The proposed project requests two years of funding to investigate the natural course of Posttraumatic Stress Disorder (PTSD). This project would be conducted using secondary data Analyses of 8 longitudinal data sets that represent a wide-range of trauma types and victim characteristics. These include: veterans of the first gulf war, motor vehicle accident survivors, community violence survivors, victims of traumatic injuries, and victims of intimate partner violence. The primary focus of the analyses will be measures of the 17 symptoms that are part of the DSM-IV definition of PTSD. The overarching goal of the project is to apply advanced statistical methods (e.g., structural equation models, growth mixture modQls, and missing data procedures) to better understand the temporal course of this disorder, and to build tools that will facilitate early treatment. The three specific aims of the proposed study of PTSD symptoms are: (1) To document the prospective relationships among the symptom clusters that comprise PTSD. Mary theories of the development of PTSD posit causal relationships among these distinct types of symptoms. We will determine which causal models are consistent with the data using cross-lagged panel analysis on these longitudinal datasets. We will also document the extent that these results are robust across a range of trauma types and participant characteristics. (2) To determine if there are clinically meaningful subtypes of posttraumatic distress that can be distinguished by unique symptom trajectories. We propose to use groWth mixture modeling to icenlify the number of trajectory types that commonly occur following trauma (e.g., quick resolution, delayed onset), to describe these trajectory types, and to explore variables that may predict membership in these classes. (3) To develop a brief, early-screening instrument that would identify those at greatest risk for subsequent PTSD. PTSD diagnosis requires that symptoms have persisted beyonjd 30-days post trauma. However, it may be possible to use a subset of symptoms measured shortly aftelr the trauma to accurately predict persons who will later develop PTSD, thereby facilitating early intervention. We propose to empirically derive an early warning instrument using these longitudinal datasets.
