FDA narrows scope of regulatory oversight for wearables, clinical decision support

In its latest efforts to make regulatory oversight a thing of the past, the Trump Administration has announced that the FDA will pull back even further on scrutinizing the worlds of wearable devices, consumer-focused health apps, and even some clinical decision support (CDS) tools used by healthcare providers: areas of the health IT marketplace that were already fraught with potential for misdirection from developers and misuse from consumers.  

“Wellness” gets a pass from the FDA’s hands-off approach to regulation

New guidance issued this week excludes certain “low risk” wellness products from being defined as devices subject to FDA oversight. The exceptions now include devices and apps “intended for maintaining or encouraging a healthy lifestyle that are unrelated to the diagnosis, cure, mitigation, prevention, or treatment of a disease or condition,” the FDA says.  

Products no longer subject to the rules around medical devices may include wearable devices and smartphone apps that provide biometric information to users and analyze that information to offer insights about a person’s wellness – as long as the manufacturer doesn’t make any claims about the product’s functionality that imply the product’s use in a medical or clinical context: a blurry and subjective line at best. 

The guidance document specifically calls out newer categories of consumer devices as no longer subject to the previous definition, such as over-the-counter continuous glucose monitors (CGMs) that are “marketed to users as a means of better understanding their insulin response to certain foods” but explicitly state they are contraindicated for people with diabetes and pre-diabetes.   

“We want to let companies know, with very clear guidance, that if their device or software is simply providing information, they can do that without FDA regulation,” FDA Commissioner Mark Makary said in an interview with Fox Business’ Varney & Co. 

“The only stipulation is if they make claims of something being medical grade… We don’t want people changing their medicines based on something that’s just a screening tool or an estimate of a physiologic parameter.” 

While the new guidance is ostensibly designed to promote innovation and bring more health tech options to the general public, as HHS Secretary Robert F. Kennedy, Jr. has previously championed, the effort adds another wrinkle in an already complex environment.  

Even low-risk wearables and other medical devices are being built with artificial intelligence at the core, which the Trump Administration has also declared a free-for-all with minimal regulatory oversight.  

Bad AI combined with devices that push their “wellness” claims to the very limit with sneakily flowery language could be bad enough. But with tens of millions of people about to become uninsured due to inaction on ACA subsidies and brutal Medicaid cuts, there’s a strong likelihood that consumers desperate for any type of healthcare at all will be turning to cheap, available, seemingly helpful devices and apps at an even higher rate. 

Those over-the-counter CGMs, for example, that are significantly cheaper and more accessible than prescription options, may quickly turn into an attractive – but clinically insufficient – workaround for people with diabetes who can’t get the appropriate care they need. These patients may end up making clinical decisions with unregulated devices and algorithms regardless of the fine print on the box, and there’s no way of telling whether using consumer-grade CGM data to manage their diabetes will be better or worse than having no data at all. 

Stay at the forefront of clinical care innovation: Join nurse and physician leaders for real-world insights, solutions, and conversations shaping care delivery at ViVE: Nurses@ViVE26 and Physicians@ViVE26.

A narrower take on what constitutes clinical decision support  

The FDA also published concurrent guidance targeting CDS tools, which similarly removes certain platforms and solutions from the pool of regulated devices.  

In order for a device to be excluded from the necessity for oversight, it must meet four criteria: 

• Not intended to acquire, process, or analyze a medical image or a signal from an in vitro diagnostic device or a pattern or signal from a signal acquisition system 
• Intended for the purpose of displaying, analyzing, or printing medical information about a patient or other medical information 
• Intended for the purpose of supporting or providing recommendations to a healthcare provider (HCP) about prevention, diagnosis, or treatment of a disease or condition 
• Intended for the purpose of enabling an HCP to independently review the basis for the recommendations that such software presents so that it is not the intent that the HCP rely primarily on any of such recommendations to make a clinical diagnosis or treatment decision regarding an individual patient 

In short, the majority of tools that aggregate and analyze non-biological data and present the resulting clinical insights as options instead of answers are no longer subject to the existing regulatory framework.   

The guidance document cites several examples of products that will not require a look by the FDA, including functions that identify drug-drug interactions or drug-allergy contraindications for specific patients, tools that analyze patient-specific symptoms to provide a prioritized list of diagnostic tests for the HCP to consider, and products that use data from a patient’s medical record to recommend prioritized treatments. 

The potential problem, of course, is that every single element of the patient’s data has to be perfect in order for these algorithms to work well enough to guide decision-making – and there’s no question that patient data is laughably far from being current, complete, accurate, and trustworthy. 

Without involvement from the FDA, healthcare organizations themselves will need to shoulder the responsibility of vetting vendors, deeply understanding the data science behind their offerings, and monitoring the quality of their outputs in the real-world clinical setting, which has not been particularly easy for health systems trying to navigate a flood of AI vendors pitching next-gen CDS products.   

In addition, both organizations and patients may have fewer options for recourse if it turns out that a flaw in the model has caused some degree of harm. 

With this regulatory step back from a class of tools that healthcare providers are implementing in droves, it’s going to become increasingly difficult to identify high-quality options and ensure patient safety. 

Healthcare leaders will need to be prepared to conduct their own thorough assessments of CDS products, including interrogating vendors on model transparency, algorithmic bias, compliance issues, and data ownership, to ensure their providers are equipped with CDS tools that are safe, secure, and effective. 

Jennifer Bresnick is a journalist and freelance content creator with a decade of experience in the health IT industry.  Her work has focused on leveraging innovative technology tools to create value, improve health equity, and achieve the promises of the learning health system.  She can be reached at [email protected].