 100 steps per minute. In the real world 100 steps per minute is a purposeful walk: you are moving like someone who is slightly late, your breathing is noticeably up, and conversation takes a little effort. Most phones and watches already report cadence, often under "steps per minute" or in the walking-pace section of a workout summary. You can also count your steps for 15 seconds and multiply by four.
Two caveats before you reorganise your week. This is observational data. Everyone in it chose their own walking speed, and the researchers adjusted for age, sex, smoking, alcohol, diet, education, socioeconomic status, medication, screen time, family history and measured sleep — which is thorough, and still cannot rule out that healthier people simply walk faster. They excluded anyone who died in the first two years and anyone with existing cardiovascular disease or cancer to limit that problem, but it does not vanish. The UK Biobank is also unrepresentative: a 5.5% response rate, skewing healthier and more affluent than the UK as a whole.[1]
And walking is not a complete programme. Step-based metrics miss cycling and swimming entirely, and they say nothing about the muscle loss that accelerates after 50, which needs resistance work. Walking briskly is one of the highest-yield, lowest-friction longevity habits available — it is not the only one.
Still, the shift in framing is worth taking seriously. For years the advice has been a volume target that most people miss and then feel defeated by. The data now say something kinder and more actionable: it is not how far you went today. It is whether, for half an hour of it, you meant it.
Common questions
How many steps per minute should I walk to live longer?
The lowest mortality risk sat near 100 steps per minute among people taking 7,500–10,000 steps a day (HR 0.57, 95% CI 0.47–0.69). That is a brisk, purposeful pace. Remember it is your fastest 30 minutes that counts, not your average all day — so short bursts of quick walking accumulate toward the same target.
Do I still need 10,000 steps a day?
No. In this analysis the lowest risk band was 7,500–10,000, not above 10,000, and the 2025 Lancet meta-analysis found the curve already flattening around 5,000–7,000. Extra steps are not harmful; the additional return just gets small. If you want a rough sense of where your overall fitness sits, the fitness age calculator is a reasonable starting point.
Can walking faster make up for walking less?
Partly — and this is where the evidence is thinnest. Under 5,000 steps a day, a cadence of 80 steps per minute was associated with a hazard ratio of 0.72 (0.54–0.97), which is a real signal. But push to 90 and the interval spans 0.32–1.32, because almost nobody in the cohort combines very low volume with high cadence. Treat "just walk fast" as promising rather than proven, and the moderate-steps-brisk-pace combination as the safer bet.