Validity/Reliability of PHQ-9 and PHQ-2 Depression Scales Among Adults Living with HIV/AIDS in Western Kenya

Background  Depression greatly burdens sub-Saharan Africa, especially populations living with HIV/AIDS, for whom few validated depression scales exist. Patient Health Questionnaire-9 (PHQ-9), a brief dual-purpose instrument yielding DSM-IV diagnoses and severity, and PHQ-2, an ultra-brief screening tool, offer advantages in resource-constrained settings. Objective  To assess the validity/reliability of PHQ-9 and PHQ-2. Design  Observational, two occasions 7 days apart. Participants  A total of 347 patients attending psychosocial support groups. Measurements  Demographics, PHQ-9, PHQ-2, general health perception rating and CD4 count. Results  Rates for PHQ-9 DSM-IV major depressive disorder (MDD), other depressive disorder (ODD) and any depressive disorder were 13%, 21% and 34%. Depression was associated with female gender, but not CD4. Construct validity was supported by: (1) a strong association between PHQ-9 and general health rating, (2) a single major factor with loadings exceeding 0.50, (3) item-total correlations exceeding 0.37 and (4) a pattern of item means similar to US validation studies. Four focus groups indicated culturally relevant content validity and minor modifications to the PHQ-9 instructions. Coefficient alpha was 0.78. Test-retest reliability was acceptable: (1) intraclass correlation 0.59 for PHQ-9 total score, (2) kappas 0.24, 0.25 and 0.38 for PHQ-9 MDD, ODD and any depressive disorder and (3) weighted kappa 0.53 for PHQ-9 depression severity categories. PHQ-2 ≥3 demonstrated high sensitivity (85%) and specificity (95%) for diagnosing any PHQ-9 depressive disorder (AUC, 0.97), and 91% and 77%, respectively, for diagnosing PHQ-9 MDD (AUC, 0.91). Psychometrics were also good within four gender/age (18–35, 36–61) subgroups. Conclusions  PHQ-9 and PHQ-2 appear valid/reliable for assessing DSM-IV depressive disorders and depression severity among adults living with HIV/AIDS in western Kenya.