Abstract
Background: This study aimed to assess the measurement properties (reliability, factor structure, and criterion validity) of the Patient Health Questionnaire (PHQ-9) as an instrument for screening major depressive disorder (MDD) in elderly primary care users in Chile. Method: About 582 participants aged between 65 and 80 years were enrolled from primary care centers. They completed the Composite International Diagnostic Interview (CIDI), a survey with sociodemographic characteristics and the PHQ-9. Results: The PHQ-9 revealed an acceptable internal consistency (ω = 0.79 [95% CI: 0.75-0.80] and α = 0.78 [95% CI: 0.75-0.81]); confirmatory factor analysis demonstrated a good fit for both 1- and 2-factor solutions. The chi-square difference test (χ(2) = 0.61, gl = 1, p = 0.43) and correlation between the somatic and the cognitive-effective latent factors were very high (r = 0.97, p < 0.001), indicating that the 1 factor model was more parsimonious. Utilizing the CIDI as the gold standard, the area under the curve (AUC) was 0.88 (SE = 0.04, 95% CI: 0.84-0.90). The optimal cut-off score of ≥ 6 yielded good sensitivity and specificity for detecting MDD (0.95 and 0.76, respectively). However, considering the clinical utility index, the cut-off score of ≥9 proved to be a more effective marker for discarding cases of MDD. Conclusion: The PHQ-9 has adequate psychometric properties for elderly primary care users. In clinical settings, it showed its greatest utility in ruling out the presence of an MDD, however, its clinical value for identifying possible cases of MDD is limited. In cases above the cut-off point, it is recommended to perform a more thorough evaluation.