153 Exploring the association between exercise adherence and recovery of low back pain factors, resulting in two models per outcome. To explore potential confounders, the association between the outcome and adherence was estimated with and without the potential confounders in the model. When the estimate of the association changed by more than 10%, the variable was added to the final model as a confounder. Furthermore, treatment group allocation from the parent trial was always included in the final model to control for the influence of the e-Exercise LBP intervention on adherence and outcomes (19). For each regression model, the assumptions of linearity, homoscedasticity, independence and normality were checked and confirmed. Multicollinearity was assessed for the final models but was not found. RESULTS A total of 208 participants were included in the parent trial (19). EXAS scores and trajectory of adherence class allocation were available for 173 participants who received two or more treatment sessions. The data were collected during a median of 4.0 treatment sessions [IQR 3.0, 6.0]. Missing data were caused by incomplete case reports forms or the absence of case reports forms from the participating physical therapists, and loss to follow up (14 patients). Demographic characteristics of the included patients can be found in Table 1. After three months, forty-five (28.5%) patients considered themselves to have recovered from LBP after a median of 4 [IQR 3, 6] treatment sessions. The median changes in the ODI and NPRS were -8 [IQR -20, -1] and -2 [IQR -4, -0.5], respectively. The mean EXAS score for all patients was 59.2 (SD 25.3), with 46.0 (SD 19.4) for the “declining adherence” class, 81.0 (SD 12.4) for the “stable adherence” class, and 39.9 for the “increasing adherence” class (SD 25.3). 7
RkJQdWJsaXNoZXIy MTk4NDMw