Forecasting weekly emergency department demand in a Portuguese private hospital - generalising to a medium-sized unit

• Autor(a) principal: Fonseca, Inês Corvo Marques da
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/154009
• topic: Healthcare; Emergency department; Machine learning; Time series; Multi-step-ahead forecasting; Domínio/Área Científica::Ciências Sociais::Economia e Gestão

The international Emergency Department (ED) overcrowding crisis affects both private and public Portuguese hospitals, which can be mitigated by an efficient medium-term operational planning. In this light, a Machine Learning multi-step-ahead predictive tool to forecast weekly ED arrivals in the largest unit of a private Portuguese healthcare provider, CUF, was developed. Linear Regression, SARIMAX and LSTM were evaluated and compared. SARIMAX, which obtained the best results, proved to have adequate predictive accuracy to support ED management. Additionally, the question of whether this model could be generalised to a medium-sized CUF ED unit was studied. Keywords: Healthcare, Emergency Department, Machine Learning, Time Series, Multi-step-ahead Forecasting, Model Generalisation.