A Busy Day in the (CDRH) Neighborhood: Updates to the CDS and General Wellness Guidance Documents

On January 6, 2026, FDA issued two revised guidance documents related to clinical decision support (CDS) software and low-risk general wellness products. In a video announcing the revised guidance documents, Commissioner Makary states that they are intended to “cut unnecessary regulation and promote innovation” in the field of artificial intelligence (although neither guidance specifically mentions artificial intelligence). The guidance documents were issued without seeking public comment, which is unusual given the more than minor changes to policy and/or statutory interpretation embodied in the revisions.

CDS Guidance

The 2026 CDS Guidance, consistent with the tenor of Commission Makary’s announcement, reflects a more expansive reading of the statute’s CDS exemption and a greater willingness to exercise “enforcement discretion” with respect to software that may not technically qualify for exemption (more about that later). This updated guidance supersedes the 2022 CDS Guidance, about which we’ve previously blogged.

As a brief refresher, the Cures Act established four criteria for Non-Device CDS software.  Below we describe the key interpretive changes made by the new CDS Guidance with respect to each criterion; this in no way captures all the changes made in the update.

Criterion 1: Non-Device CDS software functions do not acquire, process, or analyze medical images, signals from an in vitro diagnostic (IVD) device, or patterns or signals from a signal acquisition system.

The 2026 CDS Guidance reinforces that CDS software can acquire, process or analyze more than one measured signal and still meet this non-device criterion, stating that “discrete, episodic, or intermittent point-in-time physiological measurements (e.g., routine vital signs obtained at discrete clinical encounters) generally do not, by themselves, constitute a pattern” and that signal acquisition systems measure parameters “through continuous, near-continuous, or otherwise streaming measurement.”

The 2022 version of the guidance already stated that “FDA recognizes there is a continuum between a single sample and a continuous sample,” so the new language is more a clarification than a real change.

Criterion 2: Non-Device CDS software functions display, analyze or print medical information about a patient or other medical information.

The 2022 CDS Guidance raised eyebrows when it interpreted “medical information about a patient” to mean “the type of information that normally is, and generally can be, communicated between HCPs in a clinical conversation or between HCPs and patients in the context of a clinical decision, meaning that the relevance of the information to the clinical decision being made is well-understood and accepted.” It was not at all clear what type of information is normally communicated between HCPs and/or patients.

The 2026 CDS Guidance modifies this somewhat with the caveat that “[w]hether particular information is commonly discussed in a clinical conversation is not, by itself, determinative of whether it is ‘medical information about a patient’ under Criterion 2, provided the information’s relevance to patient care is supported by well-understood and accepted sources and can be appropriately understood in context.” This language is helpful in that it removes the question of what constitutes information that is “normally communicated” between a patient and an HCP and instead reflects a broader interpretation that would likely cover most topics discussed between a patient and an HCP.

Criterion 3: Non-Device CDS software functions are intended for the purpose of supporting or providing recommendations about a patient’s care to a health care professional (HCP) user.

The changes to FDA’s interpretation of Criterion 3 are perhaps the most substantive and significant, but also a bit confusing. Both the 2022 and 2026 versions of the guidance state that a software function fails Criterion 3 if it “provides a specific preventive, diagnostic or treatment output or directive.” Seemingly in support of that position, the 2022 guidance was clear that a software function that provided “a user with a single, specific, selected output or solution rather than a list of options or complete information for the user to consider” was a device. That language no longer appears in the 2026 version. Instead, the guidance now states that “if only one option is clinically appropriate and the software function otherwise meets all criteria under section 520(o)(1)(E), FDA intends to exercise enforcement discretion” for such function and provides several examples of software functions that would and would not fall within this new enforcement policy. In other words, FDA is not saying that devices that provide a single output meet Criterion 3, but rather that even though they fail Criterion 3, FDA is choosing not to regulate them. It is not clear why FDA is enacting this enforcement discretion policy rather than merely expanding its view of the types of features that meet Criterion 3. The outcome is the same for developers, although this approach gives FDA the flexibility to withdraw that discretion should it deem necessary.

Similarly, the 2026 guidance deletes a controversial statement from the 2022 version that “FDA considers software that provides information that a specific patient ‘may exhibit signs’ of a disease or condition or identifies a risk probability or risk score for a specific disease or condition as providing a specific preventive, diagnostic, or treatment output. Therefore, such software would not satisfy Criterion 3.” Although the 2026 guidance stops short of saying such functions can satisfy Criterion 3, it includes several risk score functions as examples of products over which FDA would exercise enforcement discretion.

One such example is as follows:

A software function that predicts risk of future cardiovascular events for an HCP to consider based on a patient’s weight, current and historical smoking status, blood pressure, and brain natriuretic peptide (BNP) in vitro diagnostic (IVD) test results.

What is particularly interesting here is that FDA is attempting to carve out narrow circumstances in which this type of prediction would be acceptable without marketing authorization. While the above example would be permitted, if the same functionality relied on “variant genomic data as an input that does not have established relevance to the diagnostic recommendation” then the feature would be subject to FDA oversight. Additionally, if the feature “predicts risk of a cardiovascular event in the next 24 hours” then it would also require premarket review. It is not clear what predication time frame would be appropriate to benefit from enforcement discretion, and what would require review. It is also clear that manufacturers of CDS will need to carefully consider the inputs to the feature to determine whether they fit within the enforcement discretion carve-out.

Criterion 4: Non-Device CDS software functions provide sufficient information about the basis for the recommendations to the HCP user, so that the HCP user does not rely primarily on any of the recommendations to make a clinical decision about an individual patient.

This last Criterion has always been a difficult one for developers to implement. The 2022 CDS Guidance seemed to impose disclosure requirements that were virtually impossible to meet, and it was not clear how to strike a balance between providing sufficient information to allow an HCP to review and confirm the feature’s output without needing to disclose proprietary information. The 2026 version helps clarify these disclosure requirements. It largely maintains the list of “software and labeling recommendations” that the 2022 guidance set forth in order to satisfy Criterion 4, including a recommendation in both versions that the software or labeling provide a plain language description of the software inputs and underlying algorithm that form the basis of the CDS output. However, the 2026 guidance makes several new statements suggesting that FDA is looking for less detailed labeling disclosures than previously expected.  For example, it recommends that the software or labeling include a “summary of the general approach relied upon to provide the recommendations . . . at a level of detail appropriate for the intended user and use environment, which could include, for example, the logic or methods relied upon.” Similarly, the new guidance states that “Information that enables an HCP to independently review the basis of provided recommendations is presented in a manner that promotes usability and avoids information overload, including prioritizing the most-decision-relevant information and making additional detail available as appropriate.” For companies with complex algorithms or that want to protect proprietary technology, this update is a welcome change, as the disclosed information can be more general and include only the information most important to the intended users.

The 2026 guidance also moved the discussion of automation bias that previously appeared in the section interpreting Criterion 3 to the section interpreting Criterion 4, concluding that:

FDA considers the (a) level of software automation and (b) time-critical nature of the HCP’s decision making when determining whether a software function allows an HCP to independently review the basis for the recommendations presented by the software so that they do not rely primarily on such recommendations.

Interestingly, FDA removed all but one of the references to “time-critical” from Criterion 3, but revised examples of device and non-device CDS at the end of the guidance to state that time-critical software functions failed both Criterion 3 and 4. Practically speaking, if a software function would have failed Criterion 3 due to a high level of automation or the time-critical nature of its intended use, it would already have been considered a regulated device. Clarifying that such a software function would also fail Criterion 4 doesn’t change anything.

General Wellness Guidance

CDRH also issued an update to its guidance, General Wellness: Policy for Low Risk Devices, commonly known as the “General Wellness Guidance.” Although FDA did not change its definition of a general wellness product, it greatly expanded the scope of products that can fit within that definition. This is likely to result in many more products coming to market without FDA premarket review.  Perhaps most notably, language in this revised document guts the WHOOP Warning Letter, on which we posted here. As a quick refresher, in that Warning Letter FDA stated that WHOOP’s blood pressure wearable could not be a general wellness product even if tied to wellness uses because “the product is intended to provide a measurement or estimation of a user’s blood pressure, which is inherently associated with the diagnosis of hypo- and hypertension, and is therefore intended for use in the diagnosis of a disease or other condition, or in the cure, mitigation, treatment, or prevention of disease.”

Enter the revised General Wellness Guidance, which states:

FDA may consider certain products that use non-invasive sensing (e.g. optical sensing) to estimate, infer, or output physiologic parameters (e.g. blood pressure, oxygen saturation, blood glucose, heart rate variability) to be general wellness products when such outputs are intended solely for wellness uses, and provided they:

• are non-invasive and not-implanted;
• do not involve an intervention or technology that may pose a risk to the safety of users or other persons if specific regulatory controls are not applied;
• are not intended for the diagnosis, cure, mitigation, prevention, or treatment of a disease or condition;
• are not intended to substitute for an FDA-authorized, cleared, or approved device;
• do not include claims, functionality, or outputs that prompt or guide specific clinical action or medical management; and
• do not include values that mimic those used clinically unless validated (e.g. manufacturer testing, peer-reviewed clinical literature) to reflect those values.

Products that meet the aforementioned criteria may display values, ranges, trends, baselines, or longitudinal summaries, and may contextualize these outputs in relation to sleep, activity, stress, recovery, or similar wellness domains.

As we noted in our prior blog post, FDA seemed to be on thin ice in the Warning Letter, having crafted out of nowhere the concept of a claim being “inherently associated” with a disease or condition and therefore seemingly barred from being deemed a general wellness product. Now, the agency has done a pendulum swing in the opposite direction, essentially stating that any non-invasive sensing of physiologic parameters can fit within FDA’s view of general wellness, so long as they steer clear of any disease claims and avoid claims of clinical accuracy. The guidance also states that it expects general wellness products to have labeling consistent with, and not exceeding, the general wellness intended use.

On the whole, the updates to this general wellness guidance do provide much-needed clarity regarding FDA oversight of monitoring of physiological parameters for non-medical purposes. While many large companies, such as Apple, Garmin, and Google, have long had unregulated heart rate and other monitors on their fitness watches, the basis for concluding that those were not medical devices was always a little hazy, particularly when those features alert users to heart rates above or below established thresholds. It is now clear that, so long as labeling for these products follows the recommendations set forth in the guidance, measurement of a physiological parameter alone is not enough for the product to be deemed a medical device. These updates may also help level the playing field; in our experience, smaller, more risk-averse companies would seek FDA guidance on products that seemed similar to some of the large tech company offerings, only to be told that their products were devices due to the measurement of physiological parameters, even if not intended for use in a clinical environment. Such a position would now appear to be in contrast with the updates to the General Wellness Guidance.

While the guidance provides helpful instructions as to boundaries for general wellness products, one area of potential confusion pertains to the ranges that can be utilized in a general wellness product. The guidance appears internally inconsistent in this regard. On the one hand, the guidance states that a general wellness product may “not include values that mimic those used clinically unless validated (e.g., manufacturer testing, peer-reviewed clinical literature) to reflect those values.” (Emphasis added.) However, the guidance later states that a general wellness product may not include diagnostic thresholds at all, although it may include a notification “informing a user that evaluation by a healthcare professional may be helpful when outputs fall outside ranges appropriate for general wellness use, provided that such notifications . . . do not characterize the output as abnormal . . . [and] do not include clinical thresholds . . .”

The idea of a “range appropriate for general wellness use” that is somehow distinct from clinical thresholds is novel and seems to indicate that the manufacturer would need to establish its own independent range distinct from clinical utility. It is not clear how such ranges would be determined, and whether they would be consistent across similar general wellness products. It also cuts against other language indicating that clinical values may be included so long as they are validated.

While it is helpful that the guidance explicitly states that a general wellness product can tell a user to seek input from a healthcare professional, the guidance states that the feature cannot tell the person to seek such input because a result is abnormal. Telling a person to seek medical advice without telling them why will undoubtedly create confusion for consumers.

What Does It All Mean?

If you have read this far, thank you, and you may be past the point of wanting the TLDR, but we’ll give it to you anyway. These updates, made without any public input and by the Commissioner himself, may indicate that the administration, for better or worse, is beginning to pay more attention to medical devices, or at least to digital health products. They also appear to reflect an intent to align the interests of the administration with the approach taken by CDRH, which, as we have previously noted, have been seemingly inconsistent to date.

Overall, these updates strike us as reflecting this administration’s push towards deregulation and broader uptake of wearables, but there is a question as to whether the agency has, to some extent, prioritized innovation over patient safety. We agree that FDA’s oversight of low-risk software products has historically been overly conservative at times, but software products providing risk scores or likely diagnostic outcomes to HCPs may present certain risks to patients if those features have not been properly validated.

We encourage companies to reach out to us with any questions you have in interpreting FDA’s new guidance documents and how they may apply to your products.