Psychodynamic Psychotherapy: A Quantitative, Longitudinal Perspective

Background and Purpose: The National Institute of Mental Health and others (Gabbard, Gunderson, & Fonagy, 2002; APA Division 12 Task Force, 1993) have lamented the lack of empirical research conducted by psychodynamic psychotherapists and have called for these practitioners to start demonstrating their outcomes. Gabbard, Gunderson, & Fonagy (2002) wrote about the threat posed to dynamic therapy by not demonstrating such outcomes: “[p]sychoanalytic psychotherapy is at risk of being sacrificed if scientific methods cannot be developed that will further test its practitioners' claims of efficacy” (p. 505).

Empirical studies, or efforts in this direction, are just beginning among dynamic practitioners, including some recent widely disseminated examples. Two notable examples include Leichsenring & Rabung's (2008) meta-analysis on the effectiveness of long-term psychodynamic psychotherapy published in JAMA, and Shedler's (2010) meta-analysis in American Psychologist, which gives attention to the efficacy of psychodynamic psychotherapy, broadly, demonstrating relatively large effect sizes. Method: This study used a quantitative, longitudinal methodology to look at the nature of change for 1,050 people undertaking psychodynamic psychotherapy at a psychodynamic mental health clinic, as an example of community-university collaborative research. It used a secondary data analysis to look at the nature of recovery in this psychotherapy at three month intervals, spanning 45 months, or nearly four years of data. It gave consideration to where people tended to begin (their intercept), their progress (slope), and the role of potential moderating variables such as comorbidity and medication (in the form of interaction effects), using multilevel modeling (SPSS 17.0). The psychometrically validated Outcome Questionnaire (OQ 45.2) was given at baseline and at three month variables for all clients undertaking psychotherapy at this clinic. The data set included all participants who had more than one OQ 45.2 score on record.

Results: In keeping with Leichsenring & Rabung's (2008) and Shedler's (2010) studies, this study found that people made statistically and clinically significant change, broadly. This applied across: gender, diagnoses, age, ethnicity (Spanish speaking and senior clients were looked at as subgroups of interest), income, insurance status, and length of treatment. The most pronounced symptom change and the largest effect sizes (.45) were found in the first year of treatment, and in particular, in the first three months of treatment. People largely made progress regardless of number of sessions. There was little deterioration, especially in the first year (11 - 14%). Outcomes became more variable after a year, with the percentage of clients deteriorating in relation to their own baseline score increasing, ranging from 16 – 23%.

Conclusions & Implications: This study holds the potential to begin to suggest something about the unique course of recovery and the optimal length (“dosing”) of this form of treatment for people presenting with particular diagnoses and treatment needs. It also complicates Shedler's and Leichsenring & Rabung's findings in that it shows both (1) promise of good outcomes in less than a year, and (2) smaller effect sizes in a “real life” clinic setting versus their reviews of largely prospective and efficacy studies.