Predicting Early Mortality in Adult Trauma Patients Admitted to Three Public University Hospitals in Urban India: A Prospective Multicentre Cohort Study

Background: In India alone, more than one million people die yearly due to trauma. Identification of patients at risk of early mortality is crucial to guide clinical management and explain prognosis. Prediction models can support clinical judgement, but existing models have methodological limitations. The aim of this study was to derive a vital sign based prediction model for early mortality among adult trauma patients admitted to three public university hospitals in urban India. Methods: We conducted a prospective cohort study of adult trauma patients admitted to three urban university hospitals in India between October 2013 and January 2014. The outcome measure was mortality within 24 hours. We used logistic regression with restricted cubic splines to derive our model. We assessed model performance in terms of discrimination, calibration, and optimism. Results: A total of 1629 patients were included. Median age was 35, 80% were males. Mortality between admission and 24 hours was 6%. Our final model included systolic blood pressure, heart rate, and Glasgow coma scale. Our model displayed good discrimination, with an area under the receiver operating characteristics curve (AUROCC) of 0.85. Predicted mortality corresponded well with observed mortality, indicating good calibration. Conclusion: This study showed that routinely recorded systolic blood pressure, heart rate, and Glasgow coma scale predicted early hospital mortality in trauma patients admitted to three public university hospitals in urban India. Our model needs to be externally validated before it can be applied in the clinical setting.