Running head: THE STRUCTURE OF SOMATIC SYMPTOMS IN THE PHQ-15

Purpose: Medically unexplained symptoms are the hallmark of somatoform disorders and functional somatic syndromes. Although medically unexplained symptoms represent a common phenomenon both in the general population as well as in medical settings, the exact latent structure of somatic symptoms remains largely unclear. Methods: We examined the latent structure of medically unexplained symptoms by means of the PHQ-15 questionnaire (i.e., a popular symptom checklist) and provide support for the construct validity of our model. The data was analyzed using confirmatory factor analysis in a general population sample (Study 1; N = 414) and in a sample of primary care patients (Study 2; N = 308). We compared four different latent structure models of medically unexplained symptoms: A general factor model, a correlated group factor model, a hierarchical model and a bifactor model. Results: In Study 1, a bifactor model with one general factor and four independent specific symptom factors (i.e., gastrointestinal, pain, fatigue, and cardio-pulmonary symptoms) showed the best model fit. This bifactor model was confirmed in the primary care sample (Study 2). Additionally, the model explained 59% of the variance of the Irritable Bowel Syndrome (IBS). In this structural equation model both the general factor (14%) as well as the gastrointestinal symptom factor (42%) significantly predicted the IBS. Conclusions: The findings of both studies help to clarify the latent structure of somatic symptoms in the PHQ-15. The bifactor model outperformed alternative models and demonstrated external validity in predicting IBS.