Development and validation of a brief screener for posttraumatic stress disorder risk in emergency medical settings.

OBJECTIVE: Predicting risk of posttraumatic stress disorder (PTSD) in the acute care setting is challenging given the pace and acute care demands in the emergency department (ED) and the infeasibility of using time-consuming assessments. Currently, no accurate brief screening for long-term PTSD risk is routinely used in the ED. One instrument widely used in the ED is the 27-item Immediate Stress Reaction Checklist (ISRC). The aim of this study was to develop a short screener using a machine learning approach and to investigate whether accurate PTSD prediction in the ED can be achieved with substantially fewer items than the IRSC. METHOD: This prospective longitudinal cohort study examined the development and validation of a brief screening instrument in two independent samples, a model development sample (N = 253) and an external validation sample (N = 93). We used a feature selection algorithm to identify a minimal subset of features of the ISRC and tested this subset in a predictive model to investigate if we can accurately predict long-term PTSD outcomes. RESULTS: We were able to identify a reduced subset of 5 highly predictive features of the ISRC in the model development sample (AUC = 0.80), and we were able to validate those findings in the external validation sample (AUC = 0.84) to discriminate non-remitting vs. resilient trajectories. CONCLUSION: This study developed and validated a brief 5-item screener in the ED setting, which may help to improve the diagnostic process of PTSD in the acute care setting and help ED clinicians plan follow-up care when patients are still in contact with the healthcare system. This could reduce the burden on patients and decrease the risk of chronic PTSD.

Utilization of machine learning for prediction of post-traumatic stress: a re-examination of cortisol in the prediction and pathways to non-remitting PTSD.

To date, studies of biological risk factors have revealed inconsistent relationships with subsequent post-traumatic stress disorder (PTSD). The inconsistent signal may reflect the use of data analytic tools that are ill equipped for modeling the complex interactions between biological and environmental factors that underlay post-traumatic psychopathology. Further, using symptom-based diagnostic status as the group outcome overlooks the inherent heterogeneity of PTSD, potentially contributing to failures to replicate. To examine the potential yield of novel analytic tools, we reanalyzed data from a large longitudinal study of individuals identified following trauma in the general emergency room (ER) that failed to find a linear association between cortisol response to traumatic events and subsequent PTSD. First, latent growth mixture modeling empirically identified trajectories of post-traumatic symptoms, which then were used as the study outcome. Next, support vector machines with feature selection identified sets of features with stable predictive accuracy and built robust classifiers of trajectory membership (area under the receiver operator characteristic curve (AUC)=0.82 (95% confidence interval (CI)=0.80-0.85)) that combined clinical, neuroendocrine, psychophysiological and demographic information. Finally, graph induction algorithms revealed a unique path from childhood trauma via lower cortisol during ER admission, to non-remitting PTSD. Traditional general linear modeling methods then confirmed the newly revealed association, thereby delineating a specific target population for early endocrine interventions. Advanced computational approaches offer innovative ways for uncovering clinically significant, non-shared biological signals in heterogeneous samples.