IDMP: Do You have Time to Implement MDM?
It is not quite a year since I have been looking into the Idenfication of Medicinal Products (IDMP) ISO standard, and the challenging EMA IDMP implementation deadline of July 1st, 2016. Together with HighPoint Solutions, we have proposed that using a Master Data Management (MDM) system as a product data integration layer is the best way to ensure the product data required by IDMP is consolidated and delivered to the regulator. This message has been well received with almost all the pharmaceutical companies we talked to.
During the past few months, the support for using MDM as a key part of the IDMP solution stack is growing. Supporters of using MDM now include people that have been looking into the IDMP challenge and solutions for far longer than me: independent consultants, representatives of pharma companies with active projects and leading analysts have expressed their support. At the IDMP Compliance Challenge and Regulatory Information Management conference held in Berlin last month, using MDM within the solution stack was a common theme – with a large percentage of presentations referencing the technology.
However, at this conference an objection to the use of MDM was circulating during the coffee break. Namely:
Do we really have time to implement MDM, and achieve compliance before July 2016?
First, let’s revisit why MDM is a favoured approach for IDMP. The data required for compliance is typically dispersed across 10+ internal and external data silos, building a data integration layer to prepare data for submission is the current favoured approach. It has popular support from both industry insiders and independent consultants. The data integration layer is seen as a good technical approach to overcome a number of challenges which IDMP is posing in their initial draft guidance, namely:
- Organisational: It has to make it easy for data owners to contribute to the quality of data
- Technical: It needs to integrate data from multiple systems, cleaning and resolving attributes using as much automation as possible
- Co-ordination: The layer must ensure data submitted is consistent across regulations, and also within internal transactional systems
- Timing: Projects must begin now, and pose low technical risk in order to meet the deadline.
MDM technology is an excellent fit to address these challenges (for a high level summary, see here).
So back to the time objection, it seems a bit out of place if you follow the logic:
- In order to comply with IDMP, you need to collect, cleanse, resolve and relate a diverse set of product data
- A data integration layer is the best technical architecture
- MDM is a proven fit for the data integration layer
So why would you not have time to implement MDM, if this is the best (and available) technical solution for the data collection and consolidation necessary to achieve compliance?
My feeling is the objection comes down to the definition of MDM, which is (correctly) seen as something more than technology. Expert and analyst definitions variously include the words ‘discipline’, ‘method’, ‘process’ or ‘practice’. Current consensus is that the underlying technology merely enables the processes which allow an organisation to collect and curate a single, trusted source of master data. In this light, MDM implies the need for senior level sponsorship and organisational change.
The truth is IDMP compliance needs senior level sponsorship and organisational change. In all the discussions I have had, these points come out clearly. Many pharma insiders who understand the challenge are grappling with how to get the required attention from execs that IDMP needs in order to achieve compliance. July 1 2016 is not only a deadline, it is the start of a new discipline in managing a broad range of pharma product data. This new discipline will require organisational change in order to ensure the high quality data can be produced on a continuous basis.
So the definitions of MDM actually make the case for using MDM as part of the technology stack stronger. MDM will not only provide technology for the data integration layer, but also a support structure for the new product data processes that will be required for sustained compliance. Without new product data management processes as part of the IDMP submission process, there will be few guarantees around data quality or lineage.
I fear that many of those with the objection that they don’t have time for MDM are really saying they don’t have enough time to implement IDMP as a new discipline, process or practice. This will expose them to the risk of non-compliance fines of up to 5% of revenue, and recurring fines of 2.5% of revenue.
In my mind, the challenge is not ‘do we have time to implement MDM?’, but rather ‘can we be successful both by and beyond July 2016, without implementing MDM?’ By MDM I am referring to the technology and the organisational aspects of creating and delivering complete and accurate data.