Healthcare Data Dictionaries and Master Data Management

Healthcare data dictionaries and master data management are the practices that give an organisation a single, consistent definition of its data elements and reference entities. A data dictionary documents the meaning, format, and permissible values of each data item, while master data management maintains authoritative records for core entities such as patients, providers, locations, and coded concepts, often binding local items to standard terminologies like SNOMED CT and LOINC.

Definition

A healthcare data dictionary is a documented, authoritative catalogue of an organisation's data elements — their definitions, formats, and value sets — while master data management is the discipline of maintaining a single, consistent, authoritative record of core reference entities and codes used across systems, including the binding of local items to standard terminologies.

Scope

This entry covers data dictionaries, controlled terminologies and code systems (such as SNOMED CT, LOINC, and ICD), terminology binding and mapping, and the master data management of core healthcare entities. It treats these as the semantic foundation of interoperability and a methodological topic; it does not give data-governance product or implementation guidance.

Core questions

• What is a data dictionary and what does it specify for each data element?
• How do controlled terminologies such as SNOMED CT and LOINC provide shared meaning?
• What is terminology binding, and why is mapping between code systems hard?
• What entities does master data management govern in a health organisation?

Key concepts

• Data element, value set, and permissible values
• Controlled terminology and code systems
• SNOMED CT, LOINC, and ICD
• Terminology binding and value-set definition
• Concept mapping and crosswalks
• Master data and the single source of truth
• Patient and provider identity management

Mechanisms

A data dictionary defines, for every data element, its name, meaning, data type, format, and permissible values, so that systems and people interpret the element consistently. Semantic interoperability is achieved by binding those elements to controlled terminologies: SNOMED CT supplies clinical concepts, LOINC identifies laboratory and clinical observations, and ICD classifies diagnoses, each providing stable codes with defined meanings. Where systems use different code systems, concept mapping (crosswalks) relates one terminology to another, an inherently imperfect process because granularity and scope differ. Master data management then maintains authoritative, deduplicated records for core entities — patients, providers, locations, and reference codes — giving the organisation a single source of truth that downstream systems share.

Clinical relevance

Consistent data definitions and well-governed reference data are what allow coded results, problems, and identities to be reused safely across systems and for analytics. This entry describes how data dictionaries and master data underpin semantic interoperability; it is reference material and is not guidance for configuring terminology services or for clinical coding decisions.

Evidence & guidelines

Major terminologies are maintained by recognised bodies — SNOMED CT by SNOMED International, LOINC by the Regenstrief Institute, and ICD by the World Health Organization — and their use is referenced in interoperability standards and policy. Cornet and de Keizer (2008) review the evolution and use of SNOMED, McDonald et al. (2003) describe LOINC and its design, and Benson and Grieve's textbook situates terminology within the wider interoperability stack.

History

The need for controlled medical vocabularies was recognised early in medical informatics, with SNOMED's lineage stretching back to the 1960s and developing over decades into SNOMED CT, while LOINC was created in the 1990s to give laboratory and clinical observations universal identifiers. In parallel, the database and enterprise-architecture concepts of the data dictionary and master data management were adapted to healthcare's identity and reference-data problems, becoming central to semantic interoperability.

Debates

How reliable is mapping between different terminologies?

Crosswalks between code systems such as SNOMED CT, LOINC, and ICD are limited by differences in scope and granularity, so mappings can lose or distort meaning; how far automated mapping can be trusted remains a practical concern.

Key figures

• Christopher McDonald
• Ronald Cornet
• Nicolette de Keizer
• Stanley Huff
• Daniel Vreeman

Seminal works

• cornet-dekeizer-2008
• mcdonald-2003

Frequently asked questions

What is the difference between a data dictionary and a terminology?

A data dictionary documents an organisation's own data elements and their permitted values, while a terminology such as SNOMED CT or LOINC is an external, standardised set of coded concepts; a data dictionary typically binds its elements to such terminologies to achieve shared meaning.

What does master data management govern in healthcare?

It maintains authoritative, deduplicated records for core reference entities — such as patients, providers, locations, and code sets — so that systems across the organisation share a single, consistent source of truth for those entities.