by Giuliana Miglierini
The availability of interoperable data is a “must” to ensure the smooth sharing, use and re-use of data along the entire regulatory process. A new document – the European Medicines Regulatory Network Data Standardisation Strategy – has been issued by the European Medicines Agency (EMA) and the Heads of Medicines Agencies (HMA) to provide guidance to all authorities members of the European Medicines Regulatory Network (EMRN) on how to approach issues related to the definition, adoption and implementation of international data standards.
As data contains the higher informative value, regulatory procedures are undergoing a phase of deep rethinking in order to pay greater attention to the underlying data, leaving behind the traditional assessment of regulatory documents prepared using those data. Furthermore, the secondary use of data may support optimised regulatory decision-making processes.
The EMRN Data Standardisation Strategy – one of the main deliverables of the HMA-EMA Joint Big Data Steering Group (BDSG) workplan – builds upon the EU Data Strategy and the goals of the EMRN Strategy 2025. The FAIR principles (Findable, Accessible, Interoperable and Reusable) have been identified to govern the acquisition, conservation and use of data.
A more standardised approach to be adopted by the EMRN shall make possible to reduce the current manual and technical workload needed to manage regulatory documents and data, which are often submitted using a variety of different formats. According to the Strategy, a review of the mechanisms of data governance is also needed to include the responsibilities for oversight of data standardisation work and for the periodic updating of the EMRN data strategy. Further assessment of the expected costs and benefits is needed to make available the final roadmap for the implementation and prioritisation of measures suggested by the Data Standardisation Strategy.
Adopt, adapt, develop
The document was drafted on the basis of initial internal consultations on current practices for the management of scientific data; the inventory of the current IT systems in use by members of the EMRN was also created. In a second phase, a public survey was conducted among all stakeholders; current and new standards under development were identified by analysis of a selection of Standards Development Organisations (SDOs) active in the healthcare sector.
The exercise led to the identification of three different categories of actions, respectively referred to the adoption by the EMRN of currently available standards, the adaptation of available standards to include additional requirements, and the development of entirely new standards.
A total of eight principles have been identified to guide the standardisation of data to be used for regulatory purposes. First of all, high-quality data standards developed by accredited SDOs should always be adopted, on the basis of voluntary and consensus-based processes. EMA-HMA’s suggestion is to use such standards in place of (unique) local or proprietary standards, unless not compatible with the EU legislation or otherwise impractical. This would also support the creation of the European Health Data Space (part of the EU digital strategy) and facilitate the sharing and re-use of data and the governance of data privacy according to the draft EU Data Governance Act. The principles of data protection and data ethics by design are another fundamental requirement, as well as the collaboration with stakeholders in order to be informed about their needs. An alignment and consolidation of requirements from other regulations and guidelines is also envisaged; assessment of requirements for both human and veterinary regulatory practices should be run in order to avoid duplication of efforts or creation of competing standards.
Four domains for the management of data
The EMRN Data Standardisation Strategy identifies four different domains for the adoption of common standards: Medicinal Product, Healthcare & study data, Safety & risk management and Submissions. Recommendations for each of the four domains are discussed in the document.
The Medicinal Product domain includes product information, substance information, manufacturing and quality information.
Standards such as the ISO IDMP (IDentification of Medicinal Products) – based on the HL7 FHIR (Fast Healthcare Interoperability Resources) standard – are already in place, for example, to feed EMA’s SPOR system, but additional FHIR resources would be needed to structure key product information in already existing projects. To this regard, the Strategy suggests to implement and adapt ISO IDMP specified substance group levels 2 and 4 by the associated HL7 FHIR message exchange format, so to capture active substances and excipients in a structured format. This approach would also better support inspections (together with the availability of a standardised template for all inspections data), traceability of the origin of all substances used to manufacture a certain medicinal product, and the adoption of risk-based tracking procedures to assess the interactions between product quality and inspections of manufacturing sites.
The Healthcare and study data domain refers to data acquired during clinical development (i.e. data from interventional and/or observational studies) and the development of a common data model (CDM) for individual patient data and mobile health (mHealth) devices.
Many projects are already active, e.g. the ICH M11: Clinical electronic Structured Harmonised Protocol (CeSHarP), an initiative that would need further tuning to include some additional requirements. The possible use of raw data from interventional studies is also under assessment (e.g. CDISC SDTM – Study Data Tabulation Model, and ADaM-Analysis Data Model).
A new CDM standard would also be needed to exploit the full and harmonised potential of real-world data (RWD) and metadata obtained from healthcare records and disease registries. To this instance, a particular attention should be paid to mHealth applications installed on personal smart-devices, a fast-growing segment giving rise to the availability to big quantities of data.
The Safety and risk management domain refers to pharmacovigilance activities and includes Individual case safety reports (ICSR), Product safety update reports (PSUR), and environmental risk assessment (ERA) and risk management plans (RMP).
ICSR might be optimised by adoption of the HL7 FHIR based messaging standard and the inclusion of -omics data (genomics, proteomics and metabolomics) of patients. According to the Strategy, a new structured, electronic format for PSURs built upon the ICH E2C (R2) periodic benefit- risk evaluation report guidelines would be also beneficial for the work of regulatory authorities. The CDISC SDTM8 standard for data collection of risk assessment data represents a possible option to better handle data referred to environmental risk assessments. A structured, electronic risk management plan (eRMP) format should be also created to overcome the current paper template format; the new standard may find inspiration in the ICH E2E Pharmacovigilance Planning guideline and the EU Guideline on good pharmacovigilance practices (GVP) Module V – Risk management systems.
The Submissions domain includes the dossier management and the structured application form. Version 4.0 of the electronic Common Technical Document (eCTD v4.0) is more flexible compared to the previous one; for example, it supports the re-use of Pdf documents across sequences and eCTD dossiers and the submission of other data formats within the dossiers. The wider use of structured data will make the use of Pdf files less common, in favour of electronic Application Forms (eAF) based on data exchange standards such as HL7 FHIR to manage regulatory application along the entire life cycle of a medicine.