Abstract
BackgroundPredicting Posttraumatic Stress Disorder (PTSD) is a pre-requisite for targeted prevention. Current research has identified group-level risk-indicators, many of which (e.g., head trauma, receiving opiates) concern but a subset of survivors. Identifying interchangeable sets of risk indicators may increase the efficiency of early risk assessment. The study goal is to use supervised machine learning (ML) to uncover interchangeable, maximally predictive combinations of early risk indicators.MethodsData variables (features) reflecting event characteristics, emergency department (ED) records and early symptoms were collected in 957 trauma survivors within ten days of ED admission, and used to predict PTSD symptom trajectories during the following fifteen months. A Target Information Equivalence Algorithm (TIE*) identified all minimal sets of features (Markov Boundaries; MBs) that maximized the prediction of a non-remitting PTSD symptom trajectory when integrated in a support vector machine (SVM). The predictive accuracy of each set of predictors was evaluated in a repeated 10-fold cross-validation and expressed as average area under the Receiver Operating Characteristics curve (AUC) for all validation trials.ResultsThe average number of MBs per cross validation was 800. MBs’ mean AUC was 0.75 (95% range: 0.67-0.80). The average number of features per MB was 18 (range: 12–32) with 13 features present in over 75% of the sets.ConclusionsOur findings support the hypothesized existence of multiple and interchangeable sets of risk indicators that equally and exhaustively predict non-remitting PTSD. ML’s ability to increase prediction versatility is a promising step towards developing algorithmic, knowledge-based, personalized prediction of post-traumatic psychopathology.Electronic supplementary materialThe online version of this article (doi:10.1186/s12888-015-0399-8) contains supplementary material, which is available to authorized users.
Highlights
Predicting Posttraumatic Stress Disorder (PTSD) is a pre-requisite for targeted prevention
In a previous study [23], we evaluated the ability of machine learning (ML)-based feature-selection algorithm to extract one set of early risk indicators
We evaluated the accuracy of prediction from each of these sets using support vector machines (SVMs [24])
Summary
Predicting Posttraumatic Stress Disorder (PTSD) is a pre-requisite for targeted prevention. The early identification of individuals at risk for developing posttraumatic stress disorder (PTSD) is a major clinical and public health challenge, which many studies have attempted to address (for meta-analyses, see Brewin et al [1] and Ozer et al [2]). Current findings suggest that PTSD is associated This translational gap has several reasons: Previous studies have identified risk indicators at the group level, thereby overlooking within-group heterogeneities and distinct individual paths to PTSD that emanate from the disorder’s complex multi-causal etiology [14]. Within the inherently complex and multimodal matrix of emerging post-traumatic morbidity, the relative contribution of any risk-indicator is necessarily context-dependent and does not directly translate across traumatic events and individuals exposed (e.g., female gender increases the likelihood of PTSD among survivors of physical assault, but not in accidents victims [3]). Proper risk assessment defies simple computation and requires knowledge-based, rule driven expert systems
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
