2022 Fiscal Year Final Research Report
Development of exact tests where the primary endpoint is a categorical variable
| Project/Area Number |
18K11195
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 60030:Statistical science-related
|
|---|
| Research Institution | Yokohama City University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2018-04-01 – 2023-03-31
|
| Keywords | 正確検定 / 臨床試験 |
|---|
| Outline of Final Research Achievements |
Many analysis methods for categorical variables such as binary variables use large-sample approximation. However, especially in physician-initiated clinical trials in academia, the number of patients cannot be large and the sample size is often small. Since methods using large-sample approximations are often not appropriate in such cases, the present study was undertaken to develop exact methods that do not use approximations. This study focused on the following two points:
1. to develop exact tests for the comparison of predictive values between two screening tests
2. to develop a design that simultaneously considers efficacy and safety in clinical trials to select the best treatment candidate from multiple treatment candidates
|
| Free Research Field |
医学統計
|
| Academic Significance and Societal Importance of the Research Achievements |
医学分野では，疾患の確定診断を行う前にスクリーニング検査をすることが多いが，十分な診断精度をもっているとは言えない領域もある．たとえば胃癌の腹膜播種（転移）は術前に診断する必要があるが，その診断性能には改善の余地が残されている．そこで新たな検査法を開発する際には，既存の検査法と診断性能を比べる必要があるが，上述の胃癌の例であれば，的中率による比較で性能を比較したい．これに対する比較手法はいくつか提案されているが，症例数が十分でない場合には，既存の手法では許容できない程度に誤って新規法がよいと結論づける可能性があったが，本研究によって提案された正確な手法であればそのようなことは起きない．
|
