Walking studies usually count steps. A cohort published on 13 November in the Journal of Global Health took the same seven days of wrist-worn accelerometer data that most step studies use and sorted the steps by cadence instead: steps taken at 40 per minute or faster were labelled purposeful walking, anything slower was labelled incidental. Then the authors followed 94,422 UK Biobank participants for a median of 6.87 years and asked which of the two categories showed up alongside fall injuries.
Only one of them did.
What “purposeful” meant in this dataset
The cut-off is not a fitness threshold. Forty steps a minute is a slow, steady walk, well below the pace most people would call brisk, and the paper does not present it as a target to hit. It works instead as a divider between sustained walking and the shuffling, standing and short indoor trips that also generate steps on an accelerometer.
The sample was drawn from the 103,660 UK Biobank participants who agreed to wear an Axivity AX3 device on the wrist for a week; the analysis kept those with at least 16 hours of wear on at least three days. Participants were 42 to 78 years old at baseline, average age 61.8, and 44.0% were men. Falls were not self-reported: the outcome was the first hospital inpatient record coded W00 to W19 in ICD-10, which means the study counted falls serious enough to reach a hospital, not every stumble at home. Hip fracture, coded S72.0 to S72.2, was the secondary outcome.
Where the risk curves separated
Among the 90,932 participants with no recorded fall before baseline, 2,985 had a first fall diagnosis during follow-up. The rate per 1,000 person-years fell steadily across the purposeful-walking groups: 7.93 in the slowest, then 4.97, 4.22 and 3.86. After adjustment for age, sex, deprivation, alcohol, smoking, body mass index, education, ethnicity, vision, bone mineral density, prior fracture, osteoporosis, osteoarthritis, dementia, frailty, diabetes, sleep duration, vitamin D and calcium status, and self-rated health, the hazard ratios against the lowest group (4,000 steps or fewer per day) were 0.75 (95% CI 0.60–0.92) for 4,000–5,999 steps, 0.65 (0.52–0.82) for 6,000–7,999 and 0.68 (0.54–0.86) for 8,000 or more.
Two details in that sequence matter more than the headline. First, it is not a straight line: the fourth group’s estimate sits slightly above the third, and the spline analysis describes a curve that flattens at higher volumes rather than continuing to fall. Second, the same pattern held for the number of falls recorded per person (relative risks 0.78, 0.74 and 0.72 across the three higher groups), and for hip fracture, where the ratios were 0.63 (0.39–1.02), 0.49 (0.29–0.83) and 0.36 (0.21–0.62). Expressed in absolute terms, the authors report a rate difference of about 3 to 4 fewer falls per 1,000 person-years in the three higher groups, and a seven-year cumulative risk difference of roughly 1.5 to 1.8 percentage points.
The steps that stopped mattering
Incidental steps looked protective at first. Once the models also adjusted for purposeful steps, the association disappeared, and it stayed absent in sensitivity analyses. The two step counts are correlated (r = 0.53), which is exactly why the authors ran the mutual adjustment: people who walk deliberately also accumulate more incidental movement, and the crude incidental signal appears to have been borrowing from the deliberate one.
What the paper does not show is that incidental movement is worthless. Standing up, tidying, walking to the kitchen — none of it was associated with fewer hospital-treated falls in this analysis, but the analysis was not built to detect small effects, and “no association after adjustment” is not the same finding as “no effect”. For a piece of the movement picture that is genuinely well evidenced, see our earlier report on walking intensity versus walking volume in frail older adults, where a supervised programme was tested rather than observed.
The caveats the authors list themselves
This is an observational cohort, and the paper says so repeatedly. Walking was measured once, over seven days, at baseline; participants may have moved more while being monitored; residual confounding and reverse causation cannot be excluded. The authors report E-values of 1.85, 2.17 and 2.00, which quantifies how strong an unmeasured confounder would have to be, on both the walking and the fall side, to explain the result away. That is a moderate degree of robustness, not an unassailable one.
Two further limits are worth spelling out. The subgroup analyses found statistically significant inverse associations only among men older than 60 who were not frail and not obese; in the corresponding comparison groups the estimates were not significant, which is a narrower result than the abstract’s framing implies. And the cohort is 96.5% White, drawn from volunteers in a single country, with falls ascertained from hospital records only — so a fall treated in an emergency department and sent home, or managed by a family doctor, never entered the data.
What falls guidance already recommends
Federal guidance approaches falls from a different direction, and it is worth holding the two side by side. The CDC describes falls as the leading cause of injury for adults 65 and older, with more than 14 million — about one in four — reporting a fall each year, and an age-adjusted fall death rate that rose 21% between 2018 and 2024, from 64.7 to 78.4 per 100,000 older adults. Its stated route to reducing that toll is screening for fall risk and then addressing the factors behind it: medicines that increase risk, and poor strength and balance.
MedlinePlus lists the same territory — balance problems, medicines that cause dizziness, vision problems, alcohol, leg weakness, low blood pressure, diabetes, neuropathy, slow reflexes, and hazards underfoot — and notes that regular exercise may lower risk by strengthening muscles, improving balance and keeping bones strong.
The physical activity guidelines themselves are about volume and type rather than cadence. Adults are advised to get at least 150 minutes a week of moderate-intensity activity such as brisk walking, and adults 65 and older to add two days of muscle-strengthening activity plus activities that improve balance. The CDC also notes that only one in four US adults fully meets the aerobic and muscle-strengthening guidelines, and that inadequate aerobic activity is associated with $192 billion a year in health care costs.
What a step target can and cannot claim
Nothing in this study establishes that walking more, or walking faster, prevents a fall. There is no trial here in which people were assigned a cadence and followed for injuries; the paper’s own conclusion is that the findings “may help inform future step-based recommendations”, which is a research agenda, not a recommendation. Nor does it identify a protective number: the curve flattens, the top group is not the lowest-risk group, and the authors do not propose a daily target.
What it does add is a reason to be careful with step counts as a single score. A day that reaches 8,000 steps through slow, fragmented movement is not the same exposure as a day that reaches 8,000 through several sustained walks, and studies that treat the two as interchangeable may be averaging away the part that matters. For anyone whose concern is falls specifically — particularly after a previous fall, a dizzy spell on standing, or a change in medication — the practical question is broader than a number, and worth raising with a clinician rather than settling with a phone. Our report on blood pressure recovery after standing covers one of the reasons why, and the site’s medical disclaimer explains the limits of what an article like this can tell you.
