Reproducibility and Validity of the Shanghai Women's Health Study Physical Activity Questionnaire

Charles E. Matthews,Xiao-Ou Shu,Gong Yang,Fan Jin,Barbara E. Ainsworth,Dake Liu,Yu-Tang Gao,Wei Zheng
DOI: https://doi.org/10.1093/aje/kwg255
2003-01-01
American Journal of Epidemiology
Abstract:In this investigation, the authors evaluated the reproducibility and validity of the Shanghai Women's Health Study (SWHS) physical activity questionnaire (PAQ), which was administered in a cohort study of approximately 75,000 Chinese women aged 40-70 years. Reproducibility (2-year test-retest) was evaluated using kappa statistics and intraclass correlation coefficients (ICCs). Validity was evaluated by comparing Spearman correlations (r) for the SWHS PAQ with two criterion measures administered over a period of 12 months: four 7day physical activity logs and up to 28 7-day PAQs. Women were recruited from the SWHS cohort (n = 200). Results indicated that the reproducibility of adolescent and adult exercise participation (kappa = 0.85 and kappa = 0.64, respectively) and years of adolescent exercise and adult exercise energy expenditure (ICC = 0.83 and ICC = 0.70, respectively) was reasonable. Reproducibility values for adult lifestyle activities were lower (ICC = 0.14-0.54). Significant correlations between the PAQ and criterion measures of adult exercise were observed for the first PAQ administration (physical activity log, r= 0.50; 7-day PAQ, r= 0.62) and the second PAQ administration (physical activity log, r= 0.74; 7-day PAQ, r= 0.80). Significant correlations between PAO lifestyle activities and the 7-day PAO were also noted (r = 0.33-0.88). These data indicate that the SWHS PAO is a reproducible and valid measure of exercise behaviors and that it demonstrates utility in stratifying women by levels of important lifestyle activities (e.g., housework, walking, cycling).
What problem does this paper attempt to address?