Early Identification of Post-Traumatic Stress Disorder in Trauma Patients: Development of a Multivariable Risk Prediction Model.

Abstract

The purpose of this study was to build a risk prediction model to identify trauma patients at the time of injury who are at high risk for post-traumatic stress disorder (PTSD) 1year later. Patients 18+ with operative orthopedic trauma injuries were enrolled in prospective social determinants of health cohort. Data were collected through initial surveys, medical records at time of injury, and 1-year follow-up phone screenings. Univariate analysis examined associations between factors and PTSD at 1 year. The best fit multivariable logistic regression model led to a novel PTSD risk prediction tool based on weights assigned similar to the Charlson index methods. Of 329 enrolled patients, 87 (26%) completed follow-up surveys; 58% screened positive for chronic PTSD. The best fit model predicting PTSD included age, insurance, violent mechanism, and 2 acute stress screening questions (AUC .89). Using these parameters, the maximum possible TIPPS index was 19. Those with PTSD at 1year had a mean TIPPS index of 12.9 ± 4.0, compared to 5.9 ± 4.2 for those who did not (P < .001). Traumatic injury often leads to PTSD, which can be predicted by a novel risk score incorporating age, insurance status, violent injury mechanism, and acute stress reaction symptoms. Stability in life and relationships with primary care physicians may be protective of PTSD. Diagnostic level II.