Objective Measures of Activity in the Elderly: Distribution and Associations With Demographic and Health Factors.

Participants who agreed to wear the accelerometer were more often men and less often disabled than those who did not agree, so the sample may not represent all older adults. Participants attending the Rotterdam Study visits may have had better health than the general older population. Wrist-worn accelerometers underestimated activities mainly performed with the legs, such as cycling, and could not distinguish sitting, standing and lying down. The raw accelerometer cut-points used to define activity intensity are relatively new and not firmly established. The sleep algorithm may have counted motionless awake time as sleep, which could affect estimates of sedentary time. Information on cancer, cardiovascular disease and diabetes was not up to date, so some residual confounding may remain.

Do not conclude that an activity programme will raise moderate-to-vigorous activity or reduce sedentary time in older adults. This study measured existing activity and health factors, and it did not test an intervention.