| Budget Amount *help |
¥8,320,000 (Direct Cost: ¥6,400,000、Indirect Cost: ¥1,920,000)
Fiscal Year 2014: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
Fiscal Year 2013: ¥2,470,000 (Direct Cost: ¥1,900,000、Indirect Cost: ¥570,000)
Fiscal Year 2012: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
|
|---|
| Outline of Final Research Achievements |
In this research project, we analyzed the medical data. We first built a new database over one million patients so that we could use the DPC (Diagnostic Procedure Combination) data obtained from various hospitals. Next, we developed new estimation and testing methods for the Box-Cox transformation model (BC model). The BC model is usually estimated by the maximum likelihood estimator (BC MLE) assuming normality of error terms. However, the estimator is usually inconsistent. We proposed new consistent estimators by the modification of the likelihood function and Hausman type tests for the model. Using the newly developed database and estimation and testing methods, we then analyzed the length of stay (LOS) in a hospital for cataract and diabetes patients. To control the medical expenditure by reducing LOS is a very important issue in Japan. We found out that there were large differences in LOS among hospitals even after controlling types of diseases and characteristics of patients.
|
