A patient in the MIMIC II database is uniquely identified by an integer number called Subject_ID, it can be thought of as a medical record number (MRN) normally found in hospital information systems. The basic information for any given patient is stored in the table D_PATIENTS, normally referred as the patient table throughout this document. As the database went through a careful de-identification process, the patient table only stores the patient identifier (Subject_ID), gender (sex) and date of birth (dob, shifted). Table 2.1 shows a sample content of the patient table resulted from the query in listing 2.1.
| SUBJECT_ID | SEX | DOB | DOD | HOSPITAL_EXPIRE_FLG |
|---|---|---|---|---|
| 7049 | M | 04/10/1952 | 03/18/2020 | N |
| 7060 | F | 08/01/1932 | (null) | N |
| 7072 | M | 02/22/1928 | 03/20/1999 | Y |
| 7078 | F | 11/11/1967 | 10/17/2012 | Y |
| 9181 | F | 03/11/1960 | 02/16/2007 | Y |
| 9185 | M | 02/28/1927 | (null) | N |
| 9195 | F | 12/19/1974 | (null) | N |
The date of death for patients who died in the hospital is taken to be the date of discharge. For other patients, date of death was obtained from social security death records from the US government. Where there is a conflict between the hospital data and social security, the hospital data is assumed to be more accurate. Note that patients who left the US (and subsequently died) after receiving treatment may not have their date of death recorded in the social security archives.
As shown in Figure 2.1, the patient identifier (Subject_ID) is widely used by most of the tables throughout the database to specify to which patient a given measurement or recording refers to. Figure 2.2 shows an example relating which diagnosis codes (ICD-9) were assigned to a given patient, the Subject_ID field links the ICD-9 and the patient tables. The ICD-9 table records the ICD-9 codes applied to a particular patient during a specific hospitalization period. Diagnosis-related groups (DRGs) are stored in the drgevents table and their meanings are stored in d_codeditems.
