2 Physical activity as a risk or protective factor | 23 Sensitivity analyses Two sensitivity analyses were performed to test for the robustness of our findings in a larger population with less specific data. In a first sensitivity analysis, the LASA Physical Activity Questionnaire (LAPAQ) was used to assess physical activity (108). This allowed for a sensitivity analysis on 505 subjects participating in the LASA 75 PLUS study of whom LAPAQ data was available, and thus on a larger power compared to the primary analysis using inertial sensor data. The LAPAQ subjectively assesses the frequency and duration of activities in the past 14 days. The activities included were walking, cycling, heavy household work and a first and second sport when applicable. The frequency of the activity was multiplied by the duration of the activity in minutes per day and then divided by 14 days (frequency*duration/14). In a second sensitivity analysis, binary logistic regression analyses were performed in an even larger study population of 1,002 participants of 65 years or older to further increase the power of the analyses and test for robustness of our primary findings. These regression analyses included the same four models as in the primary analyses. For this cross-sectional study, data were used from T1 (2015/2016) and T5 (2018/2019). This study population consisted of participants aged 65 years and older and was larger than the primary study population, because falls and fractures were retrospectively asked over the past 3 years instead of determined every nine months. In this population, physical activity was determined at baseline by an Actigraph tri-axial inertial sensor (Model GT3X; Actigraph, Pensacola, FL, USA).
RkJQdWJsaXNoZXIy MjY0ODMw