Development and testing of a short physical activity recall questionnaire

Purpose: To develop and test two different short telephone activity recall (STAR) questionnaires, one with closed-ended and the other with open-ended response options, that assessed overall moderate and vigorous activity in a usual week. Methods: One hundred four participants completed a 3-d test-retest study, and 88 participants completed 10-14, 24-h physical activity recalls (24PAR) and at least 7 d of objective physical activity monitoring by Actigraph during a 28-d period. Results: Consistency of classification from one administration to the next was high (65-92%), extreme inconsistencies between reports were infrequent (0-7%), and kappa values were between 0.50 and 0.75. Correlations between self-reports and criterion measures for moderate-intensity duration were between 0.30 and 0.40. Agreement between the instruments and the 24PAR for meeting the moderate or vigorous recommendations was between 60 and 70%. For the 24PAR comparisons, kappa values tended to be higher for women than men, but were of only modest strength (kappa 0.40). With the 24PAR as criterion, sensitivity of the self-report instruments was between 50 and 90%, and specificity was between 63 and 84%. Kappa values comparing the instruments with the Actigraph were low (< 0.20). Overall classification by the short instruments into meeting the recommendations was associated with higher levels of total 24PAR activity (P <= 0.01) as well as greater steps per day and counts per minute per day from the Actigraph (P <= 0.08). The open-ended instrument appeared to perform better for moderate-intensity activity, whereas the closed-ended item appeared to perform better for vigorous activity. Conclusion: The evaluated instruments had reasonable reliability and demonstrated an ability to capture important differences in overall physical activity patterns in this population, although individual classification errors were substantial.