Can a global score for anxiety and depression be obtained from the Patient Health Questionnaire (PHQ-4) in the Peruvian population that has experienced the death of a loved one? Empirical support for a unidimensional or two-dimensional model

Abstract

Introduction: Worldwide, anxiety and depression are the most common mental disorders, both in clinical settings and in the general population. The Patient Health Questionnaire-4 is the most widely used open access instrument to assess anxiety and depression in both clinical and community settings. Objective: The study aims to evaluate the psychometric evidence of the Patient Health Questionnaire (PHQ-4). Method: 1015 Peruvian individuals between 18 and 69 years old (M=26.4, SD=9.93, 58.6% women) who experienced the death of a loved one participated. Techniques derived from Classical Test Theory and Item Response Theory (IRT) were used. Results: Three models were tested: one-dimensional (CFI = 0.989; TLI = 0.966; RMSEA = 0.075), the two-factor model (CFI = 1.000; TLI = 1.000; RMSEA = 0.000) and the bifactor model which did not converge. The two-factor model was the one that came closest to a perfect model. Reliability, evaluated using the omega coefficient, was adequate for both the unidimensional model (ω = 0.88) and the two-factor model (ω anxiety = 0.83; ω depression = 0.78). The IRT analysis indicated that the PHQ-4 items are adequate indicators that can discriminate between those who do or do not present the traits of anxiety and depression. Regarding the relationship between the PHQ-4 and pandemic grief, both the unidimensional model and the two-factor model demonstrated appropriate fit indices. It was shown that the one-dimensional and two-factor models are invariant according to gender and age. Conclusion: Although the one-factor model and two correlated factors show adequate fit indices, the second had a better fit. Furthermore, this model presented adequate reliability, discrimination and a significant relationship with dysfunctional grief.