Artificial intelligence-based prediction of cardiothoracic intensive care unit length of stay: A comparative machine learning approach.

BACKGROUND: Predicting prolonged intensive care unit (ICU) length of stay (LOS) remains challenging, and traditional statistical models often fail to capture a patient's complexities. Recent artificial intelligence (AI) tools, such as supervised machine learning (ML) algorithms, can provide new insights in the field. This study aimed to develop and cross-validate a collection of supervised ML models to predict prolonged ICU LOS. METHODS: Adult patients submitted to cardiac surgery were categorized into short (up to 2 days) or prolonged (more than 2 days) ICU LOS in the cardiothoracic ICU. Information gain analysis identified the most relevant predictors of prolonged ICU LOS for incorporation into the ML algorithms. Multiple supervised ML approaches were used to predict prolonged ICU LOS, comparing the performance of each model. The clinical applicability of the developed model was assessed with the creation of an interactive web-based application. RESULTS: Prolonged ICU LOS occurred in 48.8% of the 1387 patients included in the analysis. The strongest predictors of prolonged ICU LOS were the Sequential Organ Failure Assessment (SOFA) score and Vasoactive-Inotropic Score (VIS) at 24 hours after surgery, preoperative N-terminal pro-B-type natriuretic peptide and creatinine levels, and cardiopulmonary bypass time. Compared to other models, the random forest model had better predictive performance (area under the curve, 0.804), with 67.8% sensitivity and 81.8% specificity. CONCLUSIONS: Supervised ML models offer a reliable approach for predicting prolonged ICU LOS using perioperative data in patients admitted to a cardiothoracic ICU. We highlight the potential of clinical applicability of AI-enhanced models with the creation of a widely available interactive web-based application.