RISK FACTORS FOR ANXIETY DISORDERS: COMMON AND SPECIFIC EFFECTS IN A NATIONAL SAMPLE

Anxiety disorders and major depressive disorder (MDD) are among the most common mental disorders in the United States,[1] with high individual and societal burden derived from their considerable associated work and social impairment.[2,3] Decreasing the burden of anxiety disorders and MDD is an important public health priority. Anxiety disorders often co-occur with each other and with MDD[4,5] suggesting the existence of a latent structure that may parsimoniously explain these patterns of comorbidity[6–9]. Several studies have examined the genetic liability of anxiety disorders and MDD, and consistently concluded that most of this genetic liability is exerted at the level of the latent variable,[8, 10–13] rather than at the disorder-specific level. Because anxiety disorders and MDD share a broad range of other risk factors,[14–24] an essential step in the development of more effective treatment and prevention interventions is to determine whether those factors exert their effect through the common latent variable, directly on each disorder, or through a combination of effects at both levels.[14, 16, 25] We sought tsreo address this gap in knowledge by drawing on data from a large, nationally representative survey to compare the common and the specific contribution of a set of observable risk factors to risk of anxiety disorders and MDD. We followed a conceptual model of risk factors, which was initially developed by Kendler and colleagues to examine the prevalence of MDD,[26, 27] but seems to be applicable to other internalizing disorders.[28] Because anxiety disorders often co-occur with each other and with MDD and seem to have a partially shared liability,[8, 13, 14] prior to our analyses, we hypothesized that the comorbidity structure of anxiety disorders and MDD would be well described by a single latent variable and that the risk factors postulated in Kendler’s model would exert part of their effect through this latent variable. However, because the prevalence and course of the different anxiety disorder and MDD does not fully overlap[29–31] and they also have nonshared variance, we also hypothesized that the risk factors would also partially exert their effect directly on the individual disorders.