Date of Award

2024

Document Type

Dissertation

Degree Name

Psychology (Psy.D.)

Department

Psychology

First Advisor

Raymond DiGiuseppe

Second Advisor

Imad Zaheer

Third Advisor

William Chaplin

Abstract

To determine optimal psychotherapy doses, researchers commonly rely on one of two competing theories, the Dose-Effect (DE) or the Good-Enough Level (GEL) models, to guide their research. Which model one selects is clinically meaningful, as the models have conflicting evidentiary support. Thus, it is essential to ascertain which theoretical model best fits empirical data. The present study compares the applicability and fit of the DE and GEL theories using total sessions and weeks in treatment as predictors within a community-based, unlimited-duration psychotherapy clinic. Significant determinants of treatment success were investigated. Adult participants’ symptomatology was periodically assessed using the Outcome Questionnaire 45, Second Edition (OQ 45.2). A subsample of clients who completed over two outcome measures without a significant lapse in treatment (n = 311) was analyzed via multilevel modeling. Linear, log-linear, quadratic, and cubic models were constructed and evaluated within each theoretical framework. The superior models from each theory (DE, traditional GEL, and modified GEL) were compared, and the predictive value of using sessions versus weeks was examined. Sociodemographic and clinical variables were integrated into the best-fitting model to explore potential interaction effects. Further, determinants of treatment success and deterioration were analyzed for the full sample of all clients treated (n = 434) including those with early drop out. The findings of this study suggest that the traditional, log-linear GEL model provided the best relative fit to the data for both week and session-based predictors. When compared, week-based predictors had a better model fit and were suggested to explain more of the variability than session-based predictors; however, potential inaccuracies in measuring session variables may have affected these results. Variables related to depressive symptoms (namely, hopelessness, suicidality, previous suicide attempts, and attention problems) showed a significant, positive correlation with initial symptom scores and enhanced the fit of the traditional GEL model. Furthermore, factors including low emotional stability and high symptom scores at baseline and improvements in emotional stability, hope, gratitude, and quality of life during treatment were all associated with better treatment outcomes. Conversely, low patient ratings of the therapeutic alliance during initial sessions were linked to treatment deterioration and dropout.

Download

Included in

Psychology Commons

Share

COinS