Background: Profiling treatment in Japanese hospitals has rarely been conducted systematically with an administrative database. The study aims to present descriptive statistics of medical profiling and to examine the sources of variation in resource used for six common diseases.
Methods: Administrative records for 266,677 patients were analyzed to examine variation in length of stay (LOS) and total charge (TC) by hierarchical multiple linear regression for cases of ischemic stroke, ischemic heart disease (IHD), great vessel disease (GVD), respiratory neoplasm, gastric neoplasm and colonic neoplasm.
Results: Average LOS and TC increased with disease severity and invasiveness of surgical procedure. The coefficient of determination of the full model was highest for LOS in IHD (0.432), and for TC that was highest in GVD (0.702). Among various variable sets examined, surgical procedures explained largest variance in resource use.
Conclusion: With a standardized database derived from claims data, wide audience of stakeholders in Japanese healthcare will be able to access the profiling of practice or disease variation concerned.