Validity of a questionnaire to assess historical physical activity in older women

PURPOSE: To compare historical physical activity recall to original physical activity questionnaires collected at four time points over a 17-yr period in postmenopausal women. METHODS: This study examined the recall of physical activity (PA) data collected as part of a clinical trial of a walking intervention and subsequent follow-up in 163 white postmenopausal women (74 +/- 4 yr). Physical activity levels were measured with a modified version of the Paffenbarger Physical Activity Questionnaire (1982, 1985, 1995, 1999) throughout the duration of the study. The interviewer-administered historical physical activity questionnaire (HPAQ) asked participants to recall in the year 2000 what PA they performed in 1982, 1985, 1995, and over the past year. Spearman correlation coefficients were used to compare subjects' historical recall of activity for each time period with the actual questionnaire data collected during that time period. RESULTS: Statistically significant correlations were found between the historical PA recall and the original PA questionnaires at each of the time points measured (rho = 0.39-0.62, P < 0.0001). The magnitude of the correlations increased as the time length of recall decreased. CONCLUSION: It appears that historical recall of leisure physical activity can be reasonably estimated by questionnaire over a substantial time period in older women.