| Between 35--57% of couples drop out of psychotherapy. Evidence suggests dropout is influenced by therapeutic alliance (the interpersonal relationship between client and therapist). For couples, measurement of alliance should include each spouse's perception of the self-therapist alliance, the partner-therapist alliance, and the relationships among those dimensions. Most previous studies have considered spouses' perceptions separately.; This study characterized therapeutic alliance in couples therapy by examining patterns of disagreement among the alliance dimensions. A split pattern, identified previously but inconsistently described, is defined here as the disagreement between the husband's and the wife's self-therapist alliances. Additionally, two new relational patterns were examined: (1) skew, imbalance within each spouse's view of the self-therapist and partner-therapist alliances, and (2) discrepancy, difference between spouses' views regarding husband-therapist and wife-therapist alliances. These three relational patterns and aggregate, a composite of the alliance dimensions, were examined in relationship to dropout.; The primary hypothesis was that couples who drop out of treatment would have greater split alliance patterns. Participants were 56 couples undergoing couples therapy who completed pre-treatment questionnaires regarding couple characteristics and, post-session, the Couples Therapy Alliance Scale (CTAS). Alliance patterns and aggregate alliance were constructed from CTAS scores. Following session one, 20% of couples dropped out, and another 20% did so before treatment ended.; As predicted, dropout couples had greater split alliance patterns compared to those who remained (Mann-Whitney U); there were no significant differences in skew, discrepancy or aggregate. To develop a more comprehensive preliminary model, relationships were examined among dropout, alliance and pre-therapy characteristics (age, relationship duration, distance from clinic, children, income, education, ethnicity, drug use, prior hospitalization, marital commitment, and adjustment). This model was explored using logistic regression, survival analysis, and ROC signal detection. Split alliance remained the most significant variable for identifying couples who dropped out following session one. Couples with a partner under 40 were more likely to drop out than couples over 40. This study is the first to use split alliance as the primary predictor of dropout and to determine empirically a cut-point for significant split. Implications for future research and clinical practice are discussed. |
