Supplement prediction model 125 stronger self-efficacy and outcome expectations. In turn, self-efficacy and outcome expectations are highly correlated with patient adherence to treatment [6]. Although studies showed that the TPB variables can predict exercise adherence, the explained variance in adherence is far from satisfying [1, 7-9]. Therefore, to reexamine the potentially causal psychological predictors of exercise adherence in patients with COPD, this study adopted the TPB in combination with the alliance concept. In addition, more traditional variables as reported in previous studies were included to gather knowledge on their role, and their benefit to explain adherence [3]. The aim of this supplement was to examine the predictors of exercise adherence in patients with COPD. Insights in these predictors can provide useful information for the implementation of psychological interventions to improve exercise adherence rates. Methods Statistical analyses Data were analyzed using R version 4.0.3. Baseline characteristics are presented with appropriate measures of central tendency and dispersion. For the main analyses first, Spearman’s correlations between each of the potential predictors and exercise adherence were computed. Correlations of 0.10, 0.30, and 0.50 are considered small, medium, and large effect sizes, respectively, in the behavioral sciences [10]. Variables that had significant (p < 0.05) correlations with exercise adherence were retained for a hierarchical linear multiple regression analysis (HMRA). Second, HMRA were performed. Linear as opposed to logistic regression analysis was used, because exercise adherence was viewed as a continuous outcome rather than a dichotomous outcome. Linear regression analysis produces explained variances (R2) and standardized regression coefficients (b). R2 values of 0.02, 0.15, and 0.35 are considered small, medium, and large effect sizes, respectively [10]. b’s are interpreted only relative to other b’s in the regression equation and can be considered on a ratio level scale. HMRA were conducted using stepwise regression within theoretically based blocks of variables. The order and the content of the steps were based on previous research using TPB in the exercise context and the tenets of TPB [11]. For each analysis, the order of the blocks was as follows: 1. theory of planned behavior constructs (first attitude, PBC, and subjective norm; second the preceding variables with intention), 2. psychological constructs (alliance), 3. past exercise, and fitness (exercise in history, physiotherapy in history, years of physiotherapy, days of moderate and vigorous physical activity) 4. disease variables (MRC dyspnea score, depression), and 5. demographics (gender, age, country, education, smoking
RkJQdWJsaXNoZXIy MjY0ODMw