透過您的圖書館登入
IP:3.92.96.247
  • 學位論文

應用人工智慧於白內障術後眼內炎之研究

A Study of Applying Artificial Intelligence to the Incidence of Endophthalmitis after Cataract Surgery

指導教授 : 張俊郎
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


老年人為一般白內障的好發族群,而白內障最容易引起視力不佳以及視盲,在全球統計中造成視盲最常見的原因為白內障,而白內障手術後眼睛有外傷或因身體的抵抗力下降,容易導致眼睛遭到眼內炎感染,眼內炎是令人聞之色變的可怕疾病,會讓病患視力嚴重受損甚至失明,如果不幸發生失明常會引發醫療糾紛和遺憾,眼科醫療糾紛屬於前幾名的案件,手術後發生眼內炎機率大約0.07~0.13%,機率不高,但是發生後對視力造成很大的威脅。 根據中央健康保險局醫療費用申報資料顯示,在2000年白內障手術為11萬人次,但2011年時人數大幅上升至171,068人次,成長幅度約為55%。這幾年白內障病患越來越多,而且年齡有下降的趨勢,慢慢成為流行病學領域的重要議題,讓人不得不重視這個問題,本研究使用全民健保資料庫資料做為研究對象,探討白內障手術後發生眼內炎感染情形及其危險因子,建構預測系統來預測術後眼內炎的發生,進而提升白內障手術品質。 本研究運用人工智慧方法建構六種預測評估模型,再以案例式推理技術建構評估系統,對於醫師或相關人員做為醫療診斷上輔助使用,以提升醫療服務品質。在模型中倒傳遞類神經網路優於其他模型,該模型平均測試準確率為91.35%,平均接收操作特性曲線為0.834皆是所有模型當中最佳。而粒子群演算法權重值結合案例式推理較其它權重值之結合模型有較佳之評估系統準確率達90.3%。

並列摘要


The elderly is the group prone to cataract outset, and cataract most likely causes poor vision and blindness. According to global statistics, the most common cause of blindness is cataract, while post-operative eye trauma or weakened body defenses are likely to lead to a kind of eye infection known as endophthalmitis. Endophthalmitis is a dreaded and terrible disease, as it causes severe vision impairment and even blindness in patients. In unfortunate events of blindness, medical malpractice related disputes and regret usually take place. Cases of ophthalmic related medical disputes are among the highest in terms of occurrence. The rate of post-operative endophthalmitis rate is about 0.07~0.13%, a low probability but a major threat to vision following its occurrence. According to the healthcare fee claim data of National Health Insurance Administration, Ministry of Health and Welfare (NHI), 110,000 people underwent cataract surgery in 2000, but the number increased to 171,068 in 2011, an increase of approximately 55%. In recent years, the number of cataract patients has increased, and the age has shown a downward trend. Cataract has become an important issue in the field of epidemiology, making it the problem of greater public concern. In this study, the information stored in the NHI database was adopted as the research participant in order to explore the endophthalmitis status and the risk factors, thereby constructing a predictive system that predicts occurrences of post-operative endophthalmitis and improves f the quality of cataract surgery. In this study, the Artificial Intelligence (AI) approach was adopted to construct six predictive assessment models. Through the case-based reasoning technology, the assessment system was constructed to serve as an aid for doctors or related personnel in medical diagnosis and improve the quality of medical services. The research results indicate that the back-propagation neural network was superior to the other models. The model’s average test accuracy was 91.35%, and the average receiver operating characteristics curve was 0.834, the best among all the models. On the other hand, the Particle Swarm Optimization (PSO) weight values combined with the case-based reasoning had the better assessment system accuracy ratio of 90.3%, compared to the weight values of the other combined models.

參考文獻


1.中央健康保險局,全民健康保險支付標準,取自http://www.nhi.gov.tw/webdata/webdata.aspx?menu=20&menu_id=710&webdata_id=3633,參考日期:2013/6/12。
2.中央健康保險局,眼科審查汱意事項及白內障手術事前審查作業規定,取自http://www.nhi.gov.tw/,參考日期:2013/6/14。
3.中華民國眼科醫學會,中老年視覺問題診療指引手冊,取自http://www.bhp.doh.gov.tw/BHPNet/Web/HealthTopic/TopicArticle.aspx?id=200712250416&parentid=200712250086,參考日期:2013/5/11。
4.王派洲 (2008),資料探勘概念與方法,滄海書局。
5.王偉驎、林文燦、賴政皓、陳慧敏 (2008),「應用資料探勘技術提升急診醫學檢傷分類之一致性—以台灣某醫學中心急診醫學部為例」,品質學報,第15卷,第4期,頁283-291。

被引用紀錄


邱達暐(2015)。應用兩階段隨機規劃探討半導體封裝產業之產能規劃問題〔碩士論文,國立虎尾科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0028-1308201515194200

延伸閱讀