Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Karen Fondacaro

Second Advisor

Jennifer Hurley


It is estimated that approximately one in five patients will terminate therapy early, before participating in full treatment and obtaining maximum therapeutic benefits. Millions of people are forcibly displaced as refugees each year, and therefore at increased risk for poverty, discrimination, and complex mental health needs, yet no research has yet examined rates or predictors of psychotherapy attrition among refugees. The current study draws upon a sample of refugee clients seeking treatment at a community clinic (N = 196), and a comparison group of 165 non-refugee clients at the same clinic. Logistic regression was employed to (1) compare rates of attrition between refugees and a sample of non-refugees matched for income level, sex, and age, (2) identify socioeconomic predictors of attrition in refugees, and (3) identify psychological predictors of attrition in refugees only. Exploratory analyses also examined the interactive effects of formal education and psychological symptoms among refugees only. Results indicated comparable rates of early termination regardless of refugee status, although refugees were more likely than the comparison group to experience an extended gap in early treatment. Twenty-eight percent of refugees and 35% of non-refugees stopped therapy within ten weeks; 45% of refugees and 45% of non-refugees stopped therapy within ten sessions; 14% of refugees and two percent of non-refugees experienced an extended treatment gap within ten weeks/sessions. Fifty-four percent of refugees and 47% of non-refugees experienced at least one of these early treatment disruptions. Within refugee patients’ psychotherapy, and using empirically derived symptom-based scales (somatic, cognitive, avoidance, and posttraumatic stress), age, education, and the interaction between somatic symptoms and education significantly predicted attrition. Each increasing year of age was associated with a 3.4% decrease in risk for attrition. Each education level increase (e.g., primary school to secondary school) was associated with a 24.9-30.3% decrease in attrition risk. Although somatic symptoms alone did not significantly predict attrition, the interaction was significant such that for each increase in education level, a one-point somatic score increase conferred an additional 28.8% decrease in attrition risk. More educated clients with high somatic symptoms were most likely to stay in treatment, as were less educated clients with low somatic symptoms. Findings suggest new culturally relevant considerations for enhancing treatment retention in refugees.



Number of Pages

95 p.