1 of 16

�Prototyping a Health DCAT-AP data catalogue to support population health indicator identification and quality assessment�Rob Brennan, Junli Liang and Akila Wickramasekara, ADAPT, School of Computer Science, UCD, Ireland

NeXt-generation Data Governance workshop 2025 (NXDG 2025)

rob.brennan@ucd.ie

2 of 16

Context: Data-driven Public Health for Ireland and the EU

  • HSE Health App now available
  • MyHealth@EU cross-border health services started
  • European Health Data Spaces(EHDS) Regulation
    • Published March 2025
    • Implementing acts March 2027
    • Primary use (data exchange) & Secondary use March 2029

Primary Use = use of data for the delivery of healthcare

Secondary Use = use of data for research and public interest purposes

See

https://health.ec.europa.eu/ehealth-digital-health-and-care/european-health-data-space-regulation-ehds_en

https://dssc.eu/

3 of 16

Our Use Case: �Older Adults Health and Wellbeing Profile (OAHWP)

​

​

​

​

​

4 of 16

=> Need a set of easy-to-interpret indicators (metrics) validated by clinical experts

​

5 of 16

Impact and Value of the Profile

  • Data-driven planning and resource allocation
  • Enable advocacy
  • Standardise information for other health areas
  • Safety, quality improvement and research
  • Problems
    • No standardised HSE data or metadata eco-system
    • What data will we use?
    • Can we use the data?
    • What data is most relevant?
    • Sustainability?

Example Population Health Indicator

Name:

Consistent Poverty Rate for those over 65+ (%)

Definition:

The percentage of adults aged 65 years and over who live below the poverty line and who lack basic necessities. Consistent poverty is a broader measure that considers both income poverty and the experience of deprivation (inability to afford basic necessities).

Numerator:

Number of adults (65+) who are in consistent poverty in a geographical area.

Numerator data source:

Central Statistics Office Survey on Income �and Living Conditions Table SIA61, 2024

Denominator

Number of people aged 65 and over in the geographical area.

Etc.

6 of 16

Partners

  • Clinical, Public Health, Academic Partnership
    • Clinical
      • National Clinical Programme for Older People
    • Public Health
      • Dept. of Public Health HSE Dublin South East
      • National Health Service Improvement
      • National Health Intelligence Unit
    • Academic
      • UCD School of Computer Science/ADAPT Centre (Data Workstream)
      • UCD School of Medicine (Evidence Workstream)
      • NCAD School of Design (Design Workstream)
  • Duration: Jan-Dec 2025
  • Aim: Co-design profile in partnership with older adults and key stakeholders
  • HSE Spark Innovation Programme funding

7 of 16

Profile Data Sources and Prioritisation Process

HSE PBP Core Indicator List – HIPE, (list actual datasets)

Tilda

Extracted Indicator List

CSO

Extracted indicator list

Additional Sources

Other Data Publishers

International Indicator Lists

Initial Criteria:

Age 65+;

Relevant to OA HWB Profile

Short List

(138 indicators)

Long

Indicator List

(300-400 indicators)

Final Indicator List

( Y indicators)

Final Candiate List

( X indicators)

Maximal

Candidate

Indicator

List

1200+ indicators

Screening by Project Team (Public Health)

Screening by Expert Panel

Delphi Assessment by Stakeholders

Final Screening by Expert Panel

At each assessment stage we need to know:� 1. Clinical relevance

2. Data feasibility

McGlacken, Therese, et al., “A Co-designed, Stakeholder-Led Prioritisation Framework for Developing an Older Adult Health and Wellbeing Profile in Ireland”, 18th European Public Health Conference, Helsinki 2025

8 of 16

Profile Data Sources and Prioritisation Process

HSE PBP Core Indicator List – HIPE, (list actual datasets)

Tilda

Extracted Indicator List

CSO

Extracted indicator list

Additional Sources

Other Data Publishers

International Indicator Lists

Initial Criteria:

Age 65+;

Relevant to OA HWB Profile

Short List

(138 indicators)

Quality & Feasibility Assessment

Long

Indicator List

(300-400 indicators)

Final Indicator List

( Y indicators)

Final Candiate List

( X indicators)

Maximal

Candidate

Indicator

List

1200+ indicators

Screening by Project Team (Public Health)

Screening by Expert Panel

Delphi Assessment by Stakeholders

Final Screening by Expert Panel

Quality & Feasibility Assessment

Quality & Feasibility Assessment

Quality & Feasibility Assessment

9 of 16

Our Solution: Deploy a Data Catalogue

  • Develop comprehensive metadata covering data sources as a reusable resource
  • Supports query-based solutions for many data feasibility issues
  • Lightweight meta-infrastructure
  • Problems Addressed (bold)
    • No standardised HSE data eco-system
    • What data will we use?
    • Can we use the data?
    • Is the data compatible with our desired indiocators? E.g. spatial and temporal granularity
    • Is it evidence-based?
    • Sustainability?

Data Feasibility

There are 2 key questions for data feasibility:

  1. Should I use this data? (Data reliability, relevancy, quality, …)
  2. Do we have access to the data?�(Data protection, costs, timelines, …)

- Gatto, N. M., Campbell, U. B., Rubinstein, E., Jaksa, A., Mattox, P., Mo, J., & Reynolds, R. F. (2022). The structured process to identify fit‐for‐purpose data: a data feasibility assessment framework. Clinical Pharmacology & Therapeutics, 111(1), 122-134.

10 of 16

OAHWP Data Catalogue Goals

  • Document existing national and international indicators and datasets
  • Enable assessment for quality and feasibility
  • Provide automated evidence and for OAHWP prioritisation process
  • Model and document data governance best practice
  • Validate European Health Data Spaces (EHDS) dataset metadata specification using OAHWP as a scenario

​

DCAT is a W3C standard for a data catalogue that defines metadata for

Datasets

Data services

Data catalogue

RDF-based metadata specification

Version 3 (2024)

Better

Discoverability/findability of data

Interoperability

Ease of analysis of data

DCAT–AP (Application profile for data portals in Europe) is used to describe public sector datasets within the European Union

Health DCAT-AP is a candidate extension of DCAT for EHDS https://tehdas.eu/

​

11 of 16

Metadata Ingestion, Validation, Assessment and Reporting Pipeline

Spreadsheet of metadata instances

Validation and Curation

Standards-based metadata

RDF

Data Catalogue Specification

Spreadsheet

Quality Assessment

Via SPARQL

R2RL

Mappings

Dataset quality and feasibility analysis report

Data processing pipeline

Data quality issue feedback

Domain

Experts

Health Profile

Prioritisation

Stakeholder

CSV

Quality

Semi-Automated

Web Scraping

Domain Experts,

Knowledge Engineers

Domain Experts,

Knowledge Engineers

12 of 16

Data Catalogue Structure

Indicator

Metadata Record

Denominator Data Source

Metadata Record

Numerator Data Source

Metadata Record

Indicators

Datasets

(Health DCAT-AP)

See RDF and doc https://github.com/hse-ucd/

13 of 16

Example Competency Questions for Data Catalogue

1. Completeness:

      • Are all mandatory and recommended metadata fields present?
      • Are there datasets specified for the numerator(s) and denominator(s)?
      • For each indicator with numerator and denominator data sources, do the numerator and denominator datasets have data for i) all of Ireland and ii) for the target Integrated Healthcare Area ?
      • For each indicator with data sources, do the numerator and denominator datasets have data for people aged 65 and over?

​

Similar questions for: Timeliness, Data Protection, Granularity/Precision

​

Data Catalogue Stakeholders

1. Members of the OAHWP Data Working Group: primarily interested in progress to date (completeness, quality) with creating the data catalogue

2. Participants in the OAHWP Indicator Prioritisation Process: are primarily interested in the data feasibility of using specific indicators in the OAHWP. 

​

14 of 16

Challenges Found

  • Health DCAT AP
    • Complexity: Health DCAT AP builds on many other specification
      • Developed SKOS Controlled vocabularies for many fields
    • Data Protection: additional fields needed to answer feasibility questions e.g. dpv:DataController
    • Spatial Resolution in m not appropriate for statistical data. Added SpatialAggregation property
    • Geonames regions for spatial coverage is too inflexible e.g. Irish health regions.
    • Mandatory use of WikiData for Categorical values is not appropriate
  • Data Catalogue Approach
    • Indicators needed their own concept definition
    • Tool support key for less familiar staff
    • System extremely dynamic e.g. Delphi process injected new indicators rather than just removing => automation key

​

15 of 16

Conclusions

  • EHDS will bring in a new era of unified health data and data governance for EU
  • Semantic Web metadata is a key enabler
  • Data Catalogue for EHDS and DGA currently under standardisation in CEN-CENELEC JTC 25 and ETSI Data
  • We need semantics people in there!

16 of 16

Thanks to the HSE OAHWP Data Workstream Team

  • Therese McGlacken, Stephen Barrett, Jacinta Mulroe, Teresa Bennett, Declan McKeown, Gerardine Sayers, Mary Browne, Aparna Keegan, Graham Hughes
  • And all our other collaborators in the OAHWP project

​