Budget Impact Analysis of Treatment Flow Optimization in Epilepsy Patients: Estimating Potential Impacts with Increased Referral Rate to Specialized Care.

We developed a Markov model to simulate a treatment flow of epilepsy patients who refer to specialized care from non-specialized care, and to surgery from specialized care for estimation of patient distributions and expenditures caused by increasing the referral rate for specialized care. This budget impact analysis of treatment flow optimization in epilepsy patients was performed as a long-term simulation using the Markov model by comparing the current treatment flow and the optimized treatment flow. In the model, we simulated the prognosis of new onset 5-year-old epilepsy patients (assuming to represent epilepsy occurring between 0 and 10 years of age) treated over a lifetime period. Direct costs of pharmacotherapies, management fees and surgeries are included in the analysis to evaluate the annual budget impact in Japan. In the current treatment flow, the number of refractory patients treated with four drugs by non-specialized care were estimated as 8766 and yielded JPY5.8 billion annually. However, in the optimized treatment flow, the number of patients treated with four drugs by non-specialized care significantly decreased and who continued the monotherapy increased. The costs for the four-drug therapy by non-specialized care were eliminated. Hence cost-saving of JPY9.5 billion (-5% of the current treatment flow) in total national expenditures would be expected. This study highlights that any policy decision-making for referral optimization to specialized care in appropriate epilepsy patients would be feasible with a cost-savings or very few budget impacts. However, important information in the decision-making such as transition probability to the next therapy or excuse for sensitive limitations is not available currently. Therefore, further research with reliable data such as big data analysis or a national survey with real-world treatment patterns is needed.