Risk prediction for future 6-month healthcare resource utilization in Maine

Understanding the future costs of the healthcare service utilization in patients can benefit the resource allocation management. The aim of this study is to develop a risk stratification model for healthcare resource utilization in future 6 months of patients in the state of Maine. A retrospective cohort of 1,273,114 patients was constructed to derive a decision-tree-based model to estimate the risk of resource utilization between January 1, 2013 and June 30, 2013, using the preceding 12-month clinical historical data. The model was validated with a prospective cohort of 1,358,153 patients by testing total costs between July 1, 2013 and December 31, 2013. Prospective results showed that the sensitivities of the model varied between 0.057 and 0.800, with confidence levels varying between 0.858 and 0.937 at all risk levels. Potential economic impacts of the model on healthcare resource utilization were explored. Future applications of our model will enable a more efficient resource allocation and targeted care intervention.