Estimating transdiagnostic symptom networks: The problem of "skip outs" in diagnostic interviews

Network models of the symptoms of psychological disorders provide a novel lens for examining comorbidity. Viewing symptoms as causal entities in their own right, researchers can attempt to identify specific symptoms that "bridge" diagnostic entities, providing a more granular perspective on comorbidity than the one provided by analysis at the syndromal level. Such analyses could help identify transdiagnostic targets for both research and clinical interventions. Although promising conceptually, extant work using this approach often uses structured diagnostic interview data that employ "skip outs," branching logic conditional on responses to gating questions (which may be criterial or risk markers). We demonstrate that skip outs, where screening items are asked for each disorder before assessing the remaining symptoms, can produce significant problems in interpretation of comorbidity between symptoms and, hence, transdiagnostic network models. The nature and extent of this problem is explored, and suggestions for future network studies of comorbidity are provided. (PsycINFO Database Record (c) 2018 APA, all rights reserved).