基于重症监护数据库MIMIC-Ⅱ的临床数据挖掘研究

Clinical Data Mining by Exploring Public MIMIC-Ⅱ Intensive Care Database

该文介绍了麻省理工计算生理学实验室、哈佛医学院BID医学中心以及飞利浦医疗合作建立的、开源的、多参数智能重症监护数据库Multiparameter Intelligent Monitoring in Intensive Care II:MIMIC-II,简单介绍了其结构、功能和临床应用。基于MIMIC-II临床数据库,开展了ICU脓毒血症病人血压和心率的昼夜节律性研究。该研究发现了脓毒血症患者死亡组和存活组在生理参数昼夜节律性上的显著性差异,解决了使用关系型数据库MIMIC-II开展生理参数昼夜节律性研究的很多技术问题。

This paper introduces a free and publicly open ICU database： multi-parameter intelligent monitoring in intensive care II： MIMIC-II, which has been built up and maintained by the laboratory of computational physiology at the Massachusetts Institute Technology, Beth Israel Deaconess Medical Center and Philips Healthcare over the past decade. This paper briefly introduces its infrastructure, implementation and applications in clinical studies. Clinical study pertaining to circadian variation in heart rate and blood pressure during sepsis is shown as a typical example of research performed with MIMIC-II. In this study, it was found there was significant difference in circadian variation in both heart rate and blood pressure between survival and non-survival groups in septic patients. This study tackled several important techniques necessary for the investigation of the circadian rhythm.