To ensure that the variation in physical activity in older populations was captured [22], participants were excluded if there were less than 4 valid accelerometer days (N = 11) or >1 days in-between consecutive measurement days (N = 1) [18]. A valid day was defined as a day with 10 hours of accelerometer wear time. Few days were omitted due to technical errors of
the accelerometer (N = 3) or loss of the accelerometer in the mail (N = 1) [18]. In total, data of 174 participants were analyzed.