為分析急診轉入加護病房非癌症病人之死亡風險，盡早提供醫護團隊早期辨識末期病人。採電子病歷回溯性調查設計，從2013 年8 月1 日至2018 年04 月30 日期間從急診入住內科加護病房之病人共1665 人為研究對象。並以自擬結構性調查表進行資料收集及統計。在急診CPR 其病患死亡風險相對較高，風險比為0.062；在急診即簽署DNR 者其病患死亡風險相對較高，風險比為0.265；若有使用呼吸器風險比為0.645、若符合Salpeter (2012) 末期定義風險比為0.708、而APACHE II 每增加一分，其死亡風險為1.080 倍；GCS 每減少一分，其死亡風險增加0.907 倍。急診病人有CPR、有簽署DNR、使用呼吸器、符合Salpeter(2012) 末期定義及APACHE II score、GCS 昏迷指數等變項可以預測非癌病人於加護病房之死亡率，當急診重症病人須入住加護病房時常常病情惡化快速，若能早期辨識末期病人，進而輔助臨床醫師實施有效之醫病共享決策。本文研究結果提供急診及加護病房醫護團隊決策上之參考。

The purpose of this study is to analyze the mortality risk of non-cancer patients transferred from the Emergency Room (ER) to the Intensive Care Unit to enable medical care teams to recognize end-stage patients and provide the required treatment promptly. The study design uses retrospective investigations of electronic medical records collected from August 1st, 2013 to April 30th, 2018. During that period, 1,665 patients were admitted to the ER and transferred to the Internal Medicine ICU. All data was collected and analyzed via self-drafted structured questionnaires. The mortality risk was relatively higher when patients received CPR in the ER, with a hazard ratio of 0.062. If patients signed a DNR form in the ER, the hazard ratio was at 0.265. This was followed by patients who were placed on a ventilator, with a hazard ratio of 0.645. If patients were considered at the terminal stage, as defined by Salpeter (2012), the hazard ratio was 0.708. When the APACHE II score increased one point, the mortality risk increased 1.080 times over. By decreasing the GCS score by one point, the mortality risk increased 0.907 times over. Taking into consideration the factors that play a role in predicting the mortality rate of non-cancer patients in the ICU, including patients receiving CPR, signing a DNR form, using ventilators in the ER, being identified as at the terminal stage as defined by Salpeter (2012), the APACHE II score, GCS score, and when patients have a critical need to be transferred from the ER to the ICU, their condition usually deteriorates quickly. If physicians could identify those patients at the terminal stage promptly, this would assist clinical doctors in facilitating their decision-making process and in providing suitable medical treatment. The results of this research provide a reference for physicians in the ER and ICU when making medical decisions.