The Validity of a Physical Activity Questionnaire in Rural Hispanic Diabetics

This study was designed to investigate the ability of the International Physical Activity Questionnaire (IPAQ) to measure physical activity (PA) among a group of Hispanics living in the Yakima valley, a rural area of Washington State. A random sample of sixty participants (mean age: 49.9 ± 11 years, 81% female, 78% low acculturation) was taken from an existing randomized intervention trial for diabetics.  Participants wore an accelerometer on their waist for seven days and completed both the long (IPAQ-L) and short (IPAQ-S) versions of the IPAQ the next day. Pearson correlation coefficients, the sensitivity and specificity of the IPAQ to identify individuals meeting or not meeting the national guidelines for PA, and Bland-Altman plots were employed to examine the criterion validity, classification validity, and agreement between the two measures, respectively. All comparisons were done in metabolic equivalent (MET) minutes/week. Correlation coefficients for total PA were poor for both the IPAQ-L (0.14) and the IPAQ-S (0.15). Specificity and sensitivity for the IPAQ-L was acceptable (66.7% and 76.9%, respectively). The IPAQ-S had higher specificity but lower sensitivity (100% and 36.5%, respectively). Results from the Bland-Altman plots suggest a reasonable agreement between the IPAQ-L and the accelerometer in total measured PA, but only at lower-intensity levels. The results of this study suggest that both versions of the IPAQ have poor criterion validity in measuring PA in a population of rural Hispanics; however, the IPAQ may be an acceptable tool to measure whether or not individuals meet the national guidelines for physical activity.