Patient Expenditure Prediction in Healthcare Sectors

Predicting the expenses for patient expenditure in health sectors became an important task with various applications such as provider profiling, accountable management, and medical payment Adjustment. Previous approaches mainly deals with manually designed features and linear regression-based models, which require massive medical domain knowledge and show limited predictive performance. This paper gives us a multi-view deep learning framework which can help to predict upcoming healthcare expenses at the individual level which are based on historical data. Our multi-view approach can accurately model the mixed information, including patient demographic features, medical codes, drug usages, and facility utilization. We conducted forecasting of expense tasks on a day-to-day pediatric dataset that contains more than 390,000 patients. The empirical results displays that our proposed method outperforms all baselines for medical expenses calculation. These findings help toward better preventive care and accountable care in the healthcare domain.