This recent study published in Nature Medicine sought to develop a machine-learning algorithm for prediction of persistent posttraumatic stress symptoms among ED patients. Candidate variables included acute stress symptoms and psychophysiological, immune and inflammatory markers. The resultant algorithm enabled high accurate prediction of those patients with non-remitting symptoms versus those with a resilient trajectory. Consistent with previous research, acute stress symptoms (measured using the Immediate Stress Reaction Checklist) were the most important predictor of persistent posttraumatic stress symptoms.
More information available here