J Orthop Trauma. 2021 Oct 12. doi: 10.1097/BOT.0000000000002290. Online ahead of print.
ABSTRACT
OBJECTIVE: Vital signs and laboratory values are used to guide decisions to use damage control techniques in lieu of early definitive fracture fixation. Prior models attempted to predict mortality risk but have limited utility. There is a need for a dynamic model that captures evolving physiologic changes during a trauma patient’s hospital course.
METHODS: The Parkland Trauma Index of Mortality (PTIM) is a machine-learning algorithm that uses electronic medical record (EMR) data to predict mortality within 48 hours during the first 3 days of hospitalization. It updates every hour, re-calculating as physiology changes. The model was developed using 1,935 trauma patient encounters from 2009-2014 and validated on 516 patient encounters from 2015-2016. Model performance was evaluated statistically. Data was collected retrospectively on its performance after one year of clinical use.
RESULTS: In the validation data set, PTIM accurately predicted 52 of 63 12-hour time intervals within 48 hours of mortality, for sensitivity of 82.5% (95% CI 73.1% – 91.9%). Specificity was 93.6% (95% CI 92.5% – 94.8%), and PPV was 32.5% (95% CI 25.2% – 39.7%). PTIM predicted survival for 1,608 time intervals and was incorrect only 11 times, yielding a NPV of 99.3% (95% CI 98.9% – 99.7%). AUC of the ROC curve was 0.94.During the first year of clinical use, when used in 776 patients, the last PTIM score accurately predicted 20 of the 23 12-hour time intervals within 48 hours of mortality, for a sensitivity of 86.9% (95% CI 73% – 100%). Specificity was 94.7% (95% CI 93% – 96%), and the PPV was 33.3% (95% CI 21.4% – 45%). The model predicted survival for 716 time intervals and was incorrect 3 times, yielding a NPV of 99.6% (95% CI 99.1% – 100%). AUC of the ROC curve was 0.97.
CONCLUSION: By adapting with the patient’s physiologic response to trauma and relying on EMR data alone, the PTIM overcomes many of the limitations of prior models. It may help inform decision-making for trauma patients early in their hospitalization.
LEVEL OF EVIDENCE: Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.
PMID:34653106 | DOI:10.1097/BOT.0000000000002290
