PFLPhysiological Fitness Landscape

Biomarkers · Functional Fitness

Daily Step Count

Higher values go with lower risk of death. Not measured in our NHANES panel. The curve is drawn from the published study (source below).

Daily Step Count and risk of death

Published curve higher than the NHANES fitNHANES fit higher

Risk of death relative to the curve's reference value, drawn from the published study (source below).

Published risk of death0.88×per 1000 steps per day increase (95% CI 0.83–0.93) · Dose-response meta-analysis of 7 prospective cohorts (device-measured steps, general population; 175,370 person-years)

What studies found when this marker was changed

They describe what happened in those study groups, not a recommendation. See all catalogued interventions.

Measured changes in Daily Step Count, study by study

4 effects with a numeric size, each from one study arm: the dot is the reported change, the line its 95% interval where one was given. Green points move the marker the way lower risk of death lies; amber the other way. Study results in the groups studied, not predictions for a person.

Each row is one study: what was changed, the effect it reported on this marker, and the source.
TypeInterventionReported effectEvidenceSource
behavioral12-week step-activity monitoring behavioural intervention with daily feedback, skilled coaching and progressive step goals (2-3 sessions/week) in people >=6 months after strokeSteps/day increased by 1542 (95% CI 1014-2069) from pre- to post-intervention; high-intensity walking alone 406 (95% CI -63 to 876, not significant); combined 1307 (95% CI 752-1861)RCTThompson ED, et al. Stroke. 2024. Increasing Activity After Stroke: A…
behavioral6-month lifestyle walking intervention: Garmin vivofit activity tracker plus monthly nurse telephone counselling (self-monitoring, goal-setting, action planning) in patients with HFrEFDaily step count increased by 1420 (95% CI 749 to 2091) over control at 6 months (about 25%), without improvement in 6-minute walk distanceRCTVetrovsky T, et al. Circulation. 2024. Lifestyle Walking Intervention for…
behavioralBehavioural physical-activity interventions that include motivational interviewing, vs comparators without it, in adults (97 RCTs, 27,811 participants)Increases total physical activity by SMD 0.45 (95% CI 0.33 to 0.65), reported as equivalent to 1323 extra steps/day (low certainty)meta-analysisZhu S, et al. BMJ (Clinical research ed.). 2024. Effectiveness of behavioural…
dietDrinking caffeinated coffee on randomly assigned days vs avoiding caffeine, over 14 days in ambulatory adults (case-crossover trial)10,646 vs 9665 daily steps on coffee vs no-caffeine days (mean difference 1058; 95% CI 441 to 1675); nightly sleep 36 min shorterRCTMarcus GM, et al. The New England journal of medicine. 2023. Acute Effects of…

Day-to-day noise on a wearable

How much a single day's daily steps from a consumer wearable moves around a person's own level, from months of free-living data in open datasets (same measure). ICC is the share of variation that is between people: near 1, one day tells people apart; near 0, it is mostly noise. The smallest detectable change (95%) is the change a single day, or a 7-day mean, must exceed before it is more than noise. Days to average assumes days are independent; for resting heart rate they are not, so the true number is larger.

Each row is one open dataset: how much a single day's reading varies within a person, and how many days to average.
DatasetPeople · person-daysWithin-person SDICC: one day · 7-day meanDetectable change: one day · 7-day meanDays to average
LifeSnaps67 · 4,5334,458 steps/day0.30 · 0.7512,358 · 4,671153 for 1,000 steps/day
PMData16 · 2,1074,800 steps/day0.40 · 0.8213,305 · 5,029178 for 1,000 steps/day