The EMA ISO IDMP Information Day meeting that Ian and I attended yesterday (23rd June 2015) did not bring any surprises, but it did leave me thinking that implementing IDMP is actually achievable, unlike XEVMPD three years ago.
The good thing is that EMA seem to be trying to do it right this time. Their general concept of consolidating systems and sharing/reusing data is very sensible and echoes what we at Samarind have been promoting since the beginning; that there should be a ‘single place of truth’ for each piece of data. They have clearly thought very carefully about the business side of things too; the presentations today contained lots of use cases, clarity about how the IDMP data will be used and recognition of the wishes and wants of all stakeholders – the EMA itself, industry, NCAs and even software vendors. Efficiency, accuracy, improved communications and transparency were the bywords of the day.
There is no escaping the fact that there are challenges as well as opportunities. The short timelines, the many complex dependencies between so many different organisations, the sheer amount of process and technical work that needs to be done all conspire together to make this a huge task. However here again I think common sense has prevailed.
Rather than rush to implement by the prescribed legal deadline, like they did with XEVMPD, EMA are taking a phased approach with IDMP, and bringing in timelines based on the publication of the final guidance (a bit like the FDA do) rather than a fixed deadline where no-one knows when the associated guidance will be available.
Products are being addressed first, because this information is needed for day-to-day operations and is closest to the existing XEVMPD database. They are planning to implement in five separate iterations over the next 3 years.
- Iteration 1 for products is basically the same scope as XEVMPD plus the minimum amount of extra data required to cover the 3 main IDMP ID fields, MPID (medicinal product id), PCID (packaging id) and PhPID (pharmaceutical product id). Guidance will be published Q1 2016; implementation will start 12 months afterwards in Q1 2017, and it will become mandatory 9 months after that in Q4 2017.
- Iteration 2 covers Investigational (Development) Medicinal Products. Guidance is expected Q4 2016, start of implementation Q4 2017 and it will become mandatory 6 months later in Q2 2018.
- Iteration 3 is yet to be defined (planned for 2016) but is expected to include the remaining Clinical Particulars data. There will be the same 12 + 6 month implementation cycle.
- Iteration 4 will also be defined during 2016 and is expected to include batch identifiers and the remaining EU ISO 11615 and ISO TS 20443 items.
- Iteration 5 is also yet to be defined but will cover veterinary products.
Substances will be next, partly in parallel; most functionality will be in the first iteration, for which guidance should be available Q2 2016, implementation Q2 2017 and mandated in Q4 2017 (the same time as Products). Then the other controlled vocabularies (“Referentials”) will be implemented between Q1 2016 and Q4 2016 and Organisations between Q1 2016 and Q4 2018.
These timetables, in my view, are very sensible and make the whole thing achievable, for all parties concerned. There was more good news; EMA will migrate all data from XEVMPD and provide a free tool with similar functionality to EVWeb but with a better, more up-to-date user interface.
So what about the 1st July 2016 deadline? Well, all EMA’s plans are geared around the timetables described above and, as the meeting chair said when asked what would happen if the European Commission refused to budge, “There is no plan B”. So whilst it’s not official yet, I think this is what will happen, and we can all heave a sigh of relief 🙂