Self-reported assessment of physical activity (PA) is commonly used in public health research. The present study investigated the concordance of self-reported PA assessed using the global physical activity questionnaire (GPAQ) and two different measurement approaches. Participants (n = 307, aged 30-75 years with hypertension) were recruited from a rural area in Bangladesh. We analyzed the difference between the World Health Organization (WHO) recommendations of more than 600 metabolic-equivalent time-minutes (MET-min) and the self-reported active hours, at least 2.5 h per week. Tests of sensitivity and specificity were conducted to determine concordance between the two measures. According to the WHO criteria, 255 (83%) participants were active more than 600 MET-min per week and 172 (56%) people were physically active 2.5 h or more per week, indicating a 27% difference in self-reported PA. The sensitivity, specificity, positive and negative predictive values and concordance between the two measures were 64%, 92%, 98%, 34% and 70%, respectively. Considering the WHO MET-min as the appropriate measure, 89 (35%) were false negative (FN). Older age, professionals and businesspersons were associated with a higher proportion of FN. There is a gap between self-reported PA, thus a better estimate of PA may result from combining two criteria to measure PA levels.
Keywords: Bangladesh; difference; rural area; self-reported physical activity; sensitivity and specificity.