Skip to content

Domain B: Data Semantics & Quality

Focus: Standardisation, documentation, and quality assurance

Indicators: 6 (4 Core, 2 Enhancement)

The business question

Is the data messy or ready-to-use? High maturity slashes the time spent cleaning data.

Researchers consistently report that data preparation consumes 60–80% of project time. Organisations that invest in common data models, terminology mapping, and curated research-ready datasets dramatically reduce time-to-insight and improve reproducibility. This domain measures whether data is merely available or genuinely usable at scale.


Indicator Summary

ID Indicator Type Class Unit
B.1.1 Common Data Model Adoption Core C1 Both
B.1.2 Terminology Standards Core C1 Both
B.2.1 Quality Framework & Monitoring Core B0 Both
B.2.2 Data Documentation & Metadata Core B0 Both
B.3.1 Curated Dataset Availability Enhancement O Both
B.3.2 Phenotype Library & Validation Enhancement O Both

B.1 — Analytical Interoperability

B.1.1 Common Data Model Adoption

CORE · C1 · Both

Level Description
L1 No CDM. Data in source formats with bespoke schemas.
L2 CDM (OMOP, Sentinel, PCORnet, or equivalent) evaluated. Mapping assessed. Pilot planned.
L3 Core datasets partially mapped. Coverage <50%. CDM available but not routine.
L4 Core datasets mapped to recognised CDM with >= 60% coverage. CDM maintained and refreshed. Standard tools operational.
L5 Comprehensive CDM. Externally validated. Contributing to international networks. CDM-native services.

B.1.2 Terminology Standards

CORE · C1 · Both

Level Description
L1 No consistent standards. Codes as recorded (mixture of Read, ICD-10, OPCS, local).
L2 Strategy defined. Target standards identified. Baseline assessed.
L3 Primary standards adopted (SNOMED CT, dm+d for new systems). Legacy retains original. Partial mapping.
L4 SNOMED CT, dm+d, ICD-10/OPCS consistently applied. Terminology services operational. Limitations documented.
L5 Full SNOMED CT with semantic interoperability. Advanced services. Contributing to standards.

B.2 — Data Quality

B.2.1 Quality Framework & Monitoring

CORE · B0 · Both

Level Description
L1 No framework. Issues ad-hoc. No monitoring.
L2 Framework defined. Dimensions identified. Baseline initiated.
L3 Metrics for core datasets. Annual reporting. Issues documented. Improvement underway.
L4 Comprehensive monitoring with automated checks. Metrics published. SLAs defined. Root cause analysis.
L5 Real-time monitoring. Benchmarked nationally/internationally. Certification or audit.

B.2.2 Data Documentation & Metadata

CORE · B0 · Both

Level Description
L1 Minimal documentation. Tacit knowledge. No catalogue.
L2 Initiative underway. Basic documentation. Catalogue developing.
L3 Structured metadata for core datasets. Variable quality.
L4 Comprehensive, standardised metadata. Machine-readable. Searchable catalogue integrated with access. Regular updates.
L5 Rich ecosystem with automated generation. Provenance and lineage tracking. Community contributions.

B.3 — Research-Ready Data Products

B.3.1 Curated Dataset Availability

ENHANCEMENT · O · Both

Level Description
L1 No curated datasets. Raw extracts requiring extensive cleaning.
L2 Needs assessed. Priority datasets identified. Pilot underway.
L3 Selected curated datasets. Methodology documented but not standardised.
L4 Portfolio with standard methodology. Derived variables, phenotypes, linked data. Version control.
L5 Comprehensive library. Community contributions. Automated pipelines. Benchmarked.

B.3.2 Phenotype Library & Validation

ENHANCEMENT · O · Both

Level Description
L1 No library. Researchers define from scratch.
L2 Concept established. Initial phenotypes documented. No validation.
L3 Growing library. Selected phenotypes validated. Searchable but not integrated.
L4 Comprehensive with validated definitions. Standardised validation. Integrated with access. Version control.
L5 Internationally validated. Cross-references HDR UK/OHDSI. Phenotype-as-code.