MSR32 COVID-19 Beds’ Occupancy and Hospital Complaints: A Predictive Model

Objectives COVID-19 pandemic limited the number of patients that could be promptly and adequately taken in charge. The proposed research aims at predicting the number of patients requiring any type of hospitalizations, considering not only patients affected by COVID-19, but also other severe viral diseases, including untreated chronic and frail patients, and also oncological ones, to estimate potential hospital lawsuits and complaints. Methods An unsupervised learning approach of artificial neural network’s called Self-Organizing Maps (SOM), grounding on the prediction of the existence of specific clusters and useful to predict hospital behavioral changes, has been designed to forecast the hospital beds’ occupancy, using pre and post COVID-19 time-series, and supporting the prompt prediction of litigations and potential lawsuits, so that hospital managers and public institutions could perform an impacts’ analysis to decide whether to invest resources to increase or allocate differentially hospital beds and humans capacity. Data came from the UK National Health Service (NHS) statistic and digital portals, concerning a 4-year time horizon, related to 2 pre and 2 post COVID-19 years. Results Clusters revealed two principal behaviors in selecting the resources allocation. In case of increase of non-COVID hospitalized patients, a reduction in the number of complaints (-55%) emerged. A higher number of complaints was registered (+17%) against a considerable reduction in the number of beds occupied (-26%). Based on the above, the management of hospital beds is a crucial factor which can influence the complaints trend. Conclusions The model could significantly support in the management of hospital capacity, helping decision-makers in taking rational decisions under conditions of uncertainty. In addition, this model is highly replicable also in the estimation of current hospital beds, healthcare professionals, equipment, and other resources, extremely scarce during emergency or pandemic crises, being able to be adapted for different local and national settings.