This study developed algorithms for predicting activity energy expenditure (AEE) in children (n = 24) and adults (n = 24) from the Actical activity monitor. Each participant performed 10 activities (supine resting, three sitting, three house cleaning, and three locomotion) while wearing monitors on the ankle, hip, and wrist; AEE was computed from oxygen consumption. Regression analysis, used to create AEE prediction equations based on Actical output, varied considerably for both children (R2 = .45-.75; p < .001) and adults (R2 = .14-.85; p < .008). Most of the resulting algorithms accurately predicted accumulated AEE and time within light, moderate, and vigorous intensity categories (p > .05). The Actical monitor may be useful for predicting AEE and time variables at the ankle, hip, or wrist locations.

Background:The International Physical Activity Questionnaire-Short Form (IPAQ-S) has been evaluated against accelerometer-determined physical activity measures in small homogenous samples of adults in the United States. There is limited information about the validity of the IPAQ-S in diverse US samples. Methods: 142 Blacks residing in low-income housing completed the IPAQ-S and wore an accelerometer for up to 6 days. Both 1-and 10-minute accelerometer bouts were used to define time spent in light, moderate, and vigorous physical activity. Results: We found fair agreement between the IPAQ-S and accelerometerdetermined physical activity (r = .26 for 10-minute bout, r = .36 for 1-minute bout). Correlations were higher among men than women. When we classified participants as meeting physical activity recommendations, agreement was low (kappa = .04, 10-minute; kappa = .21, 1-minute); only 25% of individuals were classified the same by both instruments (10-minute bout). Conclusions: In one of the few studies to assess the validity of a self-reported physical activity measure among Blacks, we found moderate correlations with accelerometer data, though correlations were weaker for women. Correlations were smaller when IPAQ-S data were compared using a 10-versus a 1-minute bout definition. There was limited evidence for agreement between the instruments when classifying participants as meeting physical activity recommendations.

No abstract

The present results support the concept that CRF prediction models can be used to reasonably characterize the fitness level of a cohort using data that can be obtained from a questionnaire. Accordingly, predicted CRF values may be useful as an exposure variable in large epidemiologic studies in which exercise testing is not feasible.