Intelligent load balancing strategy for mitigating hospital overcrowding during pandemic: A predictive–prescriptive approach

The coronavirus disease 2019 (COVID-19) pandemic has placed unprecedented strain on healthcare systems globally, including in Indonesia. A key challenge has been the sharp rise in patients requiring intensive care, often exceeding facility capacity. In Indonesia, COVID-19 patients tend to concentrate in primary referral hospitals, complicating patient management. This study introduces an intelligent strategy that integrates a long short-term memory (LSTM) prediction model with an integer programming (IP) optimization model to improve hospital load balancing. The LSTM model forecasts patient surges from historical data, which are then used to inform the IP model for optimizing patient transfers from overloaded to underutilized hospitals. A case study involving six referral hospitals in Balikpapan City, Indonesia, demonstrates the approach. Results show high prediction accuracy (R2 =#xthinsp;0.96) and a 40% reduction in patient load at the main referral hospital, while maintaining an average load of 71% across all hospitals. This integrated predictive–prescriptive method provides an effective solution for alleviating overcrowding and supports more responsive healthcare capacity planning during pandemics and other public health emergencies.