1. A Case for Linked Data for Medical
Devices in the IVD Market
bioMerieux INC.
Guillermo Perasso
UNC CH – CHIP – Practicum - Dec 2012

2. Content
Target Organization: bioMerieux Inc
Goals and Scope
Problem Description
Proposed Solution: Use Cases
Solution Implementation
Evolution and Future Developments.

3. Target Organization
 Target Company: bioMerieux SA
 French origin, multinational organization (~35 subs).
 In-Vitro Diagnostic (IVD) markets: Clinical and Industry.
 Key Business Application: Instrument Management.
 Manufacturing, Distribution, Technical Support.
 Several Product Lines and Market Segments.
 Different Instrument types and configurations.
 Core Information System: ERP SAP AG.
 No Semantic Web tools in IT production environments.

4. Company IS – data sources
Company’s intranet
ERP System: SAP
RDBMS: Oracle 9i
Structure /
Un-structure
information
Enterprise Content
Management System

5. Goals and Scope
 Build a business case for the implementation of Semantic Web-
based tools to support global business operations.
 Propose a simple implementation strategy for a target ontology
from SAP-based RDBMS and other sources.
 Enhance visibility of information about medical devices for stakeholders
(no SAP users).
 Improve Quality, minimize data inconsistencies and redundancies in
medical device records. Promote collaboration and data integration.

6. Problem Description
 Accessibility and Data Consistency: Multiple data sources
 ERP System databases
 Spreadsheets
 Intranet/extranet websites
 Data Visibility: SAP access required: proper authorization, access rights
 Complexity of instrument configuration management
 Number of materials
 Versions
 Types and Classifications
 Localizations
 External Coding

7. Proposed Solution: Semantic Web Tools
Publish a formal ontology for IVD devices:
 Data mapping with existing DB systems
 Advanced Queries and Search Tools
 Implement Standards:
 Languages (RDFS, SKOS, RDF, R2RML)
 Controlled Vocabularies (DC, SNOMED)
 Ontologies (FOAF, OBO, FDA)
 Data Enrichment: Semantic Annotations (unstructured data)
 Collaboration and Quality Control: Semantic Wiki, Portals.
 Advantages: Scalability, Flexibility, Interoperability (standards)

8. Proposed Solution: Semantic Web Tools
Publish a formal ontology for IVD devices:
Data mapping with existing DB systems
Advanced Queries and Search Tools
Implement Standards:
Data Enrichment: Semantic Annotations (unstructured data)
Collaboration and Quality Control: Semantic Wiki, Portals
R2RML

9. Ontology Management:
From Relational DB models to RDF
Domain
Ontology
Mapping
Ontology
Source
Ontology

10. W3C
R2RML: RDB to RDF Mapping Language
 Express customized mappings from
relational databases to RDF datasets.
 R2RML Mappings are triple maps in RDF
tailored to specific database schema and
target vocabulary.
 R2RML Processor takes a R2RML Mapping,
an Input Database and generates a Output
Dataset.

11. Source Ontology - SAP Relational
Mapping
• Defines the relational schema of the data source in a RDF form.
• RDF classes and properties to represent database schema metadata
o Tables
o Columns
Data Model for Medical Devices using SAP tables
o Primary Keys
o Foreign Keys, etc.

12. Mapping Ontology – W3C’s R2RML

13. Domain Ontology – RDF Triples
Define domain-specific terms and inter-relationships describing how underlying data to
be presented.

14. Semantic Data User Interface:
Semantic Wikis and Portal
BIOPEDIA

15. Semantic Technologies: Evolution
Deploy links with standard vocabularies
MeSH, Open Biological and Biomedical Ontologies, FDA Bioportal.
Linked Data Implementation for Instrument Installed Base:
Customer locations, Customer Service Actions, spare parts, consumables.
QA Systems: data quality, validation
Semantic Queries and Inference tools.

16. Semantic Web Solution.
Information Gain: samples
To replace an instrument with a new one, which has more capacity and require a higher level of
investment. Need to evaluate the market size for marketing purposes.
List instruments installed in hospitals and laboratories located in urban areas with more than x million
inhabitants in a certain country or region.
 Link the location of the installed base, cities and places with DBPEDIA entries to retrieve the
population in a given area. (http://dbpedia.org/About)
Require validation of instruments installed with standard technical specifications and equipments and
components not provided by the organization manufacturing facilities.
List installed instruments with accessories purchased to local vendors in subsidiaries. Report technical
specifications and vendor data.
 When possible link specifications about equipments with a formal ontology maintained by FDA.
Other links can be added for regulatory agencies in other markets (Europe, Asia, etc.)
http://bioportal.bioontology.org/ontologies/1576

17. A Case for Linked Data for Medical
Devices in the IVD Market
Semantic Technologies in the IVD Organization.
Summary of Pros and Cons:
 Flexibility – Scalabity: Simplify Implementation plans.
 Standard Ontologies, Languages, Vocabularies: Semantic
Interoperability.
 Data Protection, Data Quality.
 Data Management: Increase Overhead.