Basic Training IDMP & FHIR for NCAs staff

This training webinar, hosted by the UNICOM project, provides an in-depth, technical guide for National Competent Authority (NCA) staff on how to automatically extract structured medicinal product data from the new electronic application forms (eAFs) using ISO IDMP and FHIR standards. The central motivation is to minimize administrative effort and accelerate the transition to a data-driven medicines regulatory network, aligning with the EMA's 2025 network strategy. The session moves beyond theoretical standards, offering practical, operative instructions—including specific XPath queries—for IT and data experts to implement data extraction into their national systems.

The presentation establishes the context of the regulatory shift, moving away from proprietary PDF technology to a modern, state-of-the-art online tool (the PLM Portal, formerly "Daddy"). This new system ensures that data is received in a structured, IDMP-compatible way, leveraging EMA's SPOR services (Substance, Product, Organization, and Referential master data). The core technical challenge addressed is consuming the embedded FHIR (Fast Healthcare Interoperability Resources) message—an XML representation of the structured data—within the submitted eAF PDF. The speakers, drawing on implementation experience from the Austrian and Spanish agencies, emphasize that while the business process of submission remains unchanged, the underlying technology offers unprecedented opportunities for automation and data quality improvement.

A significant portion of the training is dedicated to demystifying the FHIR standard for regulatory data, specifically focusing on the structure of the message and how to query it using XPath. The speakers detail the "Top 10" essential medicinal product attributes that NCAs need for process control, including the PMS ID, MRP/DCP/CP number, ATC code, authorized dose form, product full name, authorization country, MA holder (Mah) details, package size, and active ingredient strength. They meticulously map these business requirements (e.g., "PMS ID") to their corresponding FHIR resources and elements (e.g., MedicinalProductDefinition > identifier with system pmsid). The methodology proposed is a "fire agnostic approach," treating the FHIR message as a standard XML document and utilizing XPath for targeted data extraction, a strategy proven effective in early NCA implementations.

The session also introduces the concept of "provenance" or "present and proposed" changes within the variation forms. Unlike old forms where changes were free text, the new system logs changes automatically, showing exactly what was selected (e.g., a specific ATC code from a list). This structured change tracking is logged in the XML, enabling systems to read and interpret the exact nature of the variation, which can then feed into workflow engines or case management tools for automated task routing and display. The speakers provide step-by-step XPath construction, demonstrating how to navigate the complex relationships between FHIR resources—such as linking the Task resource to the MedicinalProductDefinition, then to the RegulatedAuthorization, and finally to the Organization—to retrieve comprehensive product and authorization details. This practical guidance is crucial for firms like IntuitionLabs.ai that build custom data pipelines and integration layers for pharmaceutical clients.

Detailed Key Takeaways

• IDMP and FHIR Implementation is a Legal Obligation: The move to ISO IDMP standards is driven by the network strategy to 2025 and is a legal obligation stemming from pharmacovigilance regulations, necessitating immediate system readiness for structured data consumption.
• Transition from PDF to Structured Data: The new electronic application forms (eAFs) replace old PDF technology with a modern online tool (PLM Portal). While a human-readable PDF is still produced, it contains an embedded FHIR-based XML data package that NCAs must learn to consume automatically.
• FHIR Agnostic Extraction Strategy: The recommended approach for consuming the FHIR message is treating it as a standard XML document and using widely available XML processing technologies, specifically XPath, to query and extract concrete data points without relying on complex, fire-specific libraries.
• Mapping Business Needs to FHIR Resources: Key product attributes are found across four main FHIR resources: MedicinalProductDefinition (IDs, name, ATC), PackageProductDefinition (pack size), RegulatedAuthorization (MA number, country, holder), and Ingredient (substance name, strength).
• XPath Query Construction is Essential: Successful data extraction requires precise XPath queries that navigate the FHIR structure. For instance, retrieving the PMS ID involves targeting the identifier element filtered by the specific system URL associated with PMS IDs.
• Leveraging SPOR for Master Data: The new eAFs increase the usage of EMA's SPOR services. Applicants select reference products and master data (substances, organizations) from SPOR, ensuring high data quality and consistency in the submitted FHIR message.
• Handling Redundancy in Authorization Data: Authorization details (MAH, country, MA number) can be linked either to the product level or the package level, depending on the member state's regulatory approach (e.g., Italy/Portugal vs. Germany/Austria). Extraction logic must account for both paths.
• Structured Provenance Data: The new system automatically logs all changes made by the applicant (provenance). This structured change data—detailing the "present" and "proposed" values—is available in the XML, allowing for automated interpretation of variation scope and feeding into workflow automation.
• Importance of System Identifiers (Systems): Many crucial fields, such as PMS ID, ATC code, and OMS Location ID, are identified not just by their value but by a "system" attribute (a URL or code) that defines the classification or dictionary they belong to. These system identifiers must be hardcoded or managed as variables for accurate querying.
• Ingredient Strength Complexity: Extracting ingredient strength is complex as it involves multiple fields (concentration, presentation, high/low value ranges, and operators), requiring careful implementation to accurately represent quantitative data.
• Resource for Implementation: The speakers recommend referencing the ISO IDMP EU Implementation Guide (from EMA) for business context and the FHIR documentation (build.fire.org) for technical structure. They also plan to publish their XSLT code on GitHub for community use.

Tools/Resources Mentioned

• FHIR (Fast Healthcare Interoperability Resources): The standard used for inter-exchanging structured healthcare data.
• ISO IDMP (Identification of Medicinal Products): The standard being implemented for medicinal product data structure.
• XPath: The technology used to query and extract specific data elements from the XML/FHIR message.
• PLM Portal (Product Lifecycle Management Portal): The new online tool developed by EMA for creating electronic application forms (formerly known as "Daddy").
• EMA SPOR Services: Substance, Product, Organization, and Referential master data services used to ensure structured, high-quality data input.
• UNICOM Project: The pan-European project funding and driving the implementation of IDMP standards across member states.
• hl7.org/fhir: Official documentation for the FHIR standard.
• EMA ISO IDMP EU Implementation Guide: Essential business documentation for understanding the IDMP model as implemented in Europe.

Key Concepts

• Resource (FHIR): The basic container of information in a FHIR message, logically related to a business entity (e.g., MedicinalProductDefinition, Ingredient).
• Codable Concept: A fire type used to specify a code (value) with respect to a classification system (dictionary), often referencing an RMS list.
• Identifier: A fire type used to specify a unique value (like a PMS ID) against a defined system.
• Provenance: The structured tracking of changes within the application form, detailing the "present" and "proposed" state of a data element, enabling automated interpretation of variations.
• Bundle (FHIR): The root element of the FHIR message, containing all the individual resources (Task, Product, Authorization, etc.).
• NCA (National Competent Authority): Regulatory bodies within EU member states responsible for authorizing and monitoring medicinal products.