This study was conducted to evaluate the diagnostic accuracy and determine the optimum cut-off scores for clinical use of the Center for Epidemiological Studies Depression scale (CES-D) and Alcohol Use Disorders Identification Test (AUDIT) against a reference psychiatric diagnostic interview, in TB and anti-retroviral therapy (ART) patients in primary care in Zambia.
This was a cross-sectional study in 16 primary level care clinics. Consecutive sampling was used to select 649 participants who started TB treatment or ART in the preceding month. Participants were first interviewed using the CES-D and AUDIT, and subsequently with a psychiatric diagnostic interview for current major depressive disorder (MDD) and alcohol use disorders (AUDs) using the Mini-International Neuropsychiatric Interview (MINI). The diagnostic accuracy was calculated using the Area Under the Receiver Operating Characteristic curve (AUROC). The optimum cut-off scores for clinical use were calculated using sensitivity and positive predictive value (PPV).
The CES-D and AUDIT had high internal consistency (Cronbach's alpha = 0.84; 0.98 respectively). Confirmatory factor analysis showed that the four-factor CES-D model was not a good fit for the data (Tucker-Lewis Fit Index (TLI) = 0.86; standardized root-mean square residual (SRMR) = 0.06) while the two-factor AUDIT model fitted the data well (TFI = 0.99; SRMR = 0.04). Both the CES-D and AUDIT demonstrated good discriminatory ability in detecting MINI-defined current MDDs and AUDs (AUROC for CES-D = 0.78; AUDIT = 0.98 for women and 0.75 for men). The optimum CES-D cut-off score in screening for current MDD was 22 (sensitivity 73%, PPV 76%) while that of the AUDIT in screening for AUD was 24 for women (sensitivity 60%, PPV 60%), and 20 for men (sensitivity 55%, PPV 50%).
The CES-D and AUDIT showed high discriminatory ability in measuring MINI-defined current MDD and AUD respectively. They are suitable mental health screening tools for use among TB and ART patients in primary care in Zambia.