Digital health technology encompasses a variety of innovative healthcare tools including clinical decision support (CDS) and patient decision support (PDS) technology. These terms are used loosely to cover everything from a proprietary algorithm recommending optimal cancer therapies for individual patients to an app that relies on published guidelines to encourage physical activity. Some digital health products stand alone while others are accessories to conventional medical devices. As companies from Apple and Google to garage start-ups look to enter the health technology space with a broad range of medical software innovations, there are many questions related to which products will be regulated by FDA as medical devices and to what degree.

FDA has recognized the need for guidance in this area since at least 2011, and has slated digital health technology as a top priority since 2015. See FDA, Public Workshop – Mobile Medical Applications Draft Guidance; FDA, CDRH Fiscal Year 2015 (FY 2015) Proposed Guidance Development and Focused Retrospective Review of Final Guidance. But little has happened and many in industry have complained that the lack of clear regulatory guidance has hindered innovation that could benefit doctors and patients.

On December 8, 2017, FDA provided some greater clarity, issuing two draft and one final guidance. The final guidance adopts a set of principles and terminology agreed upon by an international group of medical device regulators. The two draft guidances, which are the primary focus of this article, address some of the issues left unresolved by the December 13, 2016 passage of the 21st Century Cures Act. Section 3060(a) of the Cures Act, which amended the FD&C Act to add Section 520(o), excluded certain software functions from the definition of a medical device that would be regulated by FDA. It also left substantial room for interpretation and further uncertainty.

FDA’s Function-Based Approach

The two draft guidances, “Changes to Existing Medical Software Policies Resulting from Section 3060 of the 21st Century Cures Act” (hereinafter “Draft Guidance on Changes to Medical Software Policies”), and “Clinical and Patient Decision Support Software” (hereinafter “Draft Guidance on Decision Support Software”), reflect FDA’s current thinking and recommendations for how industry may better interpret the language of the FD&C Act regarding the exclusion and inclusion of certain software functions. The first draft guidance clarifies the Agency’s interpretation of Section 520(o)(1)(A)-(D) of the FD&C Act, which discusses products with different functionality, including most significantly products for “maintaining or encouraging a healthy lifestyle.” See Draft Guidance on Changes to Medical Software Policies at 5. The second draft guidance provides greater insight to FDA’s thinking on Section 520(o)(1)(E) of the FD&C Act, which excludes from the definition of a device certain products that may be used by both healthcare providers and patients in making healthcare decisions.

Clarifying how Section 520(o)(1)(B) affects FDA’s existing General Wellness guidance, the first draft guidance states that a product will not be a medical device if it is only for encouraging or maintaining a general state of good health or a healthy lifestyle, presents a low safety risk and is not “related to the diagnosis, cure, mitigation, prevention or treatment of a disease or condition.” See Draft Guidance on Changes to Medical Software Policies at 8; see also Guidance on General Wellness: Policy for Low Risk Devices, issued July 29, 2016. In other words, an app on a watch that helps track energy consumption or output or recommends better dietary choices is not a medical device. Where, however, the software product relates a healthy lifestyle or healthy activities to something like lowering the risk of diabetes or heart disease, it is. Products that offer advice or conclusions on how healthy lifestyle choices may impact lifestyle diseases are medical devices as they are technically “related to the diagnosis, cure, mitigation, prevention or treatment of a disease or condition.” See Draft Guidance on Changes to Medical Software Policies at 8. Notwithstanding the categorization of these products as devices, however, FDA states that it intends to refrain from exercising its enforcement authority in this area, provided that the technology is low risk.

The second draft guidance provides greater insight to FDA’s thinking regarding FD&C Section 520(o)(1)(E), which excludes from the definition of device a CDS product that: (1) is not intended to take or analyze a medical image or signal (such as an MRI or ECG machine); (2) is intended to display patient or general medical information (such as practice guidelines); (3) is intended to provide recommendations for healthcare professionals; and (4) is intended to have the healthcare professional exercise his or her independent medical judgment about a specific patient without primarily relying on the CDS. See Draft Guidance on Decision Support Software at 5. The guidance also notes that PDS software with similar functionality and limitations—such as an app that helps patients know how to dose a medication consistent with its approved labeling—will not trigger FDA oversight. It also lists numerous examples of exempt and non-exempt products.

As with healthy lifestyle products, the regulatory lines are drawn mostly on functionality. If the CDS software uses publicly available information, applies that information to existing patient-specific data to recommend a diagnosis or treatment option, but leaves any final clinical decision to the individual healthcare provider, it is likely to be exempt. For PDS, the test is similar with the main difference being that the recommendation has to be one a patient could appropriately follow without expert medical consultation. In other words, software that lets doctors and patients make the kinds of healthcare decisions they have always made, but with potentially faster and more complete access to accepted relevant data via an electronic tool, will not be considered a medical device.

A Good Step – But What’s Next?

Most of the public reaction to the draft guidances thus far has been positive, viewing them as a step towards a regulatory approach that balances the need for potentially useful technological innovation with concern for patient safety and preservation of the doctor’s central role in medical decision-making. Some commentators, however, have suggested FDA has not been sufficiently nuanced in its approach and has failed to draw other potentially helpful lines.

Some have asked why FDA’s lines do not depend more on the seriousness of the disease or the degree of patient risk that could flow from the software’s recommendation. Should an app that recommends an optimal chemotherapeutic choice (even if using established guidelines) be treated the same as an app that recommends an optimal allergy medication?

Also, one may question the implications of FDA treating general wellness products as devices if they tie lifestyle choices to disease risks, but at the same time declining to use its regulatory authority over those devices. Will this create a more certain or less certain environment for innovation? And will it be a good result or not-so-good result for public health?

Another question relates to whether the complexity of the dataset and the methods for distilling information from it should be more important regulatory considerations than whether the data are publicly available. For example, two products may both rely on public information, but the one that uses only approved product labeling and practice guidelines may be a very different type of tool in the hands of a physician from the one that aggregates and wrests conclusions from various clinical and observational data. Will the physician truly have transparency regarding and be able to independently understand all of the published data utilized, when the data are updated, or how the algorithms are analyzing, weighing and interpreting the data? On the other hand, is this materially different from the current imperfect system where many clinicians likely do not have time or ability to fully review and digest the published literature?

FDA’s thinking on exempt versus non-exempt products also depends heavily on whether they dictate or merely “recommend” a course of action to the learned intermediary. But the fact that a product is characterized as making only a recommendation, which may be accepted or rejected by the healthcare provider, does not entirely predict how it will be used in clinical practice. Will the learned intermediary still exercise the same level of independent clinical judgment when faced with a technology-driven guide that is potentially part of the patient’s record? And will any shift, on balance, be good for patients?

Out of the Frying Pan and into the Fire?

For those thinking about business risks for innovators in this area, it should also be clear that a more hands-off regulatory scheme could mean greater litigation risks in the form of consumer fraud, false advertising or other claims. Just because a product with a healthy lifestyle purpose is not FDA-regulated does not make it any less vulnerable to the same kinds of claims facing manufactures of supplements, foods, and other consumer products, whose claimed benefits are often the subject of litigation challenges. In fact, less regulation may bring even greater litigation risks.

One may also imagine a host of litigation pitfalls related to product disclosures and warnings. For example, will there be lawsuits claiming that in the absence of greater regulatory oversight, doctors were not able to fully understand the strengths and limitations of new tools?

Even with unanswered questions, the use and advancement of technology in the field of digital health show no signs of slowing down. The new FDA draft guidance documents help cement and solidify the existence of CDS and PDS software as important tools in medicine. To harness the full potential this field has to offer, the industry should continue to ask questions and encourage discussion in an ever-changing era of digital health. The public comment period for the draft guidance documents closes on February 6, 2018.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.