38 Chapter 2 Post-treatment predictors A hierarchical linear regression analysis was conducted to determine post-treatment predictors for follow-up level of bulimia scale scores. All predictors were similar to the hierarchical regression analysis reported on above, with the exception that, this time, post-treatment measures were used as predictors. In addition, social embedding and level of education could not be used as predictors since these were not measured post-treatment. The final model can be seen in Table 5 and consists of 10 relevant predictors. Post-treatment bulimia scale scores, EDI interoceptive awareness, SCL-90 anxiety, SCL-90 insufficiency, and BAT lack of familiarity with one’s own body were significant positive predictors for follow-up bulimia scale scores, and NEO-PI-R conscientiousness was a marginally significant positive predictor. Lower scores on these post-treatment predictors were predictive of lower follow-up bulimia scale scores. Post-treatment EDI drive for thinness and SCL-90 hostility were negative predictors for follow-up bulimia scale scores, and SCL-90 somatic complaints and BMI were marginally significant negative predictors. Lower scores on these post-treatment predictors lead to higher follow-up bulimia scale scores4,5. See Table 5 for an overview of the final model and for test statistics for both the model and the predictors. 4 We also checked whether patients had received additional treatment (either within our clinic or elsewhere). This information was dummy coded and added as a predictor [additional treatment no/yes (0,1)] in the analysis. This led to exactly the same results, from which we can conclude that additional treatment was not a relevant factor in follow-up outcome. 5 When repeating the analysis with only female patients (n = 169), EDI interoceptive awareness, SCL90 somatic complaints, SCL-90 hostility and NEO-PI-R conscientiousness no longer reached the predetermined p < .10 criterion and were therefore excluded from the final model. One new predictor did reach the p < .10 criterion and was therefore included in the model: SCL-90 total score, which was a negative predictor. All other predictors showed results similar to those reported above (in the data on men and women).
RkJQdWJsaXNoZXIy MjY0ODMw