The impact of machine learning on physical activity-related health outcomes: A systematic review and meta-analysis.
Ezgi Hasret Kozan Cikirikci, Melek Nihal Esin
PMID 40353307WHAT IT FOUND
Adults in physical activity programmes that used machine learning took more steps per day than control groups across five randomised trials.
The programmes varied, so do not assume any single app or nurse-led plan will reproduce this.
Key findings
01Across five pooled randomised trials, physical activity programmes that used machine learning increased mean daily step count compared with control groups.
02One included trial reported no significant difference in average daily steps between the machine learning intervention group and the control group.
03The included trials used machine learning in different roles, including intervention delivery, data collection, or prediction of post-intervention frailty.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
Only five trials were pooled for daily step count, and six trials reported that outcome, so the evidence base is small. The trials used machine learning in different ways, including delivery, data collection, or prediction, so the effect of any single machine-learning component cannot be separated. Some trials had unclear or high risk of selection, performance, or detection bias. Participants varied by age and health status, including older adults, college students, people with overweight or obesity, and people with minimal physical activity, so results may not apply to all patients. Only English and Turkish studies were included, which may have missed relevant trials.
Declared interests
The authors declared no conflicts of interest and reported no specific funding.
The easy way to misread this
Do not conclude that machine learning by itself caused the step-count increase. The trials combined machine learning with other intervention components, and machine learning was used in different roles, so its separate contribution cannot be isolated.