| Project/Area Number |
15K08847
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Medical and hospital managemen
|
|---|
| Research Institution | Kochi University |
|---|
Principal Investigator |
OKUHARA YOSHIYASU 高知大学, 教育研究部医療学系連携医学部門, 教授 (40233473)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
畠山 豊 高知大学, 教育研究部医療学系連携医学部門, 准教授 (00376956)
渡部 輝明 高知大学, 教育研究部医療学系連携医学部門, 講師 (90325415)
中島 典昭 国立研究開発法人国立がん研究センター, 情報統括センター, 研究員 (00335928)
片岡 浩巳 川崎医療福祉大学, 医療技術学部, 教授 (80398049)
寺田 典生 高知大学, 教育研究部医療学系臨床医学部門, 教授 (30251531)
堀野 太郎 高知大学, 教育研究部医療学系臨床医学部門, 講師 (90448382)
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Project Status |
Completed (Fiscal Year 2017)
|
| Budget Amount *help |
¥4,940,000 (Direct Cost: ¥3,800,000、Indirect Cost: ¥1,140,000)
Fiscal Year 2017: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2016: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2015: ¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000)
|
|---|
| Keywords | 病院情報システム / 機械学習 / 病名評価 / データマイニング / 電子カルテ / 病名信頼度 |
|---|
| Outline of Final Research Achievements |
We aimed to obtain reliability of disease name from data of hospital information system.
For the definite disease name of the hospital information system, the proportion of examination, prescription and treatment being conducted in the case group and the control group was obtained, the item with the large value was selected, and the rule was obtained by the decision tree model. In addition, it was possible to identify the true disease name by finding the ratio of what kind of item taken in the model to the name of disease and classifying it by the cluster method as the characteristic quantity characterizing the item . Moreover, we applied the obtained rule to the verification data and found that some degree of accuracy can be obtained only with the order entry data. To use the narrative part of the electronic medical record as information, it turned out that applying a latent topic model is effective.
|
