Field evaluation of the Computer Science and Applications, Inc. physical activity monitor

The purpose of this study was to evaluate the Computer Science and Applications, Inc. (CSA) activity monitor to quantify physical activity in free living subjects using an activity diary as the criterion measure.

Subjects also completed a 7-d physical activity recall at the end of the monitoring period. Nine male and 10 female subjects (mean, SD) (25.0, 3.6 yr) wore the CSA monitor for 7 consecutive days. On 3 of those days, subjects completed an activity diary (2 weekdays and 1 weekend day). Total kcal per day (Dkcal(tot)) was calculated from the self-reported diary classifications of the subject's activities. For the 3 days that coincided with the diary, total number of counts accumulated per day (cnt(tot)) was obtained from the monitor.

The amount of activity per day recorded by the CSA monitor followed the same pattern of change as the activity diary. The cnt(tot) and Dkcal(tot) were significantly (P < or = 0.05) correlated on day 1 (r = 0.65), day 2 (r = 0.49), day 3 (r = 0.55), and for the 3 days pooled (r = 0.51). Subjects were classified as low, moderate, or highly active based on tertiles of kcal from the diary and counts from the CSA monitor. The percentage agreement between the CSA and the activity diary was 68.4% (Kappa = 0.53, P < 0.01). The number of minutes spent in the various activity categories were compared between instruments using an ANOVA model. The results of these analyses suggest that the CSA overestimated light activity and underestimated vigorous activity compared with the diary.

In conclusion, the CSA monitor may be useful in a field situation where total physical activity and patterns of physical activity are the desired outcomes.