| Project/Area Number |
18K14950
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 47060:Clinical pharmacy-related
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2018-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2019: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2018: ¥2,990,000 (Direct Cost: ¥2,300,000、Indirect Cost: ¥690,000)
|
|---|
| Keywords | 薬剤疫学 / 機械学習 / 時系列データ / 異常検知 / ビッグデータ / 薬剤有害事象 / データ駆動型 |
|---|
| Outline of Final Research Achievements |
We explored whether, using machine learning techniques, automated anomaly detection of drug-adverse events was feasible from large-scale database (eg, claims data). We selected two case examples; the first example was SGLT-2 inhibitor and its adverse events such as renal insufficiency or dehydration, and the second one was bisphosphonate and osteonecrosis of the jaw. Machine learning techniques successfully detected the anomaly points from a time-series data of large-scale database. However, the detected anomaly points were sensitive to the model or the administrative code selection of adverse event.
|
| Academic Significance and Societal Importance of the Research Achievements |
病名の入力に関しては医療機関・医師レベルにより方針が異なるのが日本の実情と思われる。機械学習を用いた検出においても、その精密さはこれら診療プラクティスの違いに大きく依存した。したがって、この技術を実用面で用いる前提として、病名入力の方針について何らかの統一的な指針を整備することが必要であることを示唆した研究である。
|
