Information that is consistent for the lifetime of a patient is stored in this table.

Table columns

Name Postgres data type
subject_id INTEGER NOT NULL
gender VARCHAR(1) NOT NULL
anchor_age INTEGER NOT NULL
anchor_year INTEGER NOT NULL
anchor_year_group VARCHAR(255) NOT NULL
dod TIMESTAMP(0)

Detailed Description

subject_id

subject_id is a unique identifier which specifies an individual patient. Any rows associated with a single subject_id pertain to the same individual. As subject_id is the primary key for the table, it is unique for each row.

gender

gender is the genotypical sex of the patient.

anchor_age, anchor_year, anchor_year_group

These columns provide information regarding the actual patient year for the patient admission, and the patient’s age at that time.

  • anchor_year is a shifted year for the patient.
  • anchor_year_group is a range of years - the patient’s anchor_year occurred during this range.
  • anchor_age is the patient’s age in the anchor_year. If a patient’s anchor_age is over 89 in the anchor_year then their anchor_age is set to 91, regardless of how old they actually were.
  • Example: a patient has an anchor_year of 2153, anchor_year_group of 2008 - 2010, and an anchor_age of 60.
    • The year 2153 for the patient corresponds to 2008, 2009, or 2010.
    • The patient was 60 in the shifted year of 2153, i.e. they were 60 in 2008, 2009, or 2010.
    • A patient admission in 2154 will occur in 2009-2011, an admission in 2155 will occur in 2010-2012, and so on.

dod

The de-identified date of death for the patient. Date of death is extracted from two sources: the hospital information system and the Massachusetts State Registry of Vital Records and Statistics. Individual patient records from MIMIC were matched to the vital records using a custom algorithm based on identifiers including name, social security number, and date of birth.

As a result of the linkage, out of hospital mortality is available for MIMIC-IV patients up to one year post-hospital discharge. All patient deaths occurring more than one year after hospital discharge are censored. Survival studies should incorporate this into their design.