Predictors of treatment duration and retention in a study of long-term dynamic psychotherapy: childhood adversity, adult personality, and diagnosis.

Systematic knowledge about patient characteristics that predict the duration of long-term psychotherapy is largely absent. We examined predictors of attrition, retention, and duration of long-term dynamic psychotherapy to delineate the naturalistic history of psychotherapy, specifically focusing on childhood emotional neglect and abuse, adult diagnosis, personality, and functioning as predictors. Fifty-three adults with depressive, anxiety, and/or personality disorders (PDs) were offered at least 3 years of long-term dynamic psychotherapy. The median duration of therapy for the study group was 110 sessions (95% confidence interval [CI] 52-141). Nondynamic characteristics (including demographics; most Axis I diagnoses; depression, anxiety, and distress scores; social and global functioning; and the five personality factors) did not predict number of sessions. Dysthymic disorder, presence of any PD (particularly dependent PD [DPD]), emotional neglect in childhood, and higher adaptive defense style scores predicted a greater number of sessions, while obsessive-compulsive PD (OCPD) predicted fewer sessions. Emotional neglect, DPD, higher adaptive defenses, and OCPD were each unique predictors of duration. A session frequency less than 0.7 sessions per week (36 sessions per year) was associated with a three-fold higher risk for intrinsic attrition (relative risk = 3.04, 95% CI 1.10-8.44). Childhood emotional neglect as well as adult dependency may predispose patients to remain in therapy for longer durations, while some adaptive defenses may allow patients to contain the distressing affects that might otherwise lead to early termination. It remains to be seen whether longer durations of treatments are associated with respectively greater improvement, other things being equal, and whether these findings generalize to other types of treatment.