Validation of the Bangla PHQ-4 in rural and urban community samples from Bangladesh

Abstract

Background Shortage of qualified mental health professionals and limited access to specialized care, coupled with a high burden of mental health problems, particularly anxiety and depression, is a growing public health concern in Bangladesh. A possible solution being tested is the utilization of existing community resources. Availability of a psychometrically valid and reliable screening tool for early detection of mental health problems is of paramount importance. The present study aimed at testing the psychometric properties of the Bangla version of the ultra-brief PHQ-4. Methods Employing practicing mental health professionals as data collectors, this cross-sectional study used data from 357 adults from an urban and two rural communities of Bangladesh. Results Confirmatory factor analysis indicated excellent model fit for the two-factor structure of the Bangla PHQ-4 ( χ 2  = 0.103 at p  > .01, RMSEA = 0.000, CFI = 1.000, TLI = 1.007, and SRMR = .006) for the total sample. The model demonstrated factorial invariance for sex, age, and mental health state. The scale demonstrated high internal consistency (McDonald’s ω = .869) and strong construct validity, evidenced by high correlation with the SRQ-20 (r = .747), K-6 (r = .820), self-rated mental health state (r = −.677), and interviewer-rated mental health state (r = −.680). The scale also demonstrates excellent diagnostic accuracy (AUC = .906), and acceptable sensitivity (82%) and specificity (83%) using a cutoff score ≥ 6. Conclusion This study provided evidence of satisfactory psychometric properties of the Bangla PHQ-4 in terms of factor structure, group level invariance, internal consistency, and construct validity in Bangladeshi rural and urban community samples.