Reliability and validity of PHQ-9 when administered by health workers for depression screening among women in primary care

Abstract Background Depression is twice more prevalent among women but remains unidentified in primary care. Objective We aimed to estimate the reliability and validity of PHQ-9, when administered by health workers, a cadre of public health staff, posted in primary health centres. We translated PHQ-9 to Malayalam, a language spoken by 30 million people in Kerala, India. Health workers administered PHQ-9 to women (n = 238) aged 18–60 years in a high prevalent primary care setting. Mini International Neuropsychiatric Interview (MINI) was administered by the psychiatrist on the same day to diagnose depressive disorder. Two health workers administered PHQ-9, independently, in a subset of 21 women. Results The internal consistency reliability (Cronbach’s alpha 0.89) and inter-rater reliability (intra class correlation coefficient, 0.94; 95% CI, 0.86–0.95) were high. On ROC analysis, area under curve was 0.92 (95% CI 0.88–0.96). For a cut-off score of ≥9, PHQ-9 had a sensitivity of 82.5%, (95% CI, 72.4–92.6), specificity of 90.1% (95% CI, 84.5–95.6%), positive predictive value of 73.4% (95% CI, 62.4–84.4%) and negative predictive value of 93.9% (95% CI, 90.2–97.6%). The accuracy was 88.2% (84.0–92.4%) and positive likelihood ratio was 8.3. Factor analysis supported a single factor, with eigen value above 1, with high loading for all items (0.73–0.79), except item related to appetite (0.58). This explained 62% of variance in the data. Prevalence of MINI diagnosed depressive disorders was 25%. Conclusion When administered by health workers, PHQ-9 has good reliability and at cut off score ≥9, it has good validity to identify depression in primary care.