Agreement among brief patient health questionnaire (PHQ) versions for measuring symptoms of depression in the Brazilian national health survey (PNS).

To evaluate the agreement of the 2-item (PHQ-2) and 8-item (PHQ-8) Patient Health Questionnaires compared to the PHQ-9 and clinical diagnosis for depression screening in Brazilian adults, and to examine associated sociodemographic and behavioral factors. We analyzed cross-sectional data from 148,733 participants in the 2013 and 2019 Brazilian National Health Surveys. ROC curve analyses identified optimal cutoffs for PHQ-8 and PHQ-2 against the PHQ-9 (≥ 10) and clinical indicators. Logistic regression and Ratio of Odds Ratios (ROR) compared associations across screening tools and self-reported diagnosis. The PHQ-8 (cutoff ≥ 10) showed near-perfect agreement with the PHQ-9 (AUC = 0.982). For the PHQ-2, a ≥ 3 cutoff optimized specificity (96.4%) and overall agreement (94.7%), with an AUC of 0.877. Optimal cutoffs against clinical indicators shifted lower (≥ 4 for PHQ-8; ≥ 1 for PHQ-2). Both brief versions identified associations profiles consistent with the PHQ-9 (e.g., higher odds in women, low-income individuals, and those with chronic diseases). However, ROR analyses revealed a significant healthcare gap: higher income and advancing age strongly predicted formal diagnosis, while Black/Brown populations and those with poor health were significantly underrepresented clinically compared to symptom screening. All scales showed strong reliability (α, ω > 0.85), including the PHQ-2 (r = 0.55; ρsb = 0.71). The PHQ-8 is a psychometrically equivalent alternative to the PHQ-9 in Brazil, and the PHQ-2 is a reliable, ultra-brief tool ideal for rapid, large-scale population screening. While these scales accurately capture symptom burden, the observed discrepancies with formal diagnosis highlight significant systemic barriers to mental healthcare access for vulnerable subgroups.