It has, since 2012, been a wellestablished fact that pharmaceutical companies that have medicinal products or are conducting clinical trials in the European Union (EU) will have to adhere to the ISO standards mandated in EU Regulation 520/2012. With the aim of improving overall pharmacovigilance signal detection and oversight, the European Medicines Agency (EMA) will rely on these standards (11615, 11616, 11238, 11239 & 11240) to achieve this goal.
Pharmaceutical companies that are faced with this regulatory requirement must in turn brace themselves for the implications (e.g. financially, organisationally, process, etc.) and the structural changes required to achieve compliance. A natural starting-point to further understand the impact of ISO IDMP is to analyse the data volume required for submission and, even more important, the uniqueness of the data required. But how do you evaluate this while taking into consideration the complexity of the ISO IDMP data models (ISO 11615 & ISO 11238)? Some publications have exemplified the potential data volume and complexity associated with IDMP submission to further support the establishment of a corporate business case. These publications have, however, neglected to take into consideration the data uniqueness across registrations which may significantly implicate the workload and solution scenario chosen to support IDMP compliance. To further qualify this notion, we have analysed the ISO 11615 data model (authorised products) with regard to data uniqueness and areas for potential overlap between registrations.Our initial approach involved the analysis of the ISO 11615 data model with emphasis on the data domains that potentially could be candidates for datasharing across registrations. The marketing authorisation and medicinal product domain both represent domains that we found exhibited the highest degree of data uniqueness across registrations. The data required to fulfil these data domains may be found at a national level (e.g. medicinal product name) and is furthermore tightly linked with the regulatory approval procedure (e.g. approval date, registration number, etc.) which largely is unique. The marketing authorisation class is also one of the predominant classes when it comes to transactional data (e.g. date attributes) which, due to the nature of these data, are considered highly unique. Of notable interest for this article were the areas where we potentially would find data overlaps across registrations. The manufacturer establishment (organisation), packaged medicinal product, pharmaceutical product, clinical particulars and substance IDMP domains were all speculated to represent areas where companies may benefit from data overlap (i.e. data that may be shared across IDMP submissions). Using two fictive products (A & B) and their respective regulatory data, we calculated the individual data volume to further represent the expected data overlap between IDMP submissions. Our calculation was based on a pre-analysis where we had identified areas where we would expect to find potential data overlaps due to the nature of the IDMP attribute. Please refer to Table 1 for a breakdown of the data used to prepare this article. Note that both products were intentionally very homologous in terms of the respective IDMP data. This was deliberate, to investigate the expected data overlap between two seemingly similar products (except for strength).
Our calculation showed that 2.595 data fields would be required for the submission of Product A according to the ISO 11615 information model. Due to the data overlap exhibited between the two products (1663 fields) we would expect to be able to re-use approximately 64% of the data generated as part of the initial registration of Product A. We have prepared an illustration to visualise the data uniqueness (and overlap) across the various IDMP domains (see Figure 1). In our example, both products were manufactured at the same manufacturers and similar operation types were stated in the regulatory dossier. We found that companies with a limited number of manufacturers may benefit from datasharing across registrations, since the same manufacturer may be used across multiple registrations. It is not uncommon that pharmaceutical companies aspire to have a limited number of manufacturers and our observations illustrate that this strategy may positively impact the initial IDMP submission and the subsequent maintenance of data. The same holds true for the packaged medicinal product area where a oneto-many relationship may exist between select packaging material and the respective registrations. This in turn offers the possibility to type data in once and associate all relevant registrations with the applicable packaging material.
Figure 1: Representation of data uniqueness and overlap across IDMP domains
Pharmaceutical companies with a very broad packaging material landscape will, however, be forced to type in multiple unique data attributes and associate these with the applicable registrations to support IDMP compliance. This further complicates the subsequent maintenance of data and an analysis of the packaging material landscape within the organisation may be beneficial to further qualify the complexity related to IDMP compliance in this area.In the clinical particulars domain, we would also expect that some degree of data overlap would be prevalent across registrations. It is our experience that this, however, is one of the more complex areas to definitively determine the data volume, since this greatly depends on the uniformity of the respective SmPCs and the alignment between these and the company core data sheet (CCDS). In principle, the greater the alignment between the CCDS and the respective SmPCs, the greater the data overlap. In our example, we assume a significant data overlap since we assume the products exhibit similar adverse event profile, contraindications, indications etc. It is, however, not unfamiliar that there is significant variance between SmPCs in the respective EU countries and caution should be exercised when attempting to calculate the true data volume. This national variance furthermore complicates coding of relevant sections in the SmPC against controlled reference terminologies (e.g. MedDRA) and pharmaceutical companies may benefit from starting an evaluation of the consistency across national SmPCs early to factor this into the project timeline.Within the substance domain, we again would expect that p h a r m a c e u t i c a l companies may benefit from the utilisation of data across registrations. It is expected that the implementation of ISO 11238 will significantly change the manner in which substances are registered and identified. The individual substance IDs (substance ID & specified substance ID I-III) which are a natural outcome of the implementation of ISO 11238 may, however, be used across products and would again offer the possibility for replication across registrations. G-SRS (GInAS) is proposed as a repository for this information and we would expect that all submission and maintenance activities would be performed in this system. In relation to ISO 11615, this standard would merely rely on the IDs generated in the G-SRS (GInAS) system and since some of these IDs are generated based on manufacturers, grade, physical state, etc., we can only speculate as to the overlap between registrations for this data. For small and mid-size pharmaceutical companies with a limited regulatory application landscape and potentially a homologous product and registration portfolio, the above observations are important to take into consideration since it may directly impact the business case associated with manual data entry of IDMP-relevant data elements. Furthermore, since it has been communicated by the European Medicines Agency that data from EVWeb will be migrated as part of the ISO IDMP implementation, these fields will serve as the base for IDMP and focus may be placed on the EVWeb vs. ISO IDMP gap.Conclusion:
Based on two fictive products, we have demonstrated the percentage of IDMPrelevant data overlap as well as the respective IDMP domains where data overlap may be expected. We have also included some silent observations which may be worthwhile considering when analysing the impact of ISO IDMP within your organisation. The findings are intended to start a discussion about the order of magnitude for IDMP projects and the associated factors that should be taken into consideration when scoping and estimating the internal IDMP efforts. The objective of the article is also to present observations from the analysis that may help pharmaceutical companies that are currently compiling business cases related to ISO IDMP compliance.