Prioritising Hospital Using ML

The hospital environment in particular offers the most difficult resource allocation problem for healthcare systems with demand often greatly exceeding capacity. Properly managing hospital resources whether beds or the like are crucial to drive better patient outcomes and unnecessary capacities otherwise a waste. One of the ways machine learning (ML) is solving this challenge is by incorporating real-time analytics and predictive modeling on data stored in historical formats. Therefore the current study aimed at analyzing machine learning algorithms for hospital resource prioritization. We survey multiple methods both supervised and unsupervised learning as well as a number of reinforcement models test these methods in the context of patient triage staffing prediction and resource optimization during peak loads like pandemics or natural disasters. The models are tested over several tasks based on their capacity to predict number of patient admissions resource scarcity prediction patient severity/ hospital capacity based treatment prioritization using Decision Trees and neural networks and SVM. Results. Machine learning improves decision-making processes by reducing wait times and more efficiently allocating resources to match room demand with the most adequate clinician/quirofano therefore predictive models generated by ML would lead us to tailor staff and equipment to patients needs. Yet realizing the promise of ML in hospitals is challenged by problems like data quality and integration with sewn-in hospital systems or requirements on interpretability for high-stakes medical decisions.