WHO’s ICD-11 maintenance platform frames a moving target for coders and contributors
WHO’s ICD-11 maintenance platform is not a static reference page. It is a working environment built for maintainers, contributors, and translators who need to navigate a classification that remains under active development. The site sits alongside the released browser, but its content is explicitly presented as work in progress, and that distinction shapes every function on the page. In practical terms, the platform is meant for people who shape the classification itself, not only for those who use it after publication.
The platform centers on the Foundation Component, described as the full WHO-FIC universe. That structure matters because it organizes diseases, disorders, injuries, external causes, signs and symptoms, functional descriptions, interventions, and extension codes within one multidimensional system. Entities can have multiple parents, which means a single clinical concept may sit in more than one branch of the taxonomy. Tuberculosis meningitis is the clearest example given: it belongs both to infection and to brain disease.
This design reflects a clinical reality that linear hierarchies often flatten. Many conditions do not behave like single-line textbook entries; they cross organ systems, mechanisms, and coding purposes. A maintenance platform has to preserve those overlaps if it is going to support accurate mortality and morbidity statistics.
Foundation Component hierarchy allows one condition to live in more than one place
The Foundation Component is more than a list of names. It is a network of interconnected entities and synonyms, with titles, inclusions, narrower terms, and related descriptors that can be surfaced through search. Quick search is built for speed and breadth: it searches titles, inclusions, synonyms, and narrower terms as the user types, then presents matches in a dynamic dropdown. The results are sorted by relevance and grouped within the ICD hierarchy so parent categories and children remain visually linked.
Wildcards are also available, which makes the search system useful when the exact wording is uncertain. A query such as “transp* accident” can pull up transport and accident-related entities. Advanced search goes further by letting users choose which properties to search, including title, synonym, description, inclusion, exclusion, additional information, fully specified name, and narrower term. If more than one keyword is entered, the system searches for items that contain all of them.
In a classification as large as ICD-11, search is not a convenience feature. It is the entry point for consistent coding. If a coder cannot find the correct entity quickly, the risk is not just delay; it is misclassification. For mortality and morbidity statistics, those errors propagate into surveillance, reimbursement, planning, and trend analysis.
The Coding Tool narrows the results to codable categories
WHO separates the browser experience from the Coding Tool, and the distinction is deliberate. The Coding Tool is designed to help users code with ICD-11, while the browser can show groupings or blocks as well as coded entities. The Coding Tool returns only items with codes. It also searches only the index, including titles, whereas the browser’s advanced search can look across additional properties.
The tool adds word completion, word suggestion, and chapter filtering. By default, it filters out Extension Codes and Traditional Medicine, which keeps the interface focused on mainstream statistical coding tasks. That filtering is a quiet but important feature, because it prevents noise from overwhelming the user during routine classification work.
A codebook that keeps expanding needs guardrails. The Coding Tool provides them by trimming the search field to codable categories and reducing the chance that users land on a non-codable grouping when they need a statistical output. For busy clinical coders, that separation can be the difference between a clean workflow and a frustrating search loop.
Visualization modes show how the linearization differs from the full foundation
The platform also explains how classification relationships are displayed. In the Foundation view, all nodes look the same except the selected entity, which appears with a darker background and larger circle. In Linearizations, only entities within that linearization are visible, and standard parent-child relationships appear in blue. Foundation children that are not linearization children are shown with dashed gray lines.
A third mode, Linearization within foundation, keeps the broader foundation visible while emphasizing the linearization itself. Entities included in the linearization appear as blue circles, while those outside it appear gray. Blue lines show linearization parents; gray lines show other parent-child relations in the foundation. The example on the page uses polycythaemia to show how one concept can have multiple children, some inside the linearization and some outside it.
This is not merely a visual aid. It reveals the architecture behind the code. A disease classification that supports morbidity and mortality statistics must handle both the full conceptual space and the narrower statistical slice used in release versions. The visual mode makes those layers legible to contributors who need to understand why a concept appears in one place in the foundation but not another in the release structure.
The platform also makes clear that the released ICD-11 browser is the version intended for routine use, while the maintenance site remains in motion. The content may change on an ongoing basis, and proposals are available only after login. That is the working reality of a large international classification: updates, proposals, and translations are handled in a controlled environment before they reach the public release.
A live classification system demands controlled access and ongoing revision
The maintenance platform exists for people who contribute to the classification by writing proposals or comments. Registration and login are required for those functions, and WHO points users who want the latest release to the separate ICD-11 browser. The site also links to the user guide, reference guide, training videos, and earlier releases with update history. Each of those resources signals the same operational principle: ICD-11 is maintained as a living standard, not a frozen document.
The caveat section is blunt about the intended audience. Maintainers, contributors, and translators are the primary users, and the content should not be confused with the released classification. For coders, that warning is critical. A work-in-progress classification can inform development and translation, but it is not the same as the version used for final statistical reporting.
There is a direct policy implication here. International disease classification only works when its users understand which version they are consulting and what role that version plays. The maintenance platform is built to support refinement before release, while the browser provides the stable reference for day-to-day use. In a system that underpins global mortality and morbidity data, version control is not a technical footnote. It is the foundation of comparability.
WHO’s platform therefore serves a narrow but essential function: it preserves the flexibility needed to maintain a modern, multidimensional classification while keeping the statistical core distinct from the working draft. The open question is not whether the classification will continue to evolve; it will. The real issue is how consistently contributors, translators, and coders use the platform’s layered tools so the release version remains precise, searchable, and usable when it reaches the field.