A personalized approach to treatment with patients being matched to the best-fit treatment has been proposed as one possible solution to the currently modest treatment response rates for adolescent depression. Personalized treatment involves identifying and characterizing subgroups likely to respond differently to different treatments. We investigated the feasibility of this approach, by focusing on two key risk factors that are the purported treatment targets of cognitive behavioral therapy (CBT) and interpersonal psychotherapy for depressed adolescents (IPT-A): negative unrealistic cognitions and interpersonal relationship difficulties, respectively. We sought to learn whether subgroups high and low on the two risk factors, respectively, might be identified in a large sample of depressed, treatment-seeking adolescents. Latent class analysis (LCA) was conducted on measures of the two risk factors among 431 adolescents (age 12–17) in the Treatment for Adolescents with Depression Study. LCA identified three classes: (1) adolescents with high levels of problems in both family relationships and cognitions (21.6% of sample), (2) adolescents with moderate levels of problems in both domains (52.4%), and (3) adolescents with low levels of problems in both domains (26.0%). These subgroups did not predict treatment outcome with CBT or CBT + fluoxetine (COMB). The results challenge a current assumption about how treatments could be personalized, and they support a multi-causal model of depression rather than a risk-factor-specific model. Strategies other than risk factor-based personalizing for case assignment to CBT vs. IPT-A are discussed.
PubMed: MeSH publication types
- Journal Article