Automatic infection detection based on electronic medical records

• DOI: 10.1186/s12859-018-2101-x
• Publication Date: 2018-04-11T10:28:05Z
• Authors: Huaixiao Tou; Lu Yao; Zhongyu Wei; Xiahai Zhuang; Bo Zhang
• Keywords: Medical record; Electronic medical record; Health records

Making accurate patient care decision, as early possible, is a constant challenge, especially for physicians in the emergency department. The increasing volumes of electronic medical records (EMRs) open new horizons automatic diagnosis. In this paper, we propose to use machine learning approaches infection detection based on EMRs. Five categories information are utilized prediction, including personal information, admission note, vital signs, diagnose test results and image diagnose. Experimental newly constructed EMRs dataset from department show that models can achieve decent performance with area under receiver operator characteristic curve (AUC) 0.88. Out all five types note text form makes most contribution AUC 0.87. This study provides state-of-the-art processing system automatically make decisions. It extracts features associated achieves models.