29 Predictors of outcome for cognitive behaviour therapy in binge eating disorder 2 All things considered, the body of recent research on predictors of treatment outcome for BED is limited and partly contradictory. A possible explanation for the mixed findings is the relatively small sample size used in most studies. In addition, studies tend to combine data from different interventions such as guided self-help, behavioural weight loss, CBT group therapy and individual IPT for predictor-of-outcome analyses (Masheb & Grilo, 2008; Ricca et al., 2010; Wilson et al., 2010; Zunker et al., 2011). Only a few studies report on the longer-term efficacy of treatment (Castellini et al., 2011; Fichter et al., 2008; Hilbert et al., 2007; Ricca et al, 2010; Wilson et al., 2010; Zunker et al., 2011) and only one study examined the predictive value of end-of-treatment outcomes for longer-term recovery status in BED (Lock et al., 2013). Additionally, the focus of most studies is on methodologically sound but ecologically less valid randomised controlled trials (RCTs). Applied exclusion criteria in RCTs include medical conditions that might influence eating or weight (such as diabetes) and the use of psychotropic medication (such as antidepressants) (Grilo et al., 2012; Hilbert et al., 2007; Masheb & Grilo, 2008; Ricca et al., 2010; Wilson et al., 2010). However, obese BED patients are known to suffer from depression and diabetes (Finkelstein et al., 2007; Grilo et al, 2009; Telch & Stice, 1998). A naturalistic design could take these patient groups into account. To our knowledge, only three naturalistic treatment intervention studies have been conducted, with two of these using a large sample size (Castellini et al., 2011; Deumens et al., 2012; Fichter et al., 2008). The study by Deumens et al. (2012) was conducted at our treatment centre. They examined pretreatment predictors of post-treatment outcome, using a composite score of the subscales drive for thinness, bulimia and interoceptive awareness of the Eating Disorder Inventory (EDI)-I as outcome measure in 182 BED patients. They found that being in a romantic relationship and/or living with one’s parents (‘high social embedding’) and more openness to experience predicted more improvement at post-treatment. In addition, more depressive symptoms, more agoraphobia and more extraversion were significantly related to less improvement after treatment. The present study builds on the study by Deumens et al. (2012) using partially the same population. This study, however, differs from the one by Deumens and colleagues by its use of a larger patient sample (N = 304 completers), the inclusion of follow-up measures 6 months after treatment and its use of a more specific operationalisation of binge eating pathology as outcome measure (EDI bulimia scale scores instead of the composite score used by Deumens et al.). Specifically, we investigated what factors predict who will benefit from treatment in terms of binge eating pathology and what factors at end of treatment predict outcome at follow-up. In searching for these predictors, we focussed on group CBT for BED in an intensive outpatient treatment environment. Potential predictors were chosen to study seemingly unequivocal findings (levels of eating disorder pathology, body dissatisfaction, BMI and education) and to study less unequivocal findings (level of psychopathology, personality characteristics, age and social embedding).
RkJQdWJsaXNoZXIy MjY0ODMw