Search

Deteriorating physical activity (PA) accumulation may be an indicator of premature mortality. In 698 older adults, Cox proportional hazard models were fitted to determine the interaction effect between age and three different PA metrics (total PA minutes; fragmentation defined as the # contiguous PA minutes (bouts)/total PA minutes; and proportion of time spent in

We review existing methods for quantifying fragmentation of accelerometry-measured physical activity and develop a unifying statistical framework that connects all methods by separately modeling switching between and duration of sedentary and active bouts. We analyzed the association of these metrics with mortality in National Health and Nutrition Examination Survey 2003–2006. We have identified that “Intradaily Variability” (1SD HR=1.32, 95%CI 1.18–1.48), “Active Bouts Fragmentation Index” (1SD HR=1.43, 95%CI 1.25–1.64), “Sedentary Bouts Gini Index” (1SD HR=1.15, 95%CI 1.06–1.25), and hazard rates of distribution function of durations of active and sedentary bouts (1SD HR=1.93, 1.40, 95% CI 1.59–2.34 and 1.23–1.59) independently predict mortality after adjusting for total sedentary time, total daily activity, age, sex, education, medical comorbidities, smoking status, alcohol consumption, and mobility limitations. Future studies should concentrate on developing metrics that jointly model the frequency of switching between the sedentary and active bouts as well as the bout durations.

Low daily physical activity has been linked to poor health and functional performance with aging. However, the health implications of the daily patterns/temporal accumulation of physical activity remain unknown. We created an activity fragmentation index (AFI) to account for the number and duration of daily active bouts using 7-day accelerometry data from the Baltimore Longitudinal Study of Aging (n = 734, mean age 67.8 ± 13.2 years, 50% female). Minute-epoch activity counts were dichotomized into active or sedentary states, and summarized according to the distribution of active bout durations. AFI was calculated as the reciprocal of the average duration of the active bouts. Total daily volume of physical activity was summarized using total log-transformed activity counts (TLAC). Using multiple linear regression models adjusted for age, sex, height, weight, TLAC, and disease conditions, AFI was negatively associated with gait speed (β = -0.06 m/s, p < 0.004), and positively associated with laboratory assessments of high fatigability (p < 0.001). TLAC was a significant predictor of gait speed and fatigability until AFI was added to the model, suggesting that the duration and number of active bouts throughout the day mediates the associations among TLAC, gait speed, and fatigability. Together, these results imply that accounting for fragmented daily physical activity may be more important than total volume of physical activity when assessing functional performance in older adults. Further research is needed to validate this measure and determine the longitudinal value of activity fragmentation in predicting adverse outcomes in older adults.