A Systematic Review of Validated Screening Tools for Anxiety Disorders and PTSD in Low to Middle Income Countries
Anxiety and post-traumatic stress disorder (PTSD) contribute significantly to disability adjusted life years in low-to middle-income countries (LMICs). Screening has been proposed to improve identification and management of these disorders, but little is known about the validity of screening tools for these disorders. We conducted a systematic review of validated screening tools for detecting anxiety and PTSD in LMICs. Methods: MEDLINE, EMBASE, Global Health and PsychINFO were searched
... e searched (inception-January 10, 2019). Eligible studies (1) screened for anxiety disorders and/or PTSD; (2) reported sensitivity and specificity for a given cut-off value; (3) were conducted in LMICs; and (4) compared screening results to diagnostic classifications based on a reference standard. Screening tool, cut-off, disorder, region, country, and clinical population were extracted for each included study. We assed quality using a modified version of Greenhalgh's ten item checklist. Accuracy results were organized based on screening tool, cut-off, and specific disorder. Accuracy estimates for the same cut-off for the same screening tool and disorder were combined via meta-analysis. Results Of 5343 unique citations identified, 57 articles including 75 screening tools were included. There were 44, 20 and 11 validations for anxiety, PTSD, and combined depression and anxiety, respectively. Continentally, Asia had the most validations (34). Regionally, South Asia (10) had the most validations, followed by West Asia (9) and South Africa (9). The Kessler-10 (7) and the Generalized Anxiety Disorder-7 item scale (GAD-7) (6) were the most commonly validated tools for anxiety disorders, while the Harvard Trauma Questionnaire (3) and Posttraumatic Diagnostic Scale (3) were the most commonly validated tools for PTSD. Most studies (27) had the lowest quality rating (unblinded) followed by good (21). Due to incomplete reporting, we combined only two sets of accuracy values in meta-analysis (GAD-7 cut-off ≥10; sensitivity: 76%, specificity: 64%).