Longitudinal Assessment of Post-traumatic Syndromes

点击翻译按钮获取中文摘要

Each year, more than 40 million Americans present to US emergency departments (EDs) for evaluation after trauma exposure (TE). While the majority of these individuals recover, an important subset develops adverse posttraumatic neuropsychiatric sequelae (APNS). These APNS include traditionally categorized outcomes such as posttraumatic stress disorder (PTSD), depression, minor traumatic brain injury (MTBI), and regional or widespread pain. However, these previous definitions of outcome have limited progress, and we now appreciate that the actual trajectories of APNS are multidimensional, incorporating a range of specific outcomes that may be best understood, and optimally targeted for intervention, by dividing across specific domains of functioning. This application, submitted in response to RFA-MH-16-500, proposes to identify and characterize the trajectories of the most common trauma-induced APNS within these domains of functioning using the RDoC classification system. 5,000 patients presenting to the ED after trauma will be screened, recruited, and will receive initial baseline evaluation in the ED, including blood collection and psychophysical, survey, and neurocognitive evaluation. They will be closely monitored over the next 8 weeks using innovative technologies (a wrist wearable for continuous-time monitoring of daytime physiology and sleep; a smart phone app for continuous-time monitoring of GPS and daily “flash” surveys; weekly web-based neurocognitive tests; periodic mixed-mode surveys; serial saliva collection; deep phenotyping [blood collection, fMRI, psychophysical evaluation]) and then followed less intensively using similar procedures (including deep phenotyping) over the remainder of a 52-week follow-up period. Adaptive sampling and state-of-the-art statistical methods will be used to (1) optimize precision in characterizing RDoC construct trajectories and (2) test theoretically-guided, “high yield” hypotheses evaluating the effects of pre-trauma, peritraumatic, and recovery-related factors on these trajectories and on multivariate RDoC construct trajectory profiles. The longitudinal schedule of rich, granular, multidimensional data collection in the study has been specifically designed to evaluate those constructs most important to post-TE outcomes and to test the proposed hypotheses. Ensemble machine learning methods will be used to develop tiered-targeted clinical decision support models to identify individuals at high risk of specific, common APNS outcomes. The close-knit ED research network that will undertake the study has a strong track record of prospective research on APNS and is ideally suited to carry out this exceedingly complex study. The study has been designed to be a resource for the entire field (for example, it has been designed and budgeted to collect and store a great many more biological samples at the NIMH Biorespository than we can analyze, for use by other investigators).