Janssen EP, Köhler S, Stehouwer CD, Schaper NC, Dagnelie PC, Sep SJ, Henry RM, van der Kallen CJ, Verhey FR, Schram MT.
To assess the psychometric properties and identify the best cutoff value of the Patient Health Questionnaire-9 (PHQ-9) for depression screening in individuals with type 2 diabetes mellitus (T2DM).
Observational population-based cohort study.
The Maastricht Study.
Individuals with and without T2DM (mean age 58.6 ± 8.1, 44.6% male) according to an oral glucose tolerance test (N = 2,997).
Depressive disorder and depressive symptoms were measured using the Mini-International Neuropsychiatric Interview (MINI) as the reference and the PHQ-9. Cronbach alpha, Cohen's kappa and receiver operating characteristic (ROC) analyses were used. Differences in factorial structure between participants with and without T2DM were tested using multigroup confirmatory factor analysis.
Based on the traditional PHQ-9 cutoff value, 133 (4.4%) participants had depressive symptoms (PHQ-9 score ≥10). Internal consistency of the PHQ-9 was good (Cronbach α = 0.87 with T2DM, 0.82 without T2DM), the kappa of agreement between the PHQ-9 and the MINI was moderate (0.40 with T2DM, 0.43 without T2DM). Area under the ROC curve for the PHQ-9 was 0.87 in participants with T2DM and 0.88 in those without. A PHQ-9 cutoff score of 5 provided the best sensitivity (92.3%), with acceptable specificity (70.4%), for T2DM, similar to sensitivity and specificity in individuals without T2DM. Factor analysis suggested a similar two-factor structure in both groups (affective and somatic symptoms).
Patient Health Questionnaire-9 performs well as a screening tool for depressive symptoms in individuals with and without T2DM based on the cutoff value of 5, indicating that the PHQ-9 can be used in two-stage screening in primary care to select individuals with T2DM for further psychological evaluation.