Using machine learning to predict severe hypoglycaemia in hospital.

Background Machine learning carries considerable promise to improve healthcare delivery. Clinical outcomes that are objectively measured and have serious but preventable consequences are ideal targets for prediction and intervention. Hypoglycemia, defined as a blood glucose of 3.9 mmol/L (70 mg/dL) or lower, meets these criteria. Objective To predict the risk of hypoglycemia using machine learning techniques in hospitalized patients. Methods Retrospective cohort study of patients hospitalized under general internal medicine (GIM) and cardiovascular surgery (CV) at a tertiary-care teaching hospital in Toronto, Ontario. Three models were generated using supervised machine learning: Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression, gradient boosted trees, and a recurrent neural network. Each model included baseline patient data and time-varying data. Natural language processing was used to incorporate text data from physician and nursing notes. Results We included 8,492 GIM admissions and 8,044 CV admissions. Hypoglycemia occurred in 16% of GIM admissions and 13% of CV admissions. The area under the curve for the models in the held-out validation set was approximately 0.80 on the GIM ward and 0.82 on the CV ward. When the threshold for hypoglycemia was lowered to 2.9 mmol/L (52 mg/dL), similar results were observed. Among the patients at the highest decile of risk, the positive predictive value was approximately 50% and the sensitivity was 99%. Interpretation Machine learning approaches can accurately identify patients at high risk of hypoglycemia in hospital. Future work will involve evaluating whether implementing this model with targeted clinical interventions can improve clinical outcomes. This article is protected by copyright. All rights reserved.