Research of an automatic coding system for medical terms using distributed methods of in the network
| Project/Area Number |
16590418
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Medical sociology
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
HATANO Kenji The University of Tokyo, Graduate School of Medicine, Research Associate, 大学院・医学系研究科, 助手 (60311619)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
OHE Kazuhiko The University of Tokyo, University Hospital, professor, 医学部附属病院, 教授 (40221121)
|
|---|
| Project Period (FY) |
2004 – 2005
|
| Project Status |
Completed (Fiscal Year 2005)
|
| Budget Amount *help |
¥1,700,000 (Direct Cost: ¥1,700,000)
Fiscal Year 2005: ¥700,000 (Direct Cost: ¥700,000)
Fiscal Year 2004: ¥1,000,000 (Direct Cost: ¥1,000,000)
|
|---|
| Keywords | medical term / automatic coding / ICD10 / remote method / XML Web Service / リモートメソッド |
|---|
| Research Abstract |
This research targeted in an automatic classification by the concept of the medical terms. We made trial for automatic coding of ICD10 (the tenth edition in international disease classification) classification code to the name of a disease term that was the most basic medical terms, using semantic analysis from the description of the terms and the classifications.
With the conversion of known coding rules of ICD10 into the regular expression for program processing, the rate of correct coding was insufficient (2-30 percent). For more accurate semantic analysis of the terms, we decomposed the terms into several elements (i.e. modifiers and disease names) and researched about how codes change by attachment of modifiers to disease names. 9600 patterns of ICD10 code change were extracted by resolving it to the element of the disease terms and the modifier from the standard disease name master and some regularity of patterns of code changes was found for automatic coding.
We also developed a web server that offered remote methods of XML Web service for automatic coding of disease terms (URL http://www.dis.h.u-tokyo.ac.jp/byomei/webservices) for an estimation of utility of the automatic coding.
|
Report
(3 results)
Research Products
(8 results)
