During routine hospital care, patients are billed by the hospital for diagnoses associated with their hospital stay. This table contains a record of all diagnoses a patient was billed for during their hospital stay using the ICD-9 and ICD-10 ontologies. Diagnoses are billed on hospital discharge, and are determined by trained persons who read signed clinical notes.
|Name||Postgres data type|
subject_id is a unique identifier which specifies an individual patient. Any rows associated with a single
subject_id pertain to the same individual.
hadm_id is an integer identifier which is unique for each patient hospitalization.
The priority assigned to the diagnoses. The priority can be interpreted as a ranking of which diagnoses are “important”, but many caveats to this broad statement exist. For example, patients who are diagnosed with sepsis must have sepsis as their 2nd billed condition. The 1st billed condition must be the infectious agent. There’s also less importance placed on ranking low priority diagnoses “correctly” (as there may be no correct ordering of the priority of the 5th - 10th diagnosis codes, for example).
icd_code is the International Coding Definitions (ICD) code.
There are two versions for this coding system: version 9 (ICD-9) and version 10 (ICD-10). These can be differentiated using the
In general, ICD-10 codes are more detailed, though code mappings (or “cross-walks”) exist which convert ICD-9 codes to ICD-10 codes.
Both ICD-9 and ICD-10 codes are often presented with a decimal. This decimal is not required for interpretation of an ICD code; i.e. the
icd_code of ‘0010’ is equivalent to ‘001.0’.
Was this page helpful?
Glad to hear it! Please tell us how we can improve.
Sorry to hear that. Please tell us how we can improve.