ICD-11

In one line

ICD-11 is the World Health Organization's eleventh International Classification of Diseases — the global standard for mortality and morbidity statistics, billing and reporting, redesigned from scratch as a digital-first product.

Classification vs terminology — the key distinction

A terminology like SNOMED CT records clinical meaning at the point of care, in all its richness. A classification like ICD does something different: it buckets every case into exactly one statistical category so they can be counted. You document with a terminology; you report and bill with a classification. This is why they coexist rather than compete — and why SNOMED→ICD maps matter so much.

What's new in ICD-11

ICD-11 isn't ICD-10 with more codes — it's a re-architecture:

• Foundation + linearizations. Underneath is a foundation — a semantic network of entities. From it, WHO derives linearizations (flat, mutually exclusive code lists) for specific uses: the mortality & morbidity statistics (MMS) tabulation, primary care, etc. One source of truth, many fit-for-purpose views.
• Post-coordination. A diagnosis can be a cluster: a stem code plus extension codes for laterality, severity, temporality, or the causing agent — so "fracture" becomes "fracture / left / open" without an explosion of pre-built codes.
• Digital-first. Released with a free API, a Coding Tool, and an embeddable browser — not a printed tabular list. You query it like a service.
• Broader scope. New and reorganised content: a dedicated chapter for mental health, an immune system chapter, conditions related to sexual health, gaming disorder, and traditional-medicine content. Get the two apart: Chapter 26 is TM Module 1, derived from Chinese, Japanese, Korean and Vietnamese medicine. Ayurveda, Siddha and Unani sit in TM Module 2, a separate chapter released later — that is the one shaped by Indian input.

Anatomy of a code

ICD-11 codes are alphanumeric (e.g. BA41 = acute myocardial infarction), a departure from ICD-10's letter-plus-digits. A fully specified diagnosis is often a cluster:

8B11 & <extension>   (stem & extension)
stroke, ischaemic    laterality, severity, temporality…

The & joins a stem code with extensions; the result is one precise, machine-readable diagnosis.

Where it shows up in digital health

• Vital statistics & public health — cause-of-death coding and WHO reporting run on ICD; this is its original job and remains central.
• Billing & claims — many payers and India's NHCX health-claims exchange ride on standardised diagnosis coding.
• Registries & research — cancer, injury and disease registries classify with ICD for comparability across countries and time.
• The mapping task — SNOMED→ICD maps let a clinician document once (rich terminology) while statisticians and billers get the right classification automatically.

ICD-10 → ICD-11: a long transition

WHO adopted ICD-11 for use from 2022, but countries migrate on their own timelines (India included), and the US uses its own clinical modification (ICD-10-CM) for billing. Expect both ICD-10 and ICD-11 in the wild for years — a working coder or informatician needs fluency in the transition, not just the new edition.

Common pitfalls

• Treating ICD as a clinical record — it's for counting; don't lose nuance by coding only to ICD at the point of care.
• Ignoring post-coordination — collapsing a cluster to its stem loses laterality/severity that downstream analytics need.
• Assuming a clean 1:1 map from ICD-10 — many concepts split, merge or move chapters.

Key takeaways

• ICD-11 = the global standard for counting disease (statistics, billing, reporting).
• Foundation + linearizations + post-coordination make it far more computable than ICD-10.
• It pairs with SNOMED (document) and LOINC (measure), via maps.
• The real-world skill is managing the ICD-10 ↔ ICD-11 overlap — practise it in the coding lab.