Validation of a clinical tool for depression screening

Abstract

Depression is a common disorder that requires an accurate diagnosis. The Patient Health Questionnaire (PHQ- 9) and clinical interviews are essential screening tools. This study analyzes the combined efficiency of these tools. The effectiveness of the PHQ-9 was evaluated in conjunction with three depression screening instruments by analyzing sensitivity and specificity at cut-off values between 5 and 15. Online calculators for correlation coefficients and the ROC curve were used. The Youden Index was calculated, and groups were compared. The PHQ-9 compared with MINI showed the AUC 0.972, demonstrating capacity to distinguish between positive and negative cases. Next is the semi-structured interview (AUC 0.964), followed by the fully structured interview (AUC 0.836). The highest Youden Index (0.70) was recorded for the semi-structured interview (cutoff 10), followed by MINI (0.65, cut-off 8–9), and the full interview (0.58, cut-off 8). All three standards show strong negative correlations between sensitivity and specificity (ρ, τ = –1), with the most predictable result at the fully structured interview (r= –0.943). PHQ-9 validated by MINI is the most sensitive and ideal for accurate diagnosis. The semi-structured interview provides an optimal balance between false positives and false negatives. The fully structured interview is stable but less effective in diagnosis.