In particular the update of the European regulations has brought the focus on clinical evidence and requirements for manufacturers to demonstrate clinical performance. In Europe this is typically done through having documented clinical evidence in the Technical file.
So what is Clinical Evidence”? Put simply:
Clinical Data + Expert Review = Clinical Evidence
There are a couple of key points here: The data come from a range of sources including relevant published clinical literature, in house (unpublished) clinical reports and data, postmarket feedback, external postmarket data such as registries or postmarket databases such as FDA’s MAUDE or TGA’s IRIS, and actual clinical trial data. The extent of clinical data and the need to include actual direct clinical trials depends on the risk of the device. Obviously higher risk devices require more extensive data and evaluation.
The requirement for clinical evaluation was always there for all devices in the directives from 1993. However it was contained in a very brief few words which was, shall we say, re-interpreted by manufacturers and notified bodies to be not really required for low risk devices. The 2007 revision of the MDD fixed that – adding several paragraphs of clarification (if that’s not an oxymoron) that clinical evaluation is required for every device irrespective of class and that the higher risk devices were expected to have stronger evidence.
The current draft European regulations go one step further and make the default position for Class III devices to be direct clinical trial data, unless a credible justification for an alternate approach can be provided and accepted. Notified Bodies are likely to be very conservative about such arguments.
With regard literature based evaluations, we see so many manufacturers fall down because they haven’t got a pre-defined search strategy and they cherry pick only the favourable literature. Australia’s TGA, which uses a regulatory system very closely based on the European Directives, has been particularly strict on this for years.
The minimum expectations for a properly done clinical literature review should be :
- Search databases defined in advance and justified for applicability and relevance of coverage.
- Search terms defined in advance and justified for relevance and completeness.
- Exclusion/inclusion criteria defined in advance and justified. These are not the same as search terms. Search terms generate the list of retrieved articles. Inclusion criteria allow you to inspect the article titles and abstracts and cull that list down to articles of actual relevance to the evaluation. Inclusion criteria may be around things like the specific clinical indications addressed in the article or the nature of the studies (e.g. excluding single case reports or poorly controlled studies)
- Articles must be reviewed and considered if they meet the criteria – in other words – no selection of just the favourable articles. If there are articles unfavourable to your cause they *must* be considered and addressed in the review. In other words the review must be comprehensive and balanced. There’s good odds that the reviewer will know about the unfavourable stuff anyway so it’s a really bad idea to ignore and hope they don’t notice.
And what about the “Expert Review”? A very important consideration is the expertise of the author. it’s common practice in Europe for clinical evaluations to be prepared by a product specialist within the manufacturing company. Fair enough – they probably know the device better than most. However there is a consideration of the clinical expertise of the author. Here in Australia the TGA have always insisted on the literature review and broader clinical evaluation report be written or at least reviewed and endorsed by a relevant clinical expert – i.e. an experienced clinician who would be familiar with the use of the device. Expect to see this kind of expectation appearing in more jurisdictions. The clinician doesn’t always have to be a doctor. We have helped prepare reports for dressing packs where the expert was a nurse, equipment for use in ambulances where the expert was a paramedic and reports for X-ray machines where the expert was a radiographer.
A good clinical evaluation is not just a European thing. It can be enormously helpful in a 510(k) in addressing equivalence with predicate device(s). After all – that’s what a literature review is partly doing – looking at (all the) other devices in the market and their performance as it is relevant to your device. Elsewhere – if you look at the new regulations in China – it’s really interesting that their Clinical Evaluation Report format actually includes a comparison table with predicate devices which is closely based on the FDA’s 510k comparison table approach.
Regulators around the world are looking to strengthen clincal evaluation requirements for higher risk devices – it’s an area that is a leading cause of failure of high risk device submissions and needs to be got right.
In summary, consider all of the evidence, do it in a structured and defendable way, be balanced in your approach and bring appropriate expertise to the table. Simple really…
China Clinical Evidence Guide (Devices)
China Clinical Evidence Guide for IVDs (Mandarin)
US FDA Clinical Guidances
MEDDEV 2.7/1 rev.3 Clinical evaluation: Guide for manufacturers and notified bodies
MEDDEV 2.7/2 Guide for Competent Authorities in making an assessment of clinical investigation notification
MEDDEV 2.7/4 Guidelines on Clinical investigations: a guide for manufacturers and notified bodies
Australia TGA Clinical Trials page
Japan Guidance – use of foreign clinical data in submissions
Need help with Clinical Evaluation? We can help you with literature surveys, preparation of reports or with advice on clinical trials and specific regulatory requirements in different markets. Contact us now to discuss your needs.