Abstract P268: Assessing Physical Inactivity Using Accelerometers in Comparison to International Physical Activity Questionnaire

Abstract

Introduction: Among the instruments to assess the level of physical activity in daily life (LPADL), questionnaires are cost advantageous and accelerometers are more accurate. Recent studies have shown poor agreement between these methods. Hypothesis: We tested the hypothesis that the combination of the questionnaire and the accelerometer is the best strategy to assess the prevalence of physical inactivity in adults. Objective: To evaluate and compare the prevalence of physical inactivity in adults, identified by the International Physical Activity Questionnaire (IPAQ), by triaxial accelerometry and by the combination of both. Methods: Two hundred and fifty-one participants older than 18 years were enrolled. After obtaining clinical, demographic and anthropometric data, participants underwent the following assessments: spirometry, cardiopulmonary exercise testing, body composition (bioimpedance), isokinetic muscle function, balance (force platform), and six-minute walk test. Participants who obtained the total score < 600 MET-min/wk were considered physically inactive using IPAQ. Those that performed < 150 min/wk of moderate to vigorous physical activity were considered physically inactive in the accelerometer-based method. In the combined method we considered physically inactive those who presented the IPAQ and/or the accelerometer-based criteria. Additionally, for participants who reported practicing aquatic, martial arts or cycling, only the IPAQ total score was considered. We compared the prevalence of physical inactivity and agreement between the methods. Three multivariate logistic regression models for each one of the methods were developed and predictors were mutualy compared. Results: The prevalence of physical inactivity was significantly different between the methods (IPAQ = 10%; accelerometry = 20%, and combined method = 25%). The agreement between IPAQ and accelerometry was poor (kappa = 0.152, p = 0.01). The main predictors using the IPAQ were age, fat mass, family history of cardiovascular disease, dyslipidemia, and obesity. The determinants using accelerometry were age, sex, lean body mass (LBM), family history of cardiovascular disease and smoking. The combined method selected age, sex, LBM, family history of cardiovascular disease, dyslipidemia, obesity, smoking, peak V’O2 and balance. With the exception of dyslipidemia and family history of cardiovascular disease, the combined method showed better odds ratio values. Conclusion: The combination of the IPAQ and accelerometry to determine physical inactivity was more valid when compared to the aforementioned instruments separately. The prevalence of physical inactivity and its predictors were more consistent in the combined approach in the present study. Our results suggest that the most popular methods for assessing LPADL in epidemiological studies are complementary.