Step Count as a Digital Biomarker

Step count is the number of steps a person takes, usually summarised per day. It is the most widely used measure of free-living physical activity in clinical research, because walking is the dominant form of daily movement and steps are simple to count, simple to explain and comparable across populations. The evidence behind it is unusually deep: in a pooled analysis of 15 international cohorts covering more than 47,000 adults, higher daily step counts were associated with progressively lower all-cause mortality, with the benefit levelling off at around 6,000 to 8,000 steps per day in older adults and 8,000 to 10,000 in younger adults. As a digital biomarker, step count turns a construct that questionnaires estimate roughly into an objective daily record. In clinical populations it serves as a global indicator of functional capacity, disease burden and recovery rather than a fitness statistic.

Nearly all devices count steps from accelerometer data. An algorithm detects the rhythmic acceleration pattern of walking, filters out non-walking movement and accumulates detected steps into daily totals. Research-grade actigraphs such as ActiGraph record raw accelerations processed with published, citable algorithms, while consumer devices from Apple, Garmin, Fitbit and Samsung apply proprietary on-device algorithms to the same underlying signal. Wear location matters: wrist-worn devices are convenient but also register arm movement, while waist and thigh placements track ambulation more directly. Validation studies compare device counts against manually counted steps on video or treadmill protocols. Because daily totals swing with weather, work patterns and weekends, protocols typically require a minimum daily wear time, commonly 10 waking hours, and average across at least four to seven valid days.

Step count appears across therapeutic areas as a monitoring, secondary and increasingly primary endpoint. In COPD it tracks the physical inactivity that predicts hospitalisation and mortality, and a semiautomated telecoaching programme raised daily steps in a randomised trial. In heart failure, multiple sclerosis and Parkinson's disease it captures whether treatment effects reach everyday life, and after stroke or surgery it charts recovery in a way clinic visits cannot. Step endpoints appear in more registered wearable trials than any other activity measure in the public DiMe endpoint library, including as primary outcomes. For interpretation, studies in COPD treat a change of roughly 600 to 1,100 steps per day as clinically meaningful. Trials usually pair step count with patient-reported activity questionnaires, so objective movement and perceived capacity can be read together.

Step counting degrades at the slow, shuffling gait speeds common in frail and neurological populations, where devices can miss a large share of true steps. Wrist wear inflates counts through arm movement, so totals are not directly comparable across wear locations or vendors, and algorithm updates can shift counts mid-study. A step total also says nothing about intensity on its own. Finally, being measured changes behaviour: early wear periods often show inflated activity that settles once the novelty fades.

• Paluch AE, et al. Daily steps and all-cause mortality: a meta-analysis of 15 international cohorts. Lancet Public Health. 2022. pubmed.ncbi.nlm.nih.gov
• Saint-Maurice PF, et al. Association of daily step count and step intensity with mortality among US adults. JAMA. 2020. pubmed.ncbi.nlm.nih.gov
• Demeyer H, et al. The minimal important difference in physical activity in patients with COPD. PLoS One. 2016. pubmed.ncbi.nlm.nih.gov
• Bassett DR, et al. Step counting: a review of measurement considerations and health-related applications. Sports Med. 2017. pubmed.ncbi.nlm.nih.gov
• Demeyer H, et al. Physical activity is increased by a 12-week semiautomated telecoaching programme in patients with COPD. Thorax. 2017. pubmed.ncbi.nlm.nih.gov