利用深度學習分析檢傷、住院與暫留時間之重症病患偵測系統

Abstract

急診檢傷站是所有病患來到急診的第一站，此站設立的目的是希望透過對病患基本的生命徵象的量測與對人對自身病狀的描述，由護理人員評估狀況，並對病人進行分級，在給予最合適的醫療救護的同時，也能有效地運用有效的醫療資源。儘管目前已經有許多國家都有自己的一套急診檢傷系統，但是卻仍然面對很多限制。舉例來說， 在給予急診分級時，分級的結果會十分仰賴診療的護理人員的個人經驗與知識，在大多時候的狀況，常常會有錯誤分級的情況發生。這樣的情況不僅會對有緊急狀況的病人造成不好的影響，同時也可能增加整個醫療系統的負擔。這樣的問題在台灣也不例外，儘管台灣有TTAS檢傷分級系統，但是它並沒有足夠的能力對急診室的病人去進行分級，絕大多數的病患會被分至三級。

為了要從根本上解決這個問題，我們的目標是希望能夠透過 AI 去訓練一個機器學習模型能夠有效地分析病人的病狀並進行分流。即使已經有許多機器學習的研究致力於解決急診室分流不佳造成擁塞的問題，但是大部分的研究僅僅使用最傳統的演算法，且缺乏對文字資訊處理的能力，因此在此實驗中，我們嘗試搭建一個多模態的模型，能夠同時處理病人生命徵象與文字資訊進行分析預測，期待透過多方面的資訊整合獲得一個具有分辨重症病患能力的系統。

Triage is the process of accurately assessing the patient's symptoms and providing them with proper clinical treatments in the emergency department (ED). While many countries have developed their triage process to stratify patient’s clinical severity and thus distribute medical resources, there is still some limitation of the current triage process. Since triage level is mainly performed by the experienced nurse based on the mix of subjective and objective criteria, mis-triage often happens in the ED. It can not only cause adverse effects on patients but also impose an undue burden on the healthcare delivery system. In Taiwan, although the five-level triage system, Taiwan Triage and Acuity Scale (TTAS), is adopted in the ED to help nurses to distinguish patients’ emergent conditions, it cannot prioritize patients appropriately. Most of the patients are classified as level 3. In order to address the problem, our study aims to design a deep learning prediction system that can help the physician to prioritize patients to the appropriate triage level. To our knowledge, most of the existing models are developed by traditional machine learning methods and only utilize the patient’s demographics and vital signs information. Even though the chief complaint is included in the study, it is usually treated as a categorical feature instead of a text-based feature. As a result, to grasp rich information from multi-modal features, we proposed the multi-modality model that can handle heterogeneous data by using a representation learning approach.