Mortality prediction using TRISS methodology in the Spanish ICU Trauma Registry (RETRAUCI)

Comparison of Risk Assessment Tools' Prediction of Outcomes for Penetrating Trauma

INTRODUCTION

The Trauma and Injury Severity Score (TRISS) uses anatomic and/or physiologic variables to predict mortality; however, Injury Severity Score is less reliable for penetrating trauma. National Surgical Quality Improvement Program Surgical Risk Calculator (NSQIP-SRC) and American Society of Anesthesiologists Physical Status (ASA-PS) include functional status and comorbidities. This study evaluates the accuracy of these tools in predicting mortality, length of stay (LOS), and complications for operative penetrating trauma.

METHODS

Adult penetrating trauma patients (≥18 y) who underwent surgery within 24 h of admission were included in this subgroup analysis of a prospective observational study at four trauma centers. The following three scoring models were compared: NSQIP-SRC, TRISS, and ASA-PS. Brier scores and area under the receiver-operating characteristic curve were used to compare mortality prediction. LOS prediction was assessed with linear regression and complications were evaluated with negative binomial regression. Likelihood ratio (LR) test was used to assess model fit.

RESULTS

Of 329 penetrating trauma patients, 13 (3.9%) died. The median LOS was 4 d (interquartile range 2-9), and median number of complications was zero (interquartile range 0-1). TRISS better predicted mortality than NSQIP-SRC or ASA-PS on Brier score (0.02 versus 0.03 versus 0.03) but all had similar area under the receiver-operating characteristic curve (0.93 versus 0.93 versus 0.91, P = 0.26). NSQIP-SRC and ASA-PS better predicted LOS on adjusted R2 (14.4% versus 14.1% versus 1.6%) and LR showed no difference between these two tools (P = 0.16). NSQIP-SRC best predicted complications compared to TRISS and ASA-PS (Pseudo R2: 10.3% versus 3.8% versus 5.5%; LR: P = 0.003).

CONCLUSIONS

For penetrating trauma, all three models were similarly excellent at predicting mortality. NSQIP-SRC and ASA-PS better predicted LOS and NSQIP-SRC best predicted complications, suggesting these are better tools for prognostication of outcomes for penetrating trauma.

Lack of Concordance Between Abbreviated Injury Scale and American Association for the Surgery of Trauma Organ Injury Scale in Patients with High-Grade Solid Organ Injury

BACKGROUND

The Abbreviated Injury Scale (AIS) is widely used for body region-specific injury severity. The American Association for the Surgery of Trauma Organ Injury Scale (AAST-OIS) provides organ-specific injury severity but is not included in trauma databases. Previous researchers have used AIS as a surrogate for OIS. This study aims to assess AIS-abdomen concordance with AAST-OIS grade for liver and spleen injuries, hypothesizing concordance in terms of severity (grade of OIS and AIS) and patient outcomes.

STUDY DESIGN

This retrospective study (July 2020 to June 2022) was performed at 3 trauma centers. Adult trauma patients with AAST-OIS grade III to V liver and spleen injury were included. AAST-OIS grade for each organ was compared with AIS-abdomen by evaluating the percentage of AAST-OIS grade correlating with each AIS score as well as rates of operative intervention for these injuries. Analysis was performed with chi-square tests and univariate analysis.

RESULTS

Of 472 patients, 274 had liver injuries and 205 had spleen injuries grades III to V. AAST-OIS grade III to V liver injuries had concordances rates of 85.5%, 71%, and 90.9% with corresponding AIS 3 to 5 scores. AAST-OIS grade III to V spleen injuries had concordances rates of 89.7%, 87.8%, and 87.3%, respectively. There was a statistical lack of concordance for both liver and spleen injuries (both p < 0.001). Additionally, there were higher rates of operative intervention for AAST-OIS grade IV and V liver injuries and grade III and V spleen injuries vs corresponding AIS scores (p < 0.05).

CONCLUSIONS

AIS should not be used interchangeably with OIS due to lack of concordance. AAST-OIS should be included in trauma databases to facilitate improved organ injury research and quality improvement projects.

A Prospective Multicenter Comparison of Trauma and Injury Severity Score, American Society of Anesthesiologists Physical Status, and National Surgical Quality Improvement Program Calculator's Ability to Predict Operative Trauma Outcomes

BACKGROUND

Trauma outcome prediction models have traditionally relied upon patient injury and physiologic data (eg, Trauma and Injury Severity Score [TRISS]) without accounting for comorbidities. We sought to prospectively evaluate the role of the American Society of Anesthesiologists physical status (ASA-PS) score and the National Surgical Quality Improvement Program Surgical Risk-Calculator (NSQIP-SRC), which are measurements of comorbidities, in the prediction of trauma outcomes, hypothesizing that they will improve the predictive ability for mortality, hospital length of stay (LOS), and complications compared to TRISS alone in trauma patients undergoing surgery within 24 hours.

METHODS

A prospective, observational multicenter study (9/2018-2/2020) of trauma patients ≥18 years undergoing operation within 24 hours of admission was performed. Multiple logistic regression was used to create models predicting mortality utilizing the variables within TRISS, ASA-PS, and NSQIP-SRC, respectively. Linear regression was used to create models predicting LOS and negative binomial regression to create models predicting complications.

RESULTS

From 4 level I trauma centers, 1213 patients were included. The Brier Score for each model predicting mortality was found to improve accuracy in the following order: 0.0370 for ASA-PS, 0.0355 for NSQIP-SRC, 0.0301 for TRISS, 0.0291 for TRISS+ASA-PS, and 0.0234 for TRISS+NSQIP-SRC. However, when comparing TRISS alone to TRISS+ASA-PS ( P = .082) and TRISS+NSQIP-SRC ( P = .394), there was no significant improvement in mortality prediction. NSQIP-SRC more accurately predicted both LOS and complications compared to TRISS and ASA-PS.

CONCLUSIONS

TRISS predicts mortality better than ASA-PS and NSQIP-SRC in trauma patients undergoing surgery within 24 hours. The TRISS mortality predictive ability is not improved when combined with ASA-PS or NSQIP-SRC. However, NSQIP-SRC was the most accurate predictor of LOS and complications.

Survival benefits of interventional radiology and surgical teams collaboration during primary trauma surveys: a single-centre retrospective cohort study

BACKGROUND

A team approach is essential for effective trauma management. Close collaboration between interventional radiologists and surgeons during the initial management of trauma patients is important for prompt and accurate trauma care. This study aimed to determine whether trauma patients benefit from close collaboration between interventional radiology (IR) and surgical teams during the primary trauma survey.

METHODS

A retrospective observational study was conducted between 2014 and 2021 at a single institution. Patients were assigned to an embolization group (EG), a surgery group (SG), or a combination group (CG) according to their treatment. The primary and secondary outcomes were survival at hospital discharge compared with the probability of survival (Ps) and the time course of treatment.

RESULTS

The analysis included 197 patients, consisting of 135 men and 62 women, with a median age of 56 [IQR, 38-72] years and an injury severity score of 20 [10-29]. The EG, SG, and CG included 114, 48, and 35 patients, respectively. Differences in organ injury patterns were observed between the three groups. In-hospital survival rates in all three groups were higher than the Ps. In particular, the survival rate in the CG was 15.5% higher than the Ps (95% CI: 7.5-23.6%; p < 0.001). In the CG, the median time for starting the initial procedure was 53 [37-79] min and the procedure times for IR and surgery were 48 [29-72] min and 63 [35-94] min, respectively. Those times were significantly shorter among three groups.

CONCLUSION

Close collaboration between IR and surgical teams, including the primary survey, improves the survival of severe trauma patients who require both IR procedures and surgeries by improving appropriate treatment selection and reducing the time process.

Comparison of National Surgical Quality Improvement Program Surgical Risk Calculator and Trauma and Injury Severity Score Risk Assessment Tools in Predicting Outcomes in High-Risk Operative Trauma Patients

BACKGROUND

The Trauma and Injury Severity Score (TRISS) uses anatomic/physiologic variables to predict outcomes. The National Surgical Quality Improvement Program Surgical Risk Calculator (NSQIP-SRC) includes functional status and comorbidities. It is unclear which of these tools is superior for high-risk trauma patients (American Society of Anesthesiologists Physical Status (ASA-PS) class IV or V). This study compares risk prediction of TRISS and NSQIP-SRC for mortality, length of stay (LOS), and complications for high-risk operative trauma patients.

METHODS

This is a prospective study of high-risk (ASA-PS IV or V) trauma patients (≥18 years-old) undergoing surgery at 4 trauma centers. We compared TRISS vs NSQIP-SRC vs NSQIP-SRC + TRISS for ability to predict mortality, LOS, and complications using linear, logistic, and negative binomial regression.

RESULTS

Of 284 patients, 48 (16.9%) died. The median LOS was 16 days and number of complications was 1. TRISS + NSQIP-SRC best predicted mortality (AUROC: .877 vs .723 vs .843, P = .0018) and number of complications (pseudo-R2/median error (ME) 5.26%/1.15 vs 3.39%/1.33 vs 2.07%/1.41, P < .001) compared to NSQIP-SRC or TRISS, but there was no difference between TRISS + NSQIP-SRC and NSQIP-SRC with LOS prediction (P = .43).

DISCUSSION

For high-risk operative trauma patients, TRISS + NSQIP-SRC performed better at predicting mortality and number of complications compared to NSQIP-SRC or TRISS alone but similar to NSQIP-SRC alone for LOS. Thus, future risk prediction and comparisons across trauma centers for high-risk operative trauma patients should include a combination of anatomic/physiologic data, comorbidities, and functional status.

Validation of the Trauma and Injury Severity Score for Prediction of Mortality in a Greek Trauma Population

BACKGROUND

Although the Trauma and Injury Severity Score (TRISS) has been extensively used for mortality risk adjustment in trauma, its applicability in contemporary trauma populations is increasingly questioned.

OBJECTIVE

The study aimed to evaluate the predictive performance of the TRISS in its original and revised version and compare these with a recalibrated version, including current data from a Greek trauma population.

METHODS

This is a retrospective cohort study of admitted trauma patients conducted in two tertiary Greek hospitals from January 2016 to December 2018. The model algorithm was calculated based on the Major Trauma Outcome Study coefficients (TRISSMTOS), the National Trauma Data Bank coefficients (TRISSNTDB), and reweighted coefficients of logistic regression obtained from a Greek trauma dataset (TRISSGrTD). The primary endpoint was inhospital mortality. Models' prediction was performed using discrimination and calibration statistics.

RESULTS

A total of 8,988 trauma patients were included, of whom 854 died (9.5%). The TRISSMTOS displayed excellent discrimination with an area under the curve (AUC) of 0.912 (95% CI 0.902-0.923) and comparable with TRISSNTDB (AUC = 0.908, 95% CI 0.897-0.919, p = .1195). Calibration of both models was poor (Hosmer-Lemeshow test p < .001), tending to underestimate the probability of mortality across almost all risk groups. The TRISSGrTD resulted in statistically significant improvement in discrimination (AUC = 0.927, 95% CI 0.918-0.936, p < .0001) and acceptable calibration (Hosmer-Lemeshow test p = .113).

CONCLUSION

In this cohort of Greek trauma patients, the performance of the original TRISS was suboptimal, and there was no evidence that it has benefited from its latest revision. By contrast, a strong case exists for supporting a locally recalibrated version to render the TRISS applicable for mortality prediction and performance benchmarking.

The effect of participation of interventional radiology team in a primary trauma survey on patient outcome

PURPOSE

The purpose of this study was to examine the survival benefits of a workflow in which an interventional radiology (IR) team participates in a primary trauma survey on patients with hemodynamically unstable trauma.

MATERIALS AND METHODS

A retrospective observational study was conducted between 2012 and 2019 at a single institution. Patients who underwent an IR procedure as the initial hemostasis were assigned to the hemodynamically stable group (HSG) or hemodynamically unstable group (HUG). The primary and secondary outcomes were survival at hospital discharge compared with the probability of survival (Ps) and the time course.

RESULTS

A total of 160 patients (100 men, 60 women; median age, 57.5 years [interquartile range (IQR): 31.5-72 years]) with an injury severity score of 24 (IQR: 13.75-34) were included. A total of 125 patients were included in the HSG group and 35 patients in the HUG group. The observational survival rate was significantly greater than the Ps rate by 4.9% (95% confidence interval [CI]: 1.6-8.4%; P = 0.005) in HSG and by 24.6% in HUG (95% CI: 16.9-32.3%; P < 0.001). The observational survival rate was significantly greater than Ps in HUG than in HSG (P < 0.001). The median time to initiate IR procedures and the median procedure time in HUG were 54 min [IQR: 45-66 min] and 48 min [IQR: 30-85 min], respectively; both were significantly shorter than those in the HSG.

CONCLUSION

A trauma workflow utilizing an IR team in a primary survey is associated with improved survival of patients with hemodynamically unstable trauma when compared with Ps with a shorter time course.

Machine learning techniques for mortality prediction in critical traumatic patients: anatomic and physiologic variables from the RETRAUCI study

Background

Interest in models for calculating the risk of death in traumatic patients admitted to ICUs remains high. These models use variables derived from the deviation of physiological parameters and/or the severity of anatomical lesions with respect to the affected body areas. Our objective is to create different predictive models of the mortality of critically traumatic patients using machine learning techniques.

Methods

We used 9625 records from the RETRAUCI database (National Trauma Registry of 52 Spanish ICUs in the period of 2015–2019). Hospital mortality was 12.6%. Data on demographic variables, affected anatomical areas and physiological repercussions were used. The Weka Platform was used, along with a ten-fold cross-validation for the construction of nine supervised algorithms: logistic regression binary (LR), neural network (NN), sequential minimal optimization (SMO), classification rules (JRip), classification trees (CT), Bayesian networks (BN), adaptive boosting (ADABOOST), bootstrap aggregating (BAGGING) and random forest (RFOREST). The performance of the models was evaluated by accuracy, specificity, precision, recall, F-measure, and AUC.

Results

In all algorithms, the most important factors are those associated with traumatic brain injury (TBI) and organic failures. The LR finds thorax and limb injuries as independent protective factors of mortality. The CT generates 24 decision rules and uses those related to TBI as the first variables (range 2.0–81.6%). The JRip detects the eight rules with the highest risk of mortality (65.0–94.1%). The NN model uses a hidden layer of ten nodes, which requires 200 weights for its interpretation. The BN find the relationships between the different factors that identify different patient profiles. Models with the ensemble methodology (ADABOOST, BAGGING and RandomForest) do not have greater performance. All models obtain high values ​​in accuracy, specificity, and AUC, but obtain lower values ​​in recall. The greatest precision is achieved by the SMO model, and the BN obtains the best recall, F-measure, and AUC.

Conclusion

Machine learning techniques are useful for creating mortality classification models in critically traumatic patients. With clinical interpretation, the algorithms establish different patient profiles according to the relationship between the variables used, determine groups of patients with different evolutions, and alert clinicians to the presence of rules that indicate the greatest severity.

Outcomes of very elderly trauma ICU patients. Results from the Spanish trauma ICU registry

OBJECTIVE

To analyze outcomes and factors related to mortality among very elderly trauma patients admitted to intensive care units (ICUs) participating in the Spanish trauma ICU registry.

DESIGN

A multicenter nationwide registry. Retrospective analysis. November 2012-May 2017.

SETTING

Participating ICUs.

PATIENTS

Trauma patients aged ≥80 years.

INTERVENTIONS

None.

MAIN VARIABLES OF INTEREST

The outcomes and influence of limitation of life sustaining therapy (LLST) were analyzed. Comparisons were established using the Wilcoxon test, Chi-squared test or Fisher's exact test as appropriate. Multiple logistic regression analysis was performed to analyze variables related to mortality. A p-value <0.05 was considered statistically significant.

RESULTS

The mean patient age was 83.4±3.3 years; 281 males (60.4%). Low-energy falls were the mechanisms of injury in 256 patients (55.1%). The mean ISS was 20.5±11.1, with a mean ICU stay of 7.45±9.9 days. The probability of survival based on the TRISS methodology was 69.8±29.7%. The ICU mortality rate was 15.5%, with an in-hospital mortality rate of 19.2%. The main cause of mortality was intracranial hypertension (42.7%). The ISS, the need for first- and second-tier measures to control intracranial pressure, and being admitted to the ICU for organ donation were independent mortality predictors. LLST was applied in 128 patients (27.9%). Patients who received LLST were older, with more severe trauma, and with more severe brain injury.

CONCLUSIONS

Very elderly trauma ICU patients presented mortality rates lower than predicted on the basis of the severity of injury.

Risk stratification tools in emergency general surgery

The use of risk stratification tools (RST) aids in clinical triage, decision making and quality assessment in a wide variety of medical fields. Although emergency general surgery (EGS) is characterized by a comorbid, physiologically acute patient population with disparately high rates of perioperative morbidity and mortality, few RST have been explicitly examined in this setting. We examined the available RST with the intent of identifying a tool that comprehensively reflects an EGS patients perioperative risk for death or complication. The ideal tool would combine individualized assessment with relative ease of use. Trauma Scoring Systems, Critical Care Scoring Systems, Surgical Scoring Systems and Track and Trigger Models are reviewed here, with the conclusion that Emergency Surgery Acuity Score and the American College of Surgeons National Surgical Quality Improvement Programme Universal Surgical Risk Calculator are the most applicable and appropriate for EGS.

Scoring severity in trauma: comparison of prehospital scoring systems in trauma ICU patients

PURPOSE

We evaluated the predictive ability of mechanism, Glasgow coma scale, age and arterial pressure (MGAP), Glasgow coma scale, age and systolic blood pressure (GAP), and triage-revised trauma Score (T-RTS) scores in patients from the Spanish trauma ICU registry using the trauma and injury severity score (TRISS) as a reference standard.

METHODS

Patients admitted for traumatic disease in the participating ICU were included. Quantitative data were reported as median [interquartile range (IQR), categorical data as number (percentage)]. Comparisons between groups with quantitative variables and categorical variables were performed using Student's T Test and Chi Square Test, respectively. We performed receiving operating curves (ROC) and evaluated the area under the curve (AUC) with its 95 % confidence interval (CI). Sensitivity, specificity, positive predictive and negative predictive values and accuracy were evaluated in all the scores. A value of p < 0.05 was considered significant.

RESULTS

The final sample included 1361 trauma ICU patients. Median age was 45 (30-61) years. 1092 patients (80.3 %) were male. Median ISS was 18 (13-26) and median T-RTS was 11 (10-12). Median GAP was 20 (15-22) and median MGAP 24 (20-27). Observed mortality was 17.7 % whilst predicted mortality using TRISS was 16.9 %. The AUC in the scores evaluated was: TRISS 0.897 (95 % CI 0.876-0.918), MGAP 0.860 (95 % CI 0.835-0.886), GAP 0.849 (95 % CI 0.823-0.876) and T-RTS 0.796 (95 % CI 0.762-0.830).

CONCLUSIONS

Both MGAP and GAP scores performed better than the T-RTS in the prediction of hospital mortality in Spanish trauma ICU patients. Since these are easy-to-perform scores, they should be incorporated in clinical practice as a triaging tool.