International Classification of Diseases
icdDefinition
ISO-11179 Definition
The International Classification of Diseases (ICD) is the global standard diagnostic terminology and coding system published by the World Health Organization (WHO) and adapted for clinical use in the United States by the National Center for Health Statistics (NCHS) and CMS. ICD assigns alphanumeric codes to diseases, disorders, injuries, symptoms, and causes of death, providing a universal language for health data collection, reporting, and reimbursement. The current version in US clinical use is ICD-10, which was mandated for all HIPAA-covered electronic transactions beginning October 1, 2015.
ICD-10-CM (Clinical Modification) is used for diagnosis coding in all healthcare settings, while ICD-10-PCS (Procedure Coding System) is used for inpatient procedure coding in hospitals. ICD codes are the primary clinical classification system underlying nearly every analytical workflow in US healthcare. Claims adjudication, HCC risk adjustment, HEDIS quality measures, DRG assignment for inpatient payment, public health surveillance, and clinical registry reporting all depend on accurate ICD coding.
An ICD-10-CM code can be up to seven characters: the first three characters identify the category, characters four through six add clinical specificity, and a seventh character extension is used for certain conditions such as fractures and obstetric encounters. The full ICD-10-CM code set contains over 72,000 codes, representing a dramatic expansion in specificity from the roughly 14,000 codes in ICD-9-CM. Healthcare data engineers store ICD codes as VARCHAR(7) without the decimal point separator, matching the format used in CMS reference files and HIPAA 837 transactions.
A common ingestion problem is that some source systems include the decimal (e.g., E11.9 for type 2 diabetes without complications) while others strip it (E119) — all must be normalized to the no-decimal format before joining to reference tables. Engineers maintain versioned ICD-to-HCC crosswalk tables for risk adjustment, ICD-to-DRG grouper tables for inpatient payment analysis, and ICD-to-HEDIS measure crosswalks for quality reporting. Related standards include CPT for procedure coding, SNOMED CT for clinical terminology, and the annual ICD-10-CM updates published by CMS and NCHS effective each October 1.
Standard Abbreviation
icd
Category
Production DDL — FACT_CLINICAL_EVENT
CREATE OR REPLACE TABLE FACT_CLINICAL_EVENT (
clin_evt_key INTEGER NOT NULL -- surrogate key,
mbr_key INTEGER NOT NULL -- FK to DIM_MEMBER,
prvdr_key INTEGER NOT NULL -- FK to DIM_PROVIDER,
diag_cd VARCHAR(10) -- ICD-10 diagnosis code,
hcc_cd VARCHAR(10) -- HCC category code,
proc_cd VARCHAR(10) -- procedure code,
svc_dt DATE -- service date,
pos_cd VARCHAR(2) -- place of service,
raf_scr DECIMAL(10,3) -- RAF score,
admit_dt DATE -- admission date,
dsch_dt DATE -- discharge date,
drg_cd VARCHAR(10) -- DRG code,
los_days SMALLINT -- length of stay,
readmt_ind CHAR(1) -- readmission indicator,
load_dt TIMESTAMP_NTZ NOT NULL -- load timestamp
);
Standard Snowflake DDL for the canonical clinical table. Convert to BigQuery or Databricks →
Why This Term Matters
Clinical terms are the building blocks of risk adjustment, quality measurement, and value-based care analytics. A data engineer who understands this terminology can design schemas that correctly capture patient conditions, procedures, and encounters — enabling accurate HCC scoring, HEDIS measure attribution, and CMS reporting. Misclassifying clinical fields in a data warehouse cascades into incorrect RAF scores and failed regulatory submissions.
Related Content
Related Definitions
Looking for more healthcare terms?
Browse our complete library of 100,000+ standardized healthcare data terms
Browse All Terms