To measure clinical walking speed, participants completed a standard walking test.
The median walking cadence during daily life was about 75 steps per minute, and the 95th percentile walking cadence was about 112 steps per minute.
The study found only a weak correlation between clinical walking speed and free-living walking cadence, meaning the two measures do not strongly reflect the same walking behavior.
Combining clinical walking speed with free-living walking parameters provided a small boost in predictive ability, but the effect was not significant after accounting for detailed health factors.
Overall, this study suggests that clinical walking speed tests and free-living walking cadence measurements from wrist-worn devices are both useful in predicting all-cause mortality risk.
The findings, published on October 23 in JAMA Network Open , could help doctors and researchers choose the best method for assessing health risks in different settings.
Both measures were linked to mortality risk and showed similar ability to predict it.
The study compared two ways of measuring walking: a short walking speed test conducted in a clinic and the step rate people take in their everyday lives, measured with a wrist-worn device.
New research shows that how fast you walk, whether measured in a clinical test or through daily activity, can help predict your risk of dying from any cause.
The researchers additionally conducted analyses to test how well the results held up when they excluded early deaths, looked at subgroups by age and health status, and used different definitions of walking cadence.
The study also examined differences between males and females.
The ability of walking measures to predict mortality was assessed using a statistical measure.
The models were adjusted step-by-step: starting with no adjustments, then adding sociodemographic factors, behavioral factors, and finally health-related factors.
Statistical methods were then used to estimate the risk of death associated with walking speed and cadence while accounting for other risk factors.
National death records were used to identify deaths during the follow-up period.
The researchers additionally gathered information on factors known to affect mortality, such as:
From this data, two key measures of walking cadence were calculated: the median walking cadence (the middle value of steps per minute during walking periods) and the 95th-percentile walking cadence (a higher step rate indicating brisk walking ). The study also calculated the average number of daily steps.
For free-living walking measurements, participants wore a wrist-worn GENEActiv accelerometer for 9 consecutive days. This device recorded movement at a high frequency and used special software to detect walking episodes and count steps.
To measure clinical walking speed, participants completed a standard walking test. They walked a marked 2.44-meter course at their usual pace. The test was repeated three times, and the average speed in meters per second was used for analysis. Participants who normally used a cane were allowed to use it during the test.
The researchers collected data from January 2012 to December 2013 and followed up through August 2024.
It included 4,067 participants, mostly males, with an average age of about 69 years. About one-third of the subjects were female.
The study was part of the Whitehall II accelerometer substudy in the United Kingdom.
During a median follow-up of nearly 12 years, 739 participants died.
The average clinical walking speed of participants was 1.2 meters per second. The median walking cadence during daily life was about 75 steps per minute, and the 95th percentile walking cadence was about 112 steps per minute.
The study found only a weak correlation between clinical walking speed and free-living walking cadence, meaning the two measures do not strongly reflect the same walking behavior. However, both measures were independently associated with a lower risk of death.
After adjusting for age, sex, lifestyle, and health factors, for every standard increase in clinical walking speed, the risk of death was 23% lower.
Similarly, higher median and 95th-percentile walking cadences were associated with 17% and 18% lower risks of death, respectively.
When examining how well these walking measures predicted mortality, each on its own showed modest predictive ability.
Adding clinical walking speed to a model with other known risk factors slightly improved mortality prediction, while adding free-living walking cadence also improved prediction, albeit to a similar degree to that of clinical walking speed.
Combining clinical walking speed with free-living walking parameters provided a small boost in predictive ability, but the effect was not significant after accounting for detailed health factors.
The study also found that walking speed and cadence had similar predictive power for both men and women, although the overall prediction models performed slightly better for women.
Several sensitivity analyses supported the main findings. Results were consistent when excluding early deaths, focusing on healthier or older participants, and using different ways to define walking cadence.
Overall, this study suggests that clinical walking speed tests and free-living walking cadence measurements from wrist-worn devices are both useful in predicting all-cause mortality risk. They provide similar information but reflect different aspects of walking behavior.
