Screening for depression: Rasch analysis of the structural validity of the PHQ-9 in acutely injured trauma survivors

Abstract

ObjectiveThe Patient Health Questionnaire (PHQ-9) is widely used for screening of depression in acutely injured trauma survivors. Rasch analysis has been used to evaluate its measurement characteristics. This paper aims: 1) to assess the psychometric properties of the nine PHQ-9 items and 2) to determine the structural validity of using the total PHQ-9 score as a clinical outcome measure. MethodsPHQ-9 data for 937 persons aged 18–60years admitted to 20 level 1 trauma centers in the United States were included. Good model fit indicates that all items contribute to a single underlying trait. ResultsItem 2 demonstrated misfit to the Rasch model, and six items showed disordered response categories. Ordered response categories were achieved for all nine items after modifying the original four-point scoring system into a three-point system. Person separation reliability was acceptable (0.80) for discriminating between groups of patients. Dimensionality testing supported combining the nine items into a total score. No significant differential item functioning was observed for sex and age group. ConclusionDespite some minor problems with its measurement structure, the short nine-item version of the PHQ seems to be an economic and valid instrument for the screening of depression in adults admitted to level 1 trauma centers.