Predicting hospital outpatient volume using XGBoost: a machine learning approach.

Area of Science:

• Health Informatics
• Environmental Health
• Machine Learning in Healthcare

Background:

• Hospital outpatient volume is influenced by environmental conditions and healthcare resource availability.
• Accurate prediction of outpatient demand is crucial for operational efficiency and resource allocation.
• Existing forecasting models may not fully capture the complex interplay of factors affecting outpatient volume.

Purpose of the Study:

• To develop and evaluate a predictive model for daily hospital outpatient volume using the XGBoost algorithm.
• To compare the performance of XGBoost against Seasonal AutoRegressive Integrated Moving Average with exogenous regressors (SARIMAX) and Random Forest (RF) models.
• To identify key factors influencing hospital outpatient demand.

Main Methods:

• Utilized daily climate data (temperature, precipitation, PM2.5), historical outpatient volumes, and outpatient specialist numbers from January 1, 2014, to October 31, 2024.
• Preprocessed data included handling missing values and encoding categorical variables.
• Assessed model performance using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), and R-squared (R²).

Main Results:

• The XGBoost model demonstrated superior predictive accuracy over SARIMAX and RF models.
• XGBoost achieved the lowest MAE, RMSE, and MAPE, and the highest R².
• Significant predictors included the number of outpatient specialists, temporal variables, average temperature, and PM2.5 levels.

Conclusions:

• XGBoost is a highly effective machine learning algorithm for predicting hospital outpatient volume.
• Accurate demand forecasting enables proactive resource allocation and improved hospital service capacity.
• Integrating environmental and resource data enhances the precision of outpatient volume predictions.