Part I: Four Technology Approaches for IDMP Data Management
Those people who have been reading my blog over the year must be aware that I believe a Master Data Management (MDM) based approach is the best fit for most IDMP efforts. The core features of a MDM system are an excellent match for the basic requirements of consolidating the broad scope of data required by the European regulators for IDMP.
Despite my support for a MDM based approach, I do realize IDMP compliance can be achieved through alternate technology stacks. The reasons for considering an alternate to an MDM based approach for IDMP are not based on functional capability – there is acceptance that the technology is a good fit. Rather objections are due to a small number of organizational factors:
- Too few products;
- Lack of data maturity or;
- Limited senior level sponsorship that is an important factor in the success of an MDM project.
The data maturity and sponsorship objections have been lessening with a phased approach to IDMP in Europe, as IDMP has gained greater visibility outside of Regulatory Affairs, and as there has been more time for data governance to grow during the IDMP project.
Within the next few months technology choices will be made to ensure pharmaceutical companies meet the new IDMP deadlines in Europe. So – back to question, what technology stack should you be considering?
In order to facilitate the discussion, I’ll make some basic assumptions by drawing some boxes and arrows for the major components, and giving a high level description as below:
- Data required for the full scope of IDMP compliance will be held in a variety of source systems and locations: structured, unstructured, on-premise or with partners.
- You will need to source and integrate this data from multiple systems into a single place in order to consolidate and prepare data for submission.
- The data should be of high quality; meaning it should be complete, consistent and accurate. Data quality activities will usually take place in the data consolidation layer.
- Actual submissions need to be managed on an ongoing basis, ensuring the regulators get the complete set of data, and you can capture their responses.
Technology choices align to the blue arrows and yellow boxes. Most of the discussions I have been a party to are around the yellow boxes. Broadly speaking there are four options that I see being considered for these two boxes: Manual Activities, Data Warehousing, MDM and RIMS.
All of the options require data integration – represented by the blue arrows in Figure 1. With most of the top pharmaceutical companies using PowerCenter as their corporate standard for integration, there is a strong possibility that this will be used in your IDMP project as the default integration technology. Due to the potentially 10 to 15 source systems per product for the full IDMP scope, PowerCenter’s ability to re-use mappings, and embed data quality routines will bring welcome efficiency to the data integration required.
In the first part of this blog, I will cover the Manual and Data Warehousing approaches to IDMP. These are typically being embraced by companies who aren’t also emphasising the data governance aspect of IDMP compliance.
Technology Option 1: A Manual Approach
Why choose this technology: Only advisable if you have a few products. Recently, at an IDMP breakfast in Copenhagen, 8 products were suggested as a cut-off point for a manual approach. Whilst I did not get to see the full calculations on this, the number of attributes to be maintained above 8 products did make a manual approach seem inefficient.
Challenges: Manual activities unfortunately bring along human error. Lack of formal process control can also lead to inefficiencies in terms of duplicate efforts and multiple versions that need to be synchronised . if the manual approach is right for you, you must consider embedding data quality capabilities into your processes to offset human error and lessen rework. Informatica has developed a Data Quality plug-in for Excel which will to improve process efficiency, apply controlled vocabularies and proactively reduce errors.
Technology Option 2: Data Warehousing Approach
Why choose this technology: My belief is that this approach will be adopted by companies with too many products for a manual approach but do not have a strong data governance program in place, and/or do not have an IDMP program with the influence to drive MDM adoption and improve data governance. Data Quality technology is crucial with this approach, as it will improve data consolidation efficiency through discipline (data quality rules) and proactive data quality management in what is typically a passive system (the data warehouse). Submission will be either through a RIMS system, or data exchange technology such as B2B Exchange, which is specifically designed to manage the interchange of data between two organisations.
Challenges: A lot of coding will be required to implement the required additional functionality for data consolidation that MDM has as standard. For example rules for value resolution if there are multiple values, workflows to manage exceptions, a UI to enter and edit data, etc. Even so, the data in many data warehouses will get out of sync with source systems if not managed properly.
If your data warehouse team is used to a star schema approach vs. more normalized data models, implementing the complex IDMP model may cause additional delay. Companies who start with the data warehouse approach for IDMP may find themselves re-creating many MDM features, and implementing data governance processes, but centered around a data warehouse vs. a MDM system which is a more natural fit. (Read more on my thoughts about IDMP compliance without data governance here)
Summary of Technology Options 1 and 2
From my experience these two options are primarily being considered by pharmaceutical companies who are not yet considering that data governance has a large role to play in IDMP compliance. (Or have few products, which reduces the importance of data governance). Conversations I have had with a number of consultants and analysts lead to a common conclusion: Although these solutions may be a good starting point at the moment due to low data governance maturity, the companies that choose them will probably migrate to a solution which offers a higher level of data governance within the short to medium term. I’ll cover the final two technology options in my next blog. In the meantime you can read on my views on the correlation between data governance and regulatory compliance here.