Assuming that each patient is examined by two different diagnostic tests, new and standard, we deal with in this paper the problem of comparing the two tests based on categorical scores measured by multiple readers. An exact method is developed based on conditional inference for a comparison parameter. The method is applied to the data for comparing a new test and the standard test that are developed for diagnosing whether a patient is contracted with epidermal cyst or ganglion.
For Japanese pharmaceutical companies, one of the critical issues in the design of multi-regional clinical trials (MRCT) is the number of Japanese patients. In the document issued by the Ministry of Health, Labour and Welfare (MHLW) in 2007, “Basic principles on Global Clinical Trials,” an issue was raised about the Japanese subgroup sample size and the importance of “consistency of results.” Ideas on how to calculate the sample size of a given region and the regions overall have been proposed by Uesaka (2006, 2009). For this purpose, he identified two types of consistency criteria and two types of efficacy criteria. For both efficacy and consistency these types consist of either 1) comparing the results of a specific region with those of the other regions combined, or 2) comparing the results of a specific region with those of the regions overall. Based on the efficacy criterion 2, Sekiguchi et al. (2007) examined the sample size of the Japanese subgroup in an oncology MRCT. In this study, we present a simulation that shows the relationship of Japanese subgroup sample size to both types of efficacy criteria in an oncology MRCT.
The validity of the diagnostic criteria for Metabolic Syndrome (MS) has been questioned since the syndrome was introduced. To verify their validity, the best strategy would be a cohort study. However, waist circumference (WC), which is necessary for the diagnosis of MS, was introduced to health examinations only recently. If a current regression model of WC could be transported to a previous sample, a cohort study using estimates of WC would be possible. We propose a sufficient condition for the transportability from a mathematical point of view. Using data from the Radiation Effect Research Foundation, we examined whether the proposed condition may be statistically confirmed in a real dataset. The regression model for weight, which is strongly associated with WC, was used to verify the validity of the condition. The relative risk of death from MS-related causes was obtained. The results indicated that the risk was lower with a greater WC.