Associação e concordância entre medidas de classificação do nível de atividade física em idosos comunitários obtidas pelo IPAQ e por um algoritmo desenvolvido através de dados de sensores de smartwatches

Abstract

Introduction: Measurements of physical activity in the elderly population are important because they can provide indicators for assessing health status and screening behavior. The use of questionnaires to measure the level of physical activity is problematic due to the imprecision of the information provided and the susceptibility to recording or memory bias, which can lead to overestimated or underestimated results. Contrary to this situation, the use of wearable technologies for detecting movement makes it possible to objectively quantify the level of physical activity with greater precision. Objective: To verify the presence of associations and agreement between the measures used to classify the level of physical activity in community-dwelling elderly people obtained by the International Physical Activity Questionnaire (IPAQ) and by an algorithm developed using data from smartwatch sensors. Materials and Methods: This is an observational, cross-sectional study carried out at the Cidade Madura Condominium and the Centro de Convivência do Idoso in Campina Grande-PB. An algorithm was developed in the Phyton programming language to identify active minutes and generate classifications of physical activity levels based on data from Fitbit smartwatches. The sample consisted of 56 elderly people assessed using the Fitbit, which monitored physical activity for 7 days, and then the IPAQ. The level of physical activity was classified using three different assessment methods (IPAQ, algorithm and step index). Results: A moderate correlation was observed between the data on active minutes spent on "light activities" identified by the IPAQ and the algorithm (ρ=0.331; p<0.05). There was no significant correlation between the data on total active minutes and minutes spent on moderate and vigorous activities. For the classification of the level of physical activity by the three instruments, no association and agreement was found between the IPAQ and the algorithm developed in this study. However, there was a significant association and agreement between the algorithm and the step index (κ=0.250; p<0.05). Conclusion: As it is a device that is easy to use and accept, as it provides more objective measures and allows the user to be monitored in their free-living environment, the Fitbit smartwatch proved to be a viable tool that can be used to assess aspects of physical activity in community-dwelling elderly people; however, caution is advised when using data from this device to assess physical activity levels.