Psychosocial predictors of treatment outcome for trauma-affected refugees

Background: The effects of treatment in trials with trauma-affected refugees vary considerably not only between studies but also between patients within a single study. However, we know little about why some patients benefit more from treatment, as few studies have analysed predictors of treatment outcome.
Objective: The objective of the study was to examine possible psychosocial predictors of treatment outcome for trauma-affected refugees.
Method: The participants were 195 adult refugees with posttraumatic stress disorder (PTSD) who were enrolled in a 6- to 7-month treatment programme at the Competence Centre for Transcultural Psychiatry (CTP), Denmark. The CTP Predictor Index used in the study included 15 different possible outcome predictors concerning the patients’ past, chronicity of mental health problems, pain, treatment motivation, prerequisites for engaging in psychotherapy, and social situation. The primary outcome measure was PTSD symptoms measured on the Harvard Trauma Questionnaire (HTQ). Other outcome measures included the Hopkins Symptom Check List-25, the WHO-5 Well-being Index, Sheehan Disability Scale, Hamilton Depression and Anxiety Scales, the somatisation scale of the Symptoms Checklist-90, Global Assessment of Functioning scales, and pain rated on visual analogue scales. The relations between treatment outcomes and the total score as well as subscores of the CTP Predictor Index were analysed.
Results: Overall, the total score of the CTP Predictor Index was significantly correlated to pre- to post treatment score changes on the majority of the ratings mentioned above. While employment status was the only single item significantly correlated to HTQ-score changes, a number of single items from the CTP Predictor Index correlated significantly with changes in depression and anxiety symptoms, but the size of the correlation coefficients were modest.
Conclusions: The total score of the CTP Predictor Index correlated significantly with outcomes on most of the rating scales, but correlations were modest in size, possibly due to the number of different factors influencing treatment outcome.