MIMIC III 数据库中表格信息

来源

数据库中表的基本信息

TableChildrenParentsColumnsRowsComments
admissions1811958,976Hospital admissions associated with an ICU stay.
callout22434,499Record of when patients were ready for discharge (called out), and the actual time of their discharge (or more generally, their outcome).
caregivers747,567List of caregivers associated with an ICU stay.
chartevents515330,712,483Events occuring on a patient chart.
chartevents_11538,033,561Partition of chartevents. Should not be directly queried.
chartevents_10159,584,888Partition of chartevents. Should not be directly queried.
chartevents_1115470,141Partition of chartevents. Should not be directly queried.
chartevents_1215265,413Partition of chartevents. Should not be directly queried.
chartevents_131539,066,570Partition of chartevents. Should not be directly queried.
chartevents_1415100,075,138Partition of chartevents. Should not be directly queried.
chartevents_21513,116,197Partition of chartevents. Should not be directly queried.
chartevents_31538,657,533Partition of chartevents. Should not be directly queried.
chartevents_4159,374,587Partition of chartevents. Should not be directly queried.
chartevents_51518,201,026Partition of chartevents. Should not be directly queried.
chartevents_61528,014,688Partition of chartevents. Should not be directly queried.
chartevents_715255,967Partition of chartevents. Should not be directly queried.
chartevents_81534,322,082Partition of chartevents. Should not be directly queried.
chartevents_9151,274,692Partition of chartevents. Should not be directly queried.
cptevents212573,146Events recorded in Current Procedural Terminology.
d_cpt9134High-level dictionary of the Current Procedural Terminology.
d_icd_diagnoses1414,710Dictionary of the International Classification of Diseases, 9th Revision (Diagnoses).
d_icd_procedures143,898Dictionary of the International Classification of Diseases, 9th Revision (Procedures).
d_items81012,487Dictionary of non-laboratory-related charted items.
d_labitems16753Dictionary of laboratory-related items.
datetimeevents5144,485,937Events relating to a datetime.
diagnoses_icd35651,047Diagnoses relating to a hospital admission coded using the ICD9 system.
drgcodes28125,557Hospital stays classified using the Diagnosis-Related Group system.
icustays821261,532List of ICU admissions.
inputevents_cv42217,527,935Events relating to fluid input for patients whose data was originally stored in the CareVue database.
inputevents_mv5313,618,991Events relating to fluid input for patients whose data was originally stored in the MetaVision database.
labevents3927,854,055Events relating to laboratory tests.
microbiologyevents516631,726Events relating to microbiology tests.
noteevents3112,083,180Notes associated with hospital stays.
outputevents5134,349,218Outputs recorded during the ICU stay.
patients19846,520
prescriptions3194,156,450Medicines prescribed.
procedureevents_mv525258,066Procedure start and stop times recorded for MetaVision patients.
procedures_icd35240,095Procedures relating to a hospital admission coded using the ICD9 system.
services2673,343Hospital services that patients were under during their hospital stay.
transfers313261,897Location of patients during their hospital stay.
40 Tables534728,556,685

表格中的列

在这里插入图片描述

关系图

在这里插入图片描述

其他(chartevents)

在这里插入图片描述

<think>好的,我现在需要处理用户关于MIMIC数据库的查询。用户想了解如何下载、使用教程以及数据结构。首先,我需要回忆之前提供的引用内容,特别是引用[1]和引用[2]。引用[1]提到了Windows下的安装步骤,而引用[2]详细介绍了用Python读取MIMIC-IV的方法,包括分块读取和使用dask、spark等工具。 接下来,我需要整理这些信息,分点回答下载、教程和数据结构的问题。下载部分需要提到PhysioNet的认证流程,可能用户不知道需要完成培训,这点很重要。然后安装步骤可能需要指导用户如何解压和配置数据库,引用[1]可以作为参考。 使用教程方面,引用[2]提到了Python处理的方法,特别是pandas和dask的使用,还有未来的教程计划,这部分需要强调。数据结构部分需要说明MIMIC-IV的模块,比如hosp、icu等,可能用户不清楚这些模块的具体内容,需要简要说明。 另外,用户可能需要知道如何开始使用,比如从哪些表入手,比如patients、admissions,这些是基础表。同时,处理大型数据集的技巧也很重要,比如分块读取,引用[2]中的方法可以引用。 最后,生成相关问题时要覆盖安装、数据访问、处理技巧和应用场景,这样用户可以有进一步的问题方向。需要确保所有LaTeX格式正确,比如变量用$...$,但可能这里数学公式不多,主要注意代码块的使用。回答要结构清晰,分步骤说明,保持自然的中文表达,同时正确添加引用标识。</think>### MIMIC数据库的下载、使用与数据结构详解 #### 一、MIMIC数据库下载 1. **访问权限申请** MIMIC数据库需通过PhysioNet平台申请,需完成CITI认证(伦理培训)并签署数据使用协议[^1]。 - 访问PhysioNet官网:https://physionet.org/ - 注册账号并提交“Credentialed Access”申请 2. **安装与配置** - Windows环境下可通过优快云教程《MIMIC-III 数据库安装》逐步操作,包括数据解压、PostgreSQL数据库配置等。 - 数据文件通常包含多个CSV表格,需按脚本导入数据库。 --- #### 二、使用教程与数据处理 1. **Python读取MIMIC-IV数据** ```python import pandas as pd # 分块读取超大CSV文件(适合内存不足场景) chunk_iter = pd.read_csv('mimiciv/icu/chartevents.csv', chunksize=10000) for chunk in chunk_iter: process(chunk) # 自定义处理函数 ``` - **高效工具推荐**: - `pandas`分块读取(`chunksize`参数) - `dask`分布式处理(适合TB级数据) - `PySpark`(集群环境适用)[^2] 2. **核心操作步骤** - **数据探索**:从`patients`、`admissions`表开始,关联临床事件表(如`labevents`)。 - **字段筛选**:使用`usecols`参数减少内存占用。 - **时间序列处理**:重点关注`charttime`字段的时序对齐。 --- #### 三、数据结构(以MIMIC-IV为例) MIMIC-IV分为多个模块,核心结构如下: | 模块 | 主要表格 | 关键字段示例 | |-----------|------------------------------|--------------------------------| | `hosp` | `patients`, `admissions` | `subject_id`, `gender`, `admittime` | | `icu` | `icustays`, `chartevents` | `stay_id`, `itemid`, `valuenum` | | `lab` | `labevents`, `d_labitems` | `specimen_id`, `loinc_code` | - **典型关联关系**: 患者(`patients`)→ 入院记录(`admissions`)→ ICU停留记录(`icustays`)→ 生理指标(`chartevents`)[^2] ---
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值