Diagnostic test accuracy of screening tools for post-traumatic stress disorder among refugees and asylum seekers: A systematic review and meta-analysis

Abstract

Background

Refugees and asylum seekers often experience traumatic events resulting in a high prevalence of post-traumatic stress disorder (PTSD). Undiagnosed PTSD can have detrimental effects on resettlement outcomes. Immigration medical exams provide an opportunity to screen for mental health conditions in refugee and asylum seeker populations and provide links to timely mental health care.

Objective

To assess the diagnostic accuracy of screening tools for PTSD in refugee and asylum seeker populations.

Methods

We systematically searched Medline, Embase, PsycINFO, CENTRAL and CINAHL up to 29 September 2022. We included cohort-selection or cross-sectional study designs that assessed PTSD screening tools in refugee or asylum seeker populations of all ages. All reference standards were eligible for inclusion, with a clinical interview considered the gold standard. We selected studies and extracted diagnostic test accuracy data in duplicate. Risk of bias and applicability concerns were addressed using QUADAS-2. We meta-analyzed findings using a bivariate random-effects model. We partnered with a patient representative and a clinical psychiatrist to inform review development and conduct.

Results

Our review includes 28 studies (4,373 participants) capturing 16 different screening tools. Nine of the 16 tools were developed specifically for refugee populations. Most studies assessed PTSD in adult populations, but three included studies focused on detecting PTSD in children. Nine studies looked at the Harvard Trauma Questionnaire (HTQ) with diagnostic cut-off points ranging from 1.17 to 2.5. Meta-analyses revealed a summary point sensitivity of 86.6% (95%CI 0.791; 0.917) and specificity of 78.9% (95%CI 0.639; 0.888) for these studies. After evaluation, we found it appropriate to pool other screening tools (Posttraumatic Stress Disorder Checklist, the Impact of Event Scale, and the Posttraumatic Diagnostic Scale) with the HTQ. The area under the curve for this model was 79.4%, with a pooled sensitivity of 86.2% (95%CI 0.759; 0.925) and a specificity of 72.2% (95%CI 0.616; 0.808).

Conclusions

Our review identified several screening tools that perform well among refugees and asylum seekers, but no single tool was identified as being superior. The Refugee Health Screener holds promise as a practical instrument for use in immigration medical examinations because it supports the identification of PTSD, depression, and anxiety across diverse populations. Future research should consider tool characteristics beyond sensitivity and specificity to facilitate implementation in immigration medical exams.

Registration

Open Science Framework: 10.17605/OSF.IO/PHNJV