Patient characteristics of completion and dropout of mentalization-based treatment for adolescents with conduct disorder

Abstract

Introduction: Conduct disorder (CD) is a severe mental disorder in youth. Yet, providing psychological interventions for adolescents with CD is challenging. This patient group is often characterized by risk factors for therapy dropout such as, e.g., CD symptoms and being in middle adolescence. On the other hand, little is known about characteristics of adolescents with CD who complete treatment. To gain more insight into what might become a successful therapy with adolescents with CD, this study explores baseline characteristics and drop-out occurrence in patients with CD referred to mentalization-based treatment for adolescents with CD (MBT-CD). More specifically, this study aims at identifying clusters of adolescent patients based on age, CD symptom severity and personality pathology at the beginning of treatment which may have come along with a higher or lower dropout occurrence.

Methods: Following implications of an elbow plot, a 3-means cluster-analysis was conducted on self-report baseline data of N = 32 adolescents with CD (n = 11 dropouts, n = 21 completers) who participated in a feasibility study on MBT-CD. Additionally, in an exploratory stepwise logistic regression, variables were explored as potential predictors of dropout.

Results: Cluster 1 consisted of n = 14 adolescents, of whom n = 8 (57%) dropped out. Cluster 2 consisted of n = 5 adolescents of whom 1 (20%) dropped out. Cluster 3 consisted of N = 13 adolescents, of whom only n = 2 (15%) dropped out. Cluster 2 showed descriptively the highest CD symptom severity. While adolescents in Clusters 1 and 3 showed in part similarities in CD symptom severity, personality pathology was descriptively markedly higher in Cluster 1. In the stepwise logistic regression, only intimacy personality pathology was identified as potential predictor for dropout.

Discussion: This study's exploratory findings point to different types of adolescents with CD coming along with different chances for therapists to conduct a (successful) psychotherapy. Herein, low personality functioning in the intimacy domain, rather than CD symptoms as aggressiveness, may play a crucial role. While findings are limited by the small sample size, they may be able to shed increasing light on conducting (successful) psychotherapy with a scarcely researched patient group.

Keywords: adolescence; cluster analysis; conduct disorder; dropout; mentalization-based treatment; personality pathology; stepwise logistic regression.

Figure 1

Elbow plot for visualization of numbers of clusters in relation to intra-cluster variation.

Figure 2

32 subjects in 3 clusters as identified in the cluster analysis. 0 indicates a subject, who completed treatment. 1 indicates a subject, who dropped out of treatment. Cluster 3 is depicted left, Cluster 2 in the middle, Cluster 1 on the right.