Biomaterial Database

Developing machine learning models for prediction of mortality in the medical intensive care unit. 2022

Beatriz Nistal-Nuño

Department of Anesthesiology, Complejo Hospitalario Universitario de Pontevedra. Mourente s/n, 36071, Pontevedra. Spain. Electronic address: nistalnunobeatriz7@gmail.com.

OBJECTIVE Alert of patient deterioration is essential for prompt medical intervention in the Medical Intensive Care Unit (MICU). Logistic Regression (LR) has been used for the development of most conventional severity-of-illness scoring systems to anticipate the risk of mortality in the MICU. Machine Learning (ML) models such as probabilistic graphical models and Extreme Gradient Boosting (XGB) have demonstrated improved prediction accuracy in patient outcomes compared to LR. The aim was to compare three ML models to the SAPS, SAPS II, SAPS III, SOFA, serial SOFA, LODS, and OASIS for prediction of MICU mortality. METHODS A Bayesian Network (BN), Naïve Bayes network (NB), and a XGB model were developed. 9893 adult MICU-stays from the MIMIC-III database were studied. The primary outcome was MICU mortality prediction and the secondary outcome was 1-year mortality prediction. Data analyzed consisted on routine physiological measurements collected during 5 hours in the MICU, demographic and diagnoses/procedure features. The performance was evaluated by accuracy statistics, discrimination and calibration measures. Limitations of the study were discussed. RESULTS The AUROC for MICU mortality prediction was 0.919 for XGB, 0.905 for BN, and 0.864 for NB, while the conventional systems displayed much lower values with the serial SOFA having the best value (0.814). The Diagnostic Odds Ratio was ≤7.099 for all the conventional systems, reaching values of 30.115 for XGB and 22.648 for BN. The XGB achieved a sensitivity of 0.831 and specificity of 0.86 assuring an acceptable precision (0.528), whose values were much lower for the conventional systems. The Brier score was better for the ML models, except for the NB (0.119), with 0.072 for XGB and 0.081 for BN. CONCLUSIONS The XGB and BN substantially outperformed the conventional systems for discrimination, calibration and the accuracy statistics assessed. The NB showed inferior performance to the XGB and BN but improved the discrimination and all accuracy statistics of the conventional systems except for an inferior calibration and 1-year mortality discrimination. The XGB showed the best performance among all models. These ML models have the potential to improve the monitoring of MICU patients, which must be evaluated in future studies.

UI	MeSH Term	Description	Entries
D007362	Intensive Care Units	Hospital units providing continuous surveillance and care to acutely ill patients.	ICU Intensive Care Units,Intensive Care Unit,Unit, Intensive Care
D006801	Humans	Members of the species Homo sapiens.	Homo sapiens,Man (Taxonomy),Human,Man, Modern,Modern Man
D000069550	Machine Learning	A type of ARTIFICIAL INTELLIGENCE that enable COMPUTERS to independently initiate and execute LEARNING when exposed to new data.	Transfer Learning,Learning, Machine,Learning, Transfer
D000328	Adult	A person having attained full growth or maturity. Adults are of 19 through 44 years of age. For a person between 19 and 24 years of age, YOUNG ADULT is available.	Adults
D001499	Bayes Theorem	A theorem in probability theory named for Thomas Bayes (1702-1761). In epidemiology, it is used to obtain the probability of disease in a group of people with some characteristic on the basis of the overall rate of that disease and of the likelihood of that characteristic in healthy and diseased individuals. The most familiar application is in clinical decision analysis where it is used for estimating the probability of a particular diagnosis given the appearance of some symptoms or test result.	Bayesian Analysis,Bayesian Estimation,Bayesian Forecast,Bayesian Method,Bayesian Prediction,Analysis, Bayesian,Bayesian Approach,Approach, Bayesian,Approachs, Bayesian,Bayesian Approachs,Estimation, Bayesian,Forecast, Bayesian,Method, Bayesian,Prediction, Bayesian,Theorem, Bayes
D015233	Models, Statistical	Statistical formulations or analyses which, when applied to data and found to fit the data, are then used to verify the assumptions and parameters used in the analysis. Examples of statistical models are the linear model, binomial model, polynomial model, two-parameter model, etc.	Probabilistic Models,Statistical Models,Two-Parameter Models,Model, Statistical,Models, Binomial,Models, Polynomial,Statistical Model,Binomial Model,Binomial Models,Model, Binomial,Model, Polynomial,Model, Probabilistic,Model, Two-Parameter,Models, Probabilistic,Models, Two-Parameter,Polynomial Model,Polynomial Models,Probabilistic Model,Two Parameter Models,Two-Parameter Model
D016208	Databases, Factual	Extensive collections, reputedly complete, of facts and data garnered from material of a specialized subject area and made available for analysis and application. The collection can be automated by various contemporary methods for retrieval. The concept should be differentiated from DATABASES, BIBLIOGRAPHIC which is restricted to collections of bibliographic references.	Databanks, Factual,Data Banks, Factual,Data Bases, Factual,Data Bank, Factual,Data Base, Factual,Databank, Factual,Database, Factual,Factual Data Bank,Factual Data Banks,Factual Data Base,Factual Data Bases,Factual Databank,Factual Databanks,Factual Database,Factual Databases
D017052	Hospital Mortality	A vital statistic measuring or recording the rate of death from any cause in hospitalized populations.	In-Hospital Mortality,Mortality, Hospital,Hospital Mortalities,In Hospital Mortalities,In Hospital Mortality,Inhospital Mortalities,Inhospital Mortality,Mortalities, In-house,Mortalities, Inhospital,Mortality, In-Hospital,Mortality, Inhospital,Hospital Mortalities, In,Hospital Mortality, In,In-Hospital Mortalities,In-house Mortalities,In-house Mortality,Mortalities, Hospital,Mortalities, In Hospital,Mortalities, In house,Mortalities, In-Hospital,Mortality, In Hospital,Mortality, In-house