Validity and reliability of the Baecke questionnaire against accelerometer-measured physical activity in community dwelling adults according to educational level.

Baecke questionnaire have been widely used to assess physical activity. However, the role of educational level on validity and reliability of Baecke questionnaire is still not stablished, being a factor that can potentially influence the accuracy of self-reported measures. The present study aimed to verify the validity and reliability of Baecke questionnaire for the measurement of physical activity in community dwelling adults according to education level. The sample included 251 adults (42.4±17.0 years, 55% of women). Physical activity was self-reported by Baecke questionnaire and objectively measured by accelerometer. The education level (EL) was classified by years of study into low (<8 years), medium (8-11 years) and high (>11 years). A 7-day test-retest reliability was analyzed by intraclass correlation coefficient. The relationship, agreement and validity of the Baecke questionnaire against accelerometry were analyzed by Spearman's correlation, Kappa index, and ROC curve, respectively. The reliability of Baecke questionnaire were r = 0.97 (high EL), r = 0.78 (medium EL), and r = 0.68 (low EL). Sensitivity and specificity were 77% and 71% in high EL, 54% and 80% in medium EL and 33% and 89% in the low EL. Baecke questionnaire proved to be reliable and a valid measurement of habitual physical activity in adults with medium and high EL.