陰性闌尾炎的偵測－決策樹與Alvarado計分系統之效能比較

Detective Effectiveness of Acute Appendicitis: Comparison of Decision Tree and Alvarado Score

黃德佳(Te-Chia Huang) ； 鄭滄祥(Hsang-Hsiang Cheng) ； 黃慧珠(Hui-Chu Huang) ； 曾婉婷(Wan-Ting Tseng)

《電子化優質健康照護》 2007年 (2007/11) Pp. 411-427

https://doi.org/10.29651/ECARE.200711.0411

摘要

急性闌尾炎是外科手術中最常見的疾病。雖然急性闌尾炎的治療直接而容易，但是卻難以準確地診斷急性闌尾炎的發生。目前並沒有任何的血液檢查或影像檢查可以準確地診斷出急性闌尾炎。對醫師而言，準確地檢測出該病症以減低病患因闌尾炎而導致闌尾穿孔，或降低因誤判而產生陰性闌尾切除率，藉以維護病患安全，仍是一種挑戰。
資訊科技的進步提供了另一種協助診斷的方法，本研究嘗試利用資料探勘中的決策樹分類分析法C4.5，以醫師對病患徵候症狀判斷與血液檢查的結果為基礎，期能建構一個可協助醫師正確診斷闌尾炎的分類模式，藉以提升闌尾炎診斷的醫療品質，並有效地降低陰性闌尾切除案例的發生。本研究將利用Alvarado計分法作為闌尾炎預測效能比較基準，研究結果顯示以C4.5所建構出的闌尾炎預測模式比Alvarado計分法更具有預測效能。

關鍵字

急性闌尾炎 ； 資料探勘 ； 決策樹 ； 分類分析 ； Alvarado計分法

並列摘要

Acute appendicitis is the most common disease, which needs surgical intervention, in the world. Although the treatment of acute appendicitis is simple and straightforward, its diagnosis is still difficult. No single laboratory or image examination can make the diagnosis of acute appendicitis more accurate. Therefore, it is still a challenge to physicians to reduce appendiceal perforation and negative appendectomy rates.
The information technology may provide alternative approaches to acute appendicitis diagnosis. In this study, the decision tree classification approach, C4.5, is applied to build a prognostic model to provide an economical diagnostic assistance for acute appendicitis. The experimental results showed that the performance of C4.5 outperforms that of the well known Alvarado score system. The results of this study should be helpful to the development of the computer-aided system in the other medical field.

並列關鍵字

Acute Appendicitis ； Data Mining ； Decision Tree ； Classification Analysis ； Alvarado Score