TRENDREADER(TR)TMを用いた京都大学附属病院インシデントレポート解析方法の開発
キーワード:
risk management,
patient safety,
incident/accident report,
open-ended description,
text mining.
2010 年 13 巻 p. 305-315
詳細
2010 年 13 巻 p. 305-315
Recently the necessity of the accident prevention has made the numbers of the hospitals increase which investigate and analyze incident reports.There are two ways to describe incidents and accidents; open-ended and precode-choice. It is unusual to analyze precode-choice descriptions while usual to analyze open-ended, for the costs and technical difficulty prevent from analyzing. This study investigates effectiveness for medical risk management of TR as follows: 1) Calculation of both of them to examine the usefulness of the precode-choices, 2) Observation of behaviors of the calculated out-comes which depends on time in some campaigns of the accident prevention to estimate results of campaigns, 3) Support of occupation as selecting of important reports to improve human criteria for selection.
