Is my study population the right one, or in other words: Does my clinical study population represent the patients the study product is meant for? An essential question to answer when designing a clinical study to collect clinical evidence regarding product performance. Especially in the area of medical devices, but also food, where the intended population can very broad, such is less straightforward as compared to pharmaceutical studies. And a recent publication illustrates that it is harder than it seems to choose the right clinical study population.
Key factors driving the choice of the study population, as previously summarized in a presentation, are:
Depending on the amount of existing clinical evidence, the study patient selection criteria will be more or less strict. The less data is available the more the clinical study population will be limited to protect study subject safety, and the further it deviates from the intended population. In pharma such is dealt with via the different premarket phases, but in medical devices this is not the case and typically at market release the available data is minimal. In medical devices one also needs to weight in the class of device: selection criteria for a first in man clinical study for an active implantable device will be more strict than those for a non-invasive monitoring device.
The consequence of the above, is that the clinical study population can differ substantially from the final population the product is used for. In daily practise patient selection criteria will be less strict and factors like co-morbidity and age may well affect safety. Therefore in the end safety will need to be established in daily practise. If not done pre-market, for example because the concerning product is implanted and its mode of action stretches over a period of several years, a thorough postmarket surveillance program is a must. Which is exactly one of the reasons why there is so much attention by the Notified Bodies and the MDD under revision for an adequate postmarket clinical follow-up, following the hip and PIP implant scandals.
When still choosing to collect clinical evidence on efficacy, for example for reimbursement purposes, patient selection criteria are often very strict to reduce variability and maximize product effect at the lowest possible clinical study size. When there is no safety issue, one can choose to widen patient selection criteria and stratify for patient sub-groups and analyse for confounding factors. This, however, will lead to a rise in the study sample size: One only needs to look into the medicinal study sizes to get an idea.
Regardless such efforts, however, it will still be challenging to mimic a broad population of people including all ages and co-morbidities, in a non-study setting with less extensive or other patient diagnostics and device training. Therefore daily practice trials remain necessary to evaluate to what extent it erodes device efficacy. Keeping in mind the acknowledgment in pharma that there is a gap between efficacy and effectiveness, one may even wonder whether for some medical devices, such as for example monitoring or diagnostic devices, it does not make sense to ‘skip’ efficacy studies and rather start an observational study after market release instead?
As mentioned above, widening of the clinical study population leads to a larger sample size, and then the question becomes relevant whether the concerning clinical trial is still feasible in terms of people, time and money. This embeds an ethical concern also: can we start a trial that we may not be able to finish, especially taking into account the fact that the majority of clinical trials take longer than expected? In that light I firmly believe one should always carefully weight what clinical project(s) to spend your resources on: a large efficacy study with a select patient group that will need a follow-up study in a real world setting, or a clinical study showing that the product works in the intended population in daily practise.
In conclusion one can say that it is hard to define a clinical study population that covers all aspects in one clinical study for safety, efficacy, but also for feasibility reasons. Whether such can be done or whether it is more cost-effective to run 2 or more clinical studies needs to be determined on a case by case basis.
Please do not hesitate to contact me in case you want to discuss any of the above or stay tuned with my blog: the next one on this series on clinical evidence dilemmas will follow soon.