'I just stopped going' : a mixed methods investigation into dropout from psychological treatment in adolescents with depression

While the effectiveness of psychological treatment for adolescent depression is well established, there is concern about high dropout rates. Using a mixed-methods design, this thesis sought to understand therapy dropout in the context of adolescent depression, drawing on data from a randomised controlled trial. Study 1 aimed to investigate whether dropout could be predicted from a range of child, family and treatment factors (N = 406). Increase in age and antisocial behaviour, and decrease in verbal intelligence, were found to be significant predictors of dropout. More missed sessions and poorer therapeutic alliance were also significant predictors of dropout. Study 2 aimed to investigate whether those who dropped out of therapy had poorer clinical outcomes compared with those who completed therapy. No strong evidence was found for dropouts having poorer outcomes than completers. Study 3 aimed to explore whether there was a more clinically meaningful way of classifying dropout. Interviews with adolescents (N = 32) and therapists for ‘dropout’ cases were analysed qualitatively using ideal type analysis. Three types of dropout were constructed. ‘Dissatisfied’ dropouts stopped therapy because they did not find therapy helpful; ‘got-what-they-needed’ dropouts stopped therapy because they did not feel a need to continue in therapy; and ‘troubled’ dropouts stopped therapy because they did not have the stability in their life to commit to the therapy. Study 4 aimed to investigate the role of the therapeutic alliance and rupture-repair processes in the lead up to a ‘dissatisfied’ dropout compared to other types of therapy ending. ‘Dissatisfied’ dropouts were found to have more ruptures in the therapeutic alliance, and ruptures were frequently unresolved, compared with completers and ‘got-what-they-needed’ dropouts, indicating a more difficult interaction pattern prior to ‘dissatisfied’ dropout. Together, these studies have implications for how different types of disengagement from treatment should be managed in clinical practice.