As medical apps become increasingly sophisticated, their impact on clinical decisions and outcomes is growing as well. For this reason, the attention of the FDA has turned to the subset of medical apps that fall into a category of clinical decision support (CDS) – apps that take patient-specific information, process it, and output an analysis that impacts clinical decisions.
As a group of clinicians focused on mobile health, we were fortunate to be invited to share our perspective as part of one of the two panels discussing this topic.
As we highlighted in our remarks at the opening of the second panel, the medical app market, even just the subset of apps that fall under the “clinical decision support” definition, is extraordinarily broad. From a simple medical calculator or generating preventative health recommendations to apps that could adjust pacemaker parameters or change the infusion rate on an insulin pump, the types of decisions and the risks associated with them are very broad.
And so during the second day of the FDA meeting on mobile medical apps, when the attention turned to clinical decision support, a diverse group of stakeholders emerged. And the opinions shared offered not only a window into what future regulations may look like, it also offered hints of the direction that the industry as a whole is taking.
Opening the day, the FDA highlighted three key factors it is considering as it shapes the regulatory framework for mobile medical apps:
One of the first themes to emerge during the discussion is that in tackling this diverse marketplace, it is important that the FDA keeps the scope of regulation manageable. Take for example alcohol abuse – according to a recent study, there are 200+ apps that help patients manage this condition. A number of them would technically meet the definition of CDS as they take patient specific information, provide some level of analysis, and output a recommendation.
However, the risk associated with these apps is likely to be quite low. Therefore, even if the FDA wanted to apply some regulations to these apps, excluding them from the scope of regulation is likely to be a necessity.
As for apps with slightly higher risks – for example medical calculators which generate risk scores, drug doses, and so on – one important consideration became the presence of a “learned intermediary.” Basically the idea is that there is someone between the patient and the app (a physician, nurse, etc) who either uses the app’s output as part of a larger clinical picture or can recognize erroneous output e.g. a FENa or an unrealistic insulin dose.
Another theme which emerged during the day’s discussions was the importance of the medical community taking the initiative to help answer questions of efficacy for CDS apps. For example, a far better approach than to use the “low-risk apps” method would be for medical professionals to recommend specific apps based on best-practices (certainly an approach we at iMedicalApps clearly believe in). Much of this would likely be based on expert opinion. The traditional randomized clinical trial model is, by nature, behind the curve as they take years to put together.
There was little disagreement that apps which advise patients on insulin doses, function as part of larger disease management systems, analyze/alter a pacemaker or other implanted device, and so on warrant FDA regulation on the Class II or III device level.
But what about when these types of decision support are integrated into an EHR or PHR? The current guidance suggests that the same CDS functionality, if put in an EHR, would be excluded. Additionally jurisdictionally, regulating becomes more complicated as other government entities (ONC for example) become involved.
The impact of mobile medical apps on clinical decision-making is likely to grow over the coming years. And as the major stakeholders were present this week, the FDA made clear that the challenge will be to put appropriate checks in place without slowing the pace of innovation.