Presentation (CTI)

OBJECTIVES: For effective population health management (PHM), it is useful to have benchmarks for the prevalence of different diseases. For example, if the prevalence of a certain disease in a certain municipality exceeds the benchmark, it is highly likely that PHM that reduces the prevalence of that disease will be feasible and effective. There are short-term (marginal) or long-term (average) cost reduction by lowering the prevalence. In this study, we propose both cost calculation methods and compare the results.
METHODS: Nissay Information Technology has developed and patented a method for decomposing medical expenses into each marginal cost as follows: costs associated with each disease are allocated to that disease, and not-associated costs are allocated pro rata. The average cost can be estimated as the coefficient of a regression model in which logarithm of medical expenses are the explained variable and dummy variables for each disease are the explanatory variables. Using a health insurance claims database, “Wellness-Star☆”, owned by Nippon Life Insurance, we extracted men aged 55-59 who had a confirmed diagnosis of hypertension, diabetes, or dyslipidemia in 2021, and we estimated the marginal and average cost.
RESULTS: The number of patients analyzed was 46,293. The average annual medical expenses for the subjects of the analysis were 355,060 JPY. Of this, the marginal burden for hypertension, diabetes, and dyslipidemia was 31,630 JPY, 41,550 JPY, and 17,120 JPY, respectively. The average burden was 60,200 JPY, 166,470 JPY, and 27,830 JPY, respectively.
CONCLUSIONS: The marginal cost was about one quarter of the average cost. It is reasonable that the marginal cost is smaller than the average cost. The marginal cost can be used as an indicator of the amount of medical cost reduction that should be achieved in the short term through PHM.