Dataset Information

Using the National Trauma Data Bank (NTDB) and machine learning to predict trauma patient mortality at admission.

ABSTRACT: A 400-estimator gradient boosting classifier was trained to predict survival probabilities of trauma patients. The National Trauma Data Bank (NTDB) provided 799233 complete patient records (778303 survivors and 20930 deaths) each containing 32 features, a number further reduced to only 8 features via the permutation importance method. Importantly, the 8 features can all be readily determined at admission: systolic blood pressure, heart rate, respiratory rate, temperature, oxygen saturation, gender, age and Glasgow coma score. Since death was rare, a rebalanced training set was used to train the model. The model is able to predict a survival probability for any trauma patient and accurately distinguish between a deceased and survived patient in 92.4% of all cases. Partial dependence curves (Psurvival vs. feature value) obtained from the trained model revealed the global importance of Glasgow coma score, age, and systolic blood pressure while pulse rate, respiratory rate, temperature, oxygen saturation, and gender had more subtle single variable influences. Shapley values, which measure the relative contribution of each of the 8 features to individual patient risk, were computed for several patients and were able to quantify patient-specific warning signs. Using the NTDB to sample across numerous patient traumas and hospital protocols, the trained model and Shapley values rapidly provides quantitative insight into which combination of variables in an 8-dimensional space contributed most to each trauma patient's predicted global risk of death upon emergency room admission.

SUBMITTER: Tsiklidis EJ

PROVIDER: S-EPMC7671512 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Using the National Trauma Data Bank (NTDB) and machine learning to predict trauma patient mortality at admission.

Tsiklidis Evan J EJ Sims Carrie C Sinno Talid T Diamond Scott L SL

PloS one 20201117 11

A 400-estimator gradient boosting classifier was trained to predict survival probabilities of trauma patients. The National Trauma Data Bank (NTDB) provided 799233 complete patient records (778303 survivors and 20930 deaths) each containing 32 features, a number further reduced to only 8 features via the permutation importance method. Importantly, the 8 features can all be readily determined at admission: systolic blood pressure, heart rate, respiratory rate, temperature, oxygen saturation, gend ...[more]

PMID: 33201935

Similar Datasets

Project description:Background/Objective: Traumatic intracranial hemorrhage (tICH) accounts for significant trauma morbidity and mortality. Several studies have developed prognostic models for tICH outcomes, but previous models face limitations, including poor generalizability and limited accuracy. The objective was to develop a prognostic model and determine predictors of mortality using the largest trauma database in the U.S., applying rigorous analytical methodology with true hold-out-set model validation. Methods: We identified 248,536 patients in the National Trauma Data Bank (NTDB) from 2012 to 2016 with a diagnosis code associated with tICH. For each admission, we collected demographic information, systolic blood pressure, blood alcohol level (BAL), Glasgow Coma Score (GCS), Injury Severity Score (ISS), presence of epidural/subdural/subarachnoid/intraparenchymal hemorrhage, comorbidities, complications, trauma center level, and trauma center region. Our final study population was 212,666 patients following exclusion of records with missing data. The dependent variable was patient death. Linear support vector machine (SVM) classification was carried out with recursive feature selection. Model performance was assessed using holdout 10-fold cross-validation. Results: Cross-validation demonstrated a mean accuracy of 0.792 (95% CI 0.783-0.799). Accuracy, precision, recall, and AUC were 0.827, 0.309, 0.750, and 0.791, respectively. In the final model, high ISS, advanced age, subdural hemorrhage, and subarachnoid hemorrhage were associated with increased mortality, while high GCS verbal and motor subscores, current smoker, BAL beyond the legal limit, and level 1 trauma center were associated with decreased mortality. Conclusions: A linear SVM model was developed for tICH, with nine features selected as predictors of mortality. These findings are applicable to multiple hemorrhage subtypes and may benefit the triage of high risk patients upon admission. While many studies have attempted to create models to predict mortality in TBI, we sought to confirm those predictors using modern modeling approaches, machine learning, and true hold-out test sets, using the largest available TBI database in the U.S. We find that while the predictors we identify are consistent with prior reports, overall prediction accuracy is somewhat lower than prior reports when assessed more rigorously.

Project description:ImportanceThe Undiagnosed Diseases Network (UDN) is a national network that evaluates individual patients whose signs and symptoms have been refractory to diagnosis. Providing reliable estimates of admission outcomes may assist clinical evaluators to distinguish, prioritize, and accelerate admission to the UDN for patients with undiagnosed diseases.ObjectiveTo develop computational models that effectively predict admission outcomes for applicants seeking UDN evaluation and to rank the applications based on the likelihood of patient admission to the UDN.Design, setting, and participantsThis prognostic study included all applications submitted to the UDN from July 2014 to June 2019, with 1209 applications accepted and 1212 applications not accepted. The main inclusion criterion was an undiagnosed condition despite thorough evaluation by a health care professional; the main exclusion criteria were a diagnosis that explained the objective findings or a review of the records that suggested a diagnosis. A classifier was trained using information extracted from application forms, referral letters from health care professionals, and semantic similarity between referral letters and textual description of known mendelian disorders. The admission labels were provided by the case review committee of the UDN. In addition to retrospective analysis, the classifier was prospectively tested on another 288 applications that were not evaluated at the time of classifier development.Main outcomes and measuresThe primary outcomes were whether a patient was accepted or not accepted to the UDN and application order ranked based on likelihood of admission. The performance of the classifier was assessed by comparing its predictions against the UDN admission outcomes and by measuring improvement in the mean processing time for accepted applications.ResultsThe best classifier obtained sensitivity of 0.843, specificity of 0.738, and area under the receiver operating characteristic curve of 0.844 for predicting admission outcomes among 1212 accepted and 1210 not accepted applications. In addition, the classifier can decrease the current mean (SD) UDN processing time for accepted applications from 3.29 (3.17) months to 1.05 (3.82) months (68% improvement) by ordering applications based on their likelihood of acceptance.Conclusions and relevanceA classification system was developed that may assist clinical evaluators to distinguish, prioritize, and accelerate admission to the UDN for patients with undiagnosed diseases. Accelerating the admission process may improve the diagnostic journeys for these patients and serve as a model for partial automation of triaging or referral for other resource-constrained applications. Such classification models make explicit some of the considerations that currently inform the use of whole-genome sequencing for undiagnosed disease and thereby invite a broader discussion in the clinical genetics community.

Project description:ObjectiveSeveral prognostic models of suicide risk have been published; however, few have been implemented in Japan using longitudinal cohort data. The aim of this study was to identify suicide risk factors for suicidal ideation in the Japanese population and to develop a machine-learning model to predict suicide risk in Japan.Materials and methodsData was obtained from Wave1 Time 1 (November 2016) and Time 2 (March 2017) of the National Survey for Stress and Health in Japan, were incorporated into a suicide risk prediction machine-learning model, trained using 65 items related to trauma and stress. The study included 3,090 and 2,163 survey respondents >18 years old at Time 1 and Time 2, respectively. The mean (standard deviation, SD) age was 44.9 (10.9) years at Time 1 and 46.0 (10.7) years at Time 2. We analyzed the participants with increased suicide risk at Time 2 survey. Model performance, including the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity, were also analyzed.ResultsThe model showed a good performance (AUC = 0.830, 95% confidence interval = 0.795-0.866). Overall, the model achieved an accuracy of 78.8%, sensitivity of 75.4%, specificity of 80.4%, positive predictive value of 63.4%, and negative predictive value of 87.9%. The most important risk factor for suicide risk was the participants' Suicidal Ideation Attributes Scale score, followed by the Sheehan Disability Scale score, Patient Health Questionnaire-9 scores, Cross-Cutting Symptom Measure (CCSM-suicidal ideation domain, Dissociation Experience Scale score, history of self-harm, Generalized Anxiety Disorder-7 score, Post-Traumatic Stress Disorder check list-5 score, CCSM-dissociation domain, and Impact of Event Scale-Revised scores at Time 1.ConclusionsThis prognostic study suggests the ability to identify patients at a high risk of suicide using an online survey method. In addition to confirming several well-known risk factors of suicide, new risk measures related to trauma and trauma-related experiences were also identified, which may help guide future clinical assessments and early intervention approaches.

Project description:Risk assessment of in-hospital mortality of patients at the time of hospitalization is necessary for determining the scale of required medical resources for the patient depending on the patient's severity. Because recent machine learning application in the clinical area has been shown to enhance prediction ability, applying this technique to this issue can lead to an accurate prediction model for in-hospital mortality prediction. In this study, we aimed to generate an accurate prediction model of in-hospital mortality using machine learning techniques. Patients 18 years of age or older admitted to the University of Tokyo Hospital between January 1, 2009 and December 26, 2017 were used in this study. The data were divided into a training/validation data set (n = 119,160) and a test data set (n = 33,970) according to the time of admission. The prediction target of the model was the in-hospital mortality within 14 days. To generate the prediction model, 25 variables (age, sex, 21 laboratory test items, length of stay, and mortality) were used to predict in-hospital mortality. Logistic regression, random forests, multilayer perceptron, and gradient boost decision trees were performed to generate the prediction models. To evaluate the prediction capability of the model, the model was tested using a test data set. Mean probabilities obtained from trained models with five-fold cross-validation were used to calculate the area under the receiver operating characteristic (AUROC) curve. In a test stage using the test data set, prediction models of in-hospital mortality within 14 days showed AUROC values of 0.936, 0.942, 0.942, and 0.938 for logistic regression, random forests, multilayer perceptron, and gradient boosting decision trees, respectively. Machine learning-based prediction of short-term in-hospital mortality using admission laboratory data showed outstanding prediction capability and, therefore, has the potential to be useful for the risk assessment of patients at the time of hospitalization.

Dataset Information

Using the National Trauma Data Bank (NTDB) and machine learning to predict trauma patient mortality at admission.

Publications

Using the National Trauma Data Bank (NTDB) and machine learning to predict trauma patient mortality at admission.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets