Algorithm measured sitting in overweight postmenopausal women

Researchers tested a sitting-detection algorithm in a randomised trial of 406 postmenopausal women aged 55 or older who were overweight. It analysed hip-worn movement-sensor data collected during everyday life. The trial had three groups: healthy living as the control, reducing sitting, and increasing breaks from sitting. Researchers compared measurements at two timepoints with a reference device and a standard method based on movement thresholds.

The algorithm’s accuracy measures exceeded 84.8%. In the reducing-sitting group, it estimated an intervention-versus-control change in daily sedentary time of 38.2 minutes, compared with 65.1 minutes for the reference device and 0.28 minutes for the standard method. In the group increasing breaks, it estimated a change of four breaks per day, compared with 22 for the reference device and fewer than one for the standard method.