A Machine Learning Prediction Model of Respiratory Failure Within 48 Hours of Patient Admission for COVID-19: Model
Siavash Bolourani1, Max Brenner1, Ping Wang1
1Feinstein Institutes for Medical Research, Northwell Health, Manhasset, NY, United States.
Insights
Machine learning accurately predicts COVID-19 respiratory failure within 48 hours. An XGBoost model identified high-risk patients using emergency department data, improving clinical decision-making for better patient outcomes.
Area of Science:
- Medical Informatics
- Computational Biology
- Public Health
Background:
- Predicting respiratory failure in COVID-19 patients is crucial for resource allocation and reducing mortality.
- Machine learning (ML) offers potential for improved clinical decision support in managing COVID-19 complexity.
Purpose of the Study:
- To develop an ML model for predicting respiratory failure within 48 hours of hospital admission.
- To utilize emergency department data for early identification of at-risk COVID-19 patients.
Main Methods:
- Trained and validated three ML models (two XGBoost, one logistic regression) using data from 11,525 COVID-19 patients.
- Employed cross-hospital validation and compared model performance against the Modified Early Warning Score.
- Utilized clinical and laboratory data commonly collected in the emergency department.
Main Results:
- The XGBoost model achieved the highest accuracy (0.919, AUC=0.77), outperforming other models and the Modified Early Warning Score.
- Key predictors included oxygen delivery type, age, Emergency Severity Index, respiratory rate, serum lactate, and demographics.
- The model demonstrated strong predictive performance for 48-hour respiratory failure.
Conclusions:
- The developed XGBoost model shows high predictive accuracy for COVID-19 respiratory failure.
- The model's clinical plausibility supports its potential use in clinical practice for early risk stratification.
- This ML approach can aid in timely intervention for patients at high risk of deterioration.
Background:
Predicting early respiratory failure due to COVID-19 can help triage patients to higher levels of care, allocate scarce resources, and reduce morbidity and mortality by appropriately monitoring and treating the patients at greatest risk for deterioration. Given the complexity of COVID-19, machine learning approaches may support clinical decision making for patients with this disease.
Objective:
Our objective is to derive a machine learning model that predicts respiratory failure within 48 hours of admission based on data from the emergency department.
Methods:
Data were collected from patients with COVID-19 who were admitted to Northwell Health acute care hospitals and were discharged, died, or spent a minimum of 48 hours in the hospital between March 1 and May 11, 2020. Of 11,525 patients, 933 (8.1%) were placed on invasive mechanical ventilation within 48 hours of admission. Variables used by the models included clinical and laboratory data commonly collected in the emergency department. We trained and validated three predictive models (two based on XGBoost and one that used logistic regression) using cross-hospital validation. We compared model performance among all three models as well as an established early warning score (Modified Early Warning Score) using receiver operating characteristic curves, precision-recall curves, and other metrics.
Results:
The XGBoost model had the highest mean accuracy (0.919; area under the curve=0.77), outperforming the other two models as well as the Modified Early Warning Score. Important predictor variables included the type of oxygen delivery used in the emergency department, patient age, Emergency Severity Index level, respiratory rate, serum lactate, and demographic characteristics.
Conclusions:
The XGBoost model had high predictive accuracy, outperforming other early warning scores. The clinical plausibility and predictive ability of XGBoost suggest that the model could be used to predict 48-hour respiratory failure in admitted patients with COVID-19.
Related Concept Videos
01:29Acute Respiratory Failure-V
Ensure that patients are monitored continuously for their response to therapy, including changes in...
01:21Acute Respiratory Failure-I
Definition: It is defined by specific criteria based on blood gas measurements. Hypoxemia happens when the partial pressure of oxygen (PaO2) falls below 60 mmHg. At the same time,...
01:21Acute Respiratory Failure-II
The underlying physiological abnormalities that contribute to hypoxemic respiratory failure include:
01:23Acute Respiratory Failure-IV
01:30Acute Respiratory Failure-III
01:19Respiratory Assessment: Purpose and Indications
Objectives and Importance:
The primary goal of respiratory assessment is to evaluate patients at early risk of clinical deterioration. Since respiratory distress often precedes other signs of declining health, breathing patterns and sounds become a...