利用開放資料建立藥物腎毒性警示系統之雛形－以健康存摺為例

黃仁甫 ； 劉德明

《TANET2017 臺灣網際網路研討會》 2017年 (2017/10) Pp. 1728-1734

https://doi.org/10.6728/TANET.201710.0306

摘要

背景：台灣的末期腎臟病盛行率與發生率皆高居世界第一，如何改善此趨勢是所有醫藥專家學者的努力方向，而防止藥物腎毒性是其中的焦點課題之一；過去病患的就診資料因分散於不同醫療院所，難以一窺全貌。台灣中央健康保險署於民國一百零三年九月推出的「健康存摺」系統，除讓民眾得以存取自身的跨院所就診資料，亦給予完整評估用藥與腎功能關係的機會。研究目的：建立一藥物腎毒性警示參考系統之雛形，可供醫師與病患在跨院就醫過程中所使用，借此提升電子病歷的使用意義。方法：本研究首先透過「健康存摺」系統所提供的資料下載服務，依照腎臟疾病臨床治療指引分析就診資料中的血液檢驗結果以判定病患的腎功能等級。同時運用中央健保署藥物資訊查詢系統來辨識用藥紀錄中的藥物成份；此外，基於藥物腎毒性文獻建立藥物成分及藥物腎毒性規則庫，最後結合腎功能資料與用藥紀錄，根據藥物腎毒性規則庫的判斷產出自動化報告。結果：本研究透過整合多項嚴謹與準確的開放資料所建立的流程模型，可自動彙整跨院臨床就診資料並自動擷取與更新資料源，經虛擬測試個案驗證，成功產出自動化藥物腎毒性分析報告；本系統的個人化藥物腎毒性分析報告可提供醫師與病患於臨床診療時，特別於跨院與基層診所就醫時的完整參考。

關鍵字

藥物腎毒性 ； 健康存摺 ； 開放資料 ； 臨床資料 ； 醫學資訊

並列摘要

Background: Both prevalence and incidence of end-stage renal disease in Taiwan rank first globally. To prevent drug nephrotoxicity is one of the critical issues while all medical professionals aim to improve the trend. Previously, the patients' medical records scattered in different institutions, which formed a barrier to access the complete medical records. The National Health Insurance (NHI) Administration in Taiwan has launched My Health Bank in September 2013, which allows individuals to obtain their personal cross-institutional medical records. Additionally, the system also provided an opportunity to comprehensively evaluate the association of drug history and renal function. Aim: This study aims to set up a prototype warning system for drug nephrotoxicity for doctors and patients in the cross-institutional environment to achieve the meaningful uses of electronic health records. Methods: Initially, this study downloaded patients' medical records from the My Health Bank. The results of biochemistry tests were analyzed and patients' renal function was categorized according to clinical guideline for renal diseases. Meanwhile, ingredients of drugs were also analyzed based on the drug information system of NHI and a rule base for drug nephrotoxicity which was established from literatures. Finally, a personal report was automatically produced after applying the rule base on the data of renal function and drug ingredients. Results: After integrating several trusted and accurate clinical open data, this study set up a prototype system which can process cross-enterprise medical records, capture and update the data sources automatically. The results of test cases demonstrated that this prototype warning system could generate a personalized report of drug nephrotoxicity. The personalized report of drug nephrotoxicity could provide a comprehensive reference for doctors and patients, especially in the cross-institutional environment.

並列關鍵字

drug nephrotoxicity ； My Health Bank ； open data ； medical record ； medical informatics