Prediction of Length of Stay for Elective Procedures in Irish Hospitals

Length of stay is a measure of hospital service delivery efficiency used throughout the world. The ability to predict a patient’s length of stay has the potential to influence management decisions on when other patients may be scheduled for surgery and other elective procedures. Additionally, it would allow for the enhancement of resource allocation for nursing and support staff based on the expected patient dependency and for how long. This research project attempted to predict whether a patient will stay longer than the typical length of stay for a scheduled procedure. A variety of predictive models were developed with the final model, a gradient boosted decision tree, correctly identifying approximately 4 out of 5 long stay patients. The research output is developed to the standard of being suitable for deployment in one of Ireland’s largest acute hospitals and the methodology such that it could be implemented in other Irish hospitals with appropriate datasets