Generalisable prediction model of surgical case duration: multicentre development and temporal validation

The recent multicentre study in Japan aimed to create a generalizable prediction model for surgical case duration, addressing the limitations of existing models that often rely on site-specific inputs. By analyzing data from two general hospitals over a four-year period, the researchers included a wide range of preoperative predictors, such as patient demographics and surgical context. They employed advanced machine learning techniques, including elastic-net and random forest models, to enhance the accuracy of predictions. The model underwent rigorous validation in 2024, demonstrating consistent performance across different centers and years. This development is significant as it can improve operating room scheduling, leading to better resource management and potentially enhanced patient outcomes.