The AdHOC (age, head injury, oxygenation, circulation) score: a simple assessment tool for early assessment of severely injured patients with major fractures

The role of big data management, data registries, and machine learning algorithms for optimizing safe definitive surgery in trauma: a review

Digital data processing has revolutionized medical documentation and enabled the aggregation of patient data across hospitals. Initiatives such as those from the AO Foundation about fracture treatment (AO Sammelstudie, 1986), the Major Trauma Outcome Study (MTOS) about survival, and the Trauma Audit and Research Network (TARN) pioneered multi-hospital data collection. Large trauma registries, like the German Trauma Registry (TR-DGU) helped improve evidence levels but were still constrained by predefined data sets and limited physiological parameters. The improvement in the understanding of pathophysiological reactions substantiated that decision making about fracture care led to development of patient's tailored dynamic approaches like the Safe Definitive Surgery algorithm. In the future, artificial intelligence (AI) may provide further steps by potentially transforming fracture recognition and/or outcome prediction. The evolution towards flexible decision making and AI-driven innovations may be of further help. The current manuscript summarizes the development of big data from local databases and subsequent trauma registries to AI-based algorithms, such as Parkland Trauma Mortality Index and the IBM Watson Pathway Explorer.

Parkland Trauma Index of Mortality in Orthopaedic Trauma Patients: An Initial Report

OBJECTIVES

The extent and timing of surgery in severely injured patients remains an unsolved problem in orthopaedic trauma. Different laboratory values or scores have been used to try to predict mortality and estimate physiological reserve. The Parkland Trauma Index of Mortality (PTIM) has been validated as an electronic medical record-integrated algorithm to help with operative timing in trauma patients. The aim of this study was to report our initial experience with PTIM and how it relates to other scores.

METHODS

A retrospective chart review of level 1 and level 2 trauma patients admitted to our institution between December 2020 and November 2022 was conducted. Patients scored with PTIM with orthopaedic injuries were included in this study. Exclusion criteria were patients younger than 18 years.

RESULTS

Seven hundred seventy-four patients (246 female patients) with a median age of 40.5 (18-101) were included. Mortality was 3.1%. Patients in the PTIM high-risk category (≥0.5) had a 20% mortality rate. The median PTIM was 0.075 (0-0.89) and the median Injury Severity Score (ISS) was 9.0 (1-59). PTIM (P < 0.001) and ISS (P < 0.001) were significantly lower in surviving patients. PTIM was mentioned in 7.6% of cases, and in 1.7% of cases, providers indicated an action in response to the PTIM. PTIM and ISS were significantly higher in patients with documented PTIM.

CONCLUSION

PTIM is better at predicting mortality compared with ISS. Our low rate of PTIM documentation in provider notes highlights the challenges of implementing a new algorithm.

LEVEL OF EVIDENCE

Level III, retrospective cohort study.

Keeping it simple: the value of mortality prediction after trauma with basic indices like the Reverse Shock Index multiplied by Glasgow Coma Scale

BACKGROUND

Identification of trauma patients at significant risk of death in the prehospital setting is challenging. The prediction probability of basic indices like vital signs, Shock Index (SI), SI multiplied by age (SIA) or the GCS is limited and more complex scores are not feasible on-scene. The Reverse SI multiplied by GCS score (rSIG) has been proposed as a triage tool to identify trauma patients with an increased risk of dying at EDs. Age adjustment (rSIG/A) displayed no advantage.We aim to (1) validate the accuracy of the rSIG in predicting death or early transfusion in a large trauma registry population, and (2) determine if the rSIG is valid for evaluation of trauma patients in the prehospital setting.

METHODS

70 829 trauma patients were retrieved from the TraumaRegister DGU database (time period between 2008 and 2017). The area under the receiver operating characteristic curve (AUROC) was calculated to measure the ability of SI, SIA, rSIG and rSIG divided by age (rSIG/A) to predict in-hospital mortality from data at the time of hospital arrival and solely from prehospital data.

RESULTS

The rSIG at time of hospital admission was not sufficiently predictive for clinical decision-making. However, rSIG calculated solely from prehospital data accurately predicted risk of death. Using prehospital data, the AUROC for mortality of rSIG/A was the highest (0.85; CI: 0.85 to 0.86), followed by rSIG (0.76; CI: 0.75 to 0.77), SIA (0.71; CI: 0.70 to 0.71) and SI (0.48; CI: 0.47 to 0.49).

CONCLUSION

The prehospital rSIG/A can be a useful adjunct for the prehospital evaluation of trauma patients and their allocation to trauma centres or trauma team activation. However, we could not confirm that the rSIG at hospital admission is a reliable tool for risk stratification.

Parkland Trauma Index of Mortality: Real-Time Predictive Model for Trauma Patients

OBJECTIVE

Vital signs and laboratory values are used to guide decisions to use damage control techniques in lieu of early definitive fracture fixation. Previous 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 data to predict mortality within 48 hours during the first 3 days of hospitalization. It updates every hour, recalculating as physiology changes. The model was developed using 1935 trauma patient encounters from 2009 to 2014 and validated on 516 patient encounters from 2015 to 2016. Model performance was evaluated statistically. Data were collected retrospectively on its performance after 1 year of clinical use.

RESULTS

In the validation data set, PTIM accurately predicted 52 of the sixty-three 12-hour time intervals within 48 hours of mortality, for sensitivity of 82.5% [95% confidence interval (CI), 73.1%-91.9%]. The specificity was 93.6% (95% CI, 92.5%-94.8%), and the positive predictive value (PPV) was 32.5% (95% CI, 25.2%-39.7%). PTIM predicted survival for 1608 time intervals and was incorrect only 11 times, yielding a negative predictive value of 99.3% (95% CI, 98.9%-99.7%). The area under the curve of the receiver operating characteristic 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 twenty-three 12-hour time intervals within 48 hours of mortality, for sensitivity of 86.9% (95% CI, 73%-100%). The specificity was 94.7% (95% CI, 93%-96%), and the positive predictive value was 33.3% (95% CI, 21.4%-45%). The model predicted survival for 716 time intervals and was incorrect 3 times, yielding a negative predictive value of 99.6% (95% CI, 99.1%-100%). The area under the curve of the receiver operating characteristic curve was 0.97.

CONCLUSIONS

By adapting with the patient's physiologic response to trauma and relying on electronic medical record data alone, the PTIM overcomes many of the limitations of previous models. It may help inform decision-making for trauma patients early in their hospitalization.

LEVEL OF EVIDENCE

Prognostic Level I. See Instructions for Authors for a complete description of levels of evidence.

Independent Predictors of Mortality in Torso Trauma Injuries

Noncompressible torso injuries (NCTIs) represent a trauma-related condition with high lethality. This study’s aim was to identify potential prediction factors of mortality in this group of trauma patients at a Level 1 trauma center in Italy. Materials and Methods: A total of 777 patients who had sustained a noncompressible torso injury (NCTI) and were admitted to the Niguarda Trauma Center in Milan from 2010 to 2019 were included. Of these, 166 patients with a systolic blood pressure (SBP) <90 mmHg were considered to have a noncompressible torso hemorrhage (NCTH). Demographic data, mechanism of trauma, pre-hospital and in-hospital clinical conditions, diagnostic/therapeutic procedures, and survival outcome were retrospectively recorded. Results: Among the 777 patients, 69% were male and 90.2% sustained a blunt trauma with a median age of 43 years. The comparison between survivors and non-survivors pointed out a significantly lower pre-hospital Glasgow coma scale (GCS) and SBP (p < 0.001) in the latter group. The multivariate backward regression model identified age, pre-hospital GCS and injury severity score (ISS) (p < 0.001), pre-hospital SBP (p = 0.03), emergency department SBP (p = 0.039), performance of torso contrast enhanced computed tomography (CeCT) (p = 0.029), and base excess (BE) (p = 0.008) as independent predictors of mortality. Conclusions: Torso trauma patients who were hemodynamically unstable in both pre- and in-hospital phases with impaired GCS and BE had a greater risk of death. The detection of independent predictors of mortality allows for the timely identification of a subgroup of patients whose chances of survival are reduced.