Predicting the need for intubation in the first 24 h after critical care admission using machine learning approaches

Abstract

Early and accurate prediction of the need for intubation may provide more time for preparation and increase safety margins by avoiding high risk late intubation. This study evaluates whether machine learning can predict the need for intubation within 24 h using commonly available bedside and laboratory parameters taken at critical care admission. We extracted data from 2 large critical care databases (MIMIC-III and eICU-CRD). Missing variables were imputed using autoencoder. Machine learning classifiers using logistic regression and random forest were trained using 60% of the data and tested using the remaining 40% of the data. We compared the performance of logistic regression and random forest models to predict intubation in critically ill patients. After excluding patients with limitations of therapy and missing data, we included 17,616 critically ill patients in this retrospective cohort. Within 24 h of admission, 2,292 patients required intubation, whilst 15,324 patients were not intubated. Blood gas parameters (PaO2, PaCO2, HCO3−), Glasgow Coma Score, respiratory variables (respiratory rate, SpO2), temperature, age, and oxygen therapy were used to predict intubation. Random forest had AUC 0.86 (95% CI 0.85–0.87) and logistic regression had AUC 0.77 (95% CI 0.76–0.78) for intubation prediction performance. Random forest model had sensitivity of 0.88 (95% CI 0.86–0.90) and specificity of 0.66 (95% CI 0.63–0.69), with good calibration throughout the range of intubation risks. The results showed that machine learning could predict the need for intubation in critically ill patients using commonly collected bedside clinical parameters and laboratory results. It may be used in real-time to help clinicians predict the need for intubation within 24 h of intensive care unit admission.

Development and Internal Validation of Supervised Machine Learning Algorithms for Predicting the Risk of Surgical Site Infection Following Minimally Invasive Transforaminal Lumbar Interbody Fusion

“…In its most basic sense, machine learning uses programmed algorithms that learn and optimize their operations by analyzing input data to make predictions within an acceptable range (15)(16)(17)(18). In this study, 10-fold cross-validation was adopted, which means patients were randomly divided into a training set and a validation set at a ratio of 9:1 in each round.…”

A multi-task Gaussian process self-attention neural network for real-time prediction of the need for mechanical ventilators in COVID-19 patients

“…Our model is generic, and the observational time window setting depends on the specific application. In a critical scenario where the hospitals fall short of medical beds and ventilators, it is crucial to predict early (like within the first 24-72 hours after hospital admission) about a patient’s future condition, see also [24] , [25] , [26] for similar settings. Another interesting application would be to use a patient’s 72 hours of data counting backward from the time point of the event of intubation happens.…”

The mNUTRIC Score as a Tool to Predict Mortality and Increased Resource Utilization in Intensive Care Patients with Sepsis.

“…28,29 A similar approach was employed for predicting the need for intubation. [30][31][32] Siu et al used machine learning to develop a model predicting the need for intubation during the rst 24 hours after ICU admission. The parameters required for the model were as follows: blood gas results, the Glasgow Coma Score, respiratory rate, oxygen saturation, temperature, age, and parameters of oxygen therapy.…”

“…The reported AUC of the model was 0.86 (95% CI 0.85-0.87). 30 The approaches presented above often required advanced automated electronical real-time data collection and analysis. This kind of know-how is not always available in an ICU setting.…”