The U.S. Food and Drug Administration (FDA) is most commonly known for its work in regulating food, drugs, biologics, medical devices, cosmetics, veterinary products, and tobacco products. As health care continues to modernize and employ digital tools to assist in evaluating, treating, and educating patients, the FDA is responsible for regulating many emerging digital health technologies that represent the future of health care. This article describes recent advances in health care devices and focuses on the FDA’s involvement in regulating these devices. Immediately following this summary on page 16 is an interview that the authors, Vinita Mujumdar, JD, and Haley Jeffcoat, MPH, conducted March 3 with Bakul Patel, MSc, MBA, Director, the FDA’s Digital Health Center of Excellence (DHCoE).
The health technology explosion
In today’s digital world, the use of technologies such as smartphones, social media, Internet applications, and more are commonplace and have reshaped many aspects of life—including the way we communicate, work, shop, gather information, and so on. These digital advancements also expanded the ways individuals can monitor their personal health and wellness, and offer health care providers new opportunities to monitor, treat, and engage with patients. These advancements are leading to a convergence of information, technology, and connectivity to improve health care and health outcomes.
With tools such as electronic health records (EHRs), registries, smart medical devices, wearables, sensors, and so on, physicians can gather more patient data than ever before. By improving our ability to access, exchange, integrate, and consume data through increased interoperability, these patient data can be shared and cooperatively used across delivery systems, as well as beyond organizational, regional, and national boundaries. But with this increase in data, caring for patients will require more robust knowledge management. Moreover, the ability to track patients and their care more closely with increased data and knowledge allows for greater identification of variances in care processes and patient outcomes. Conformance with preferred clinical pathways can be assured, however, by defining standardized clinical processes and appropriate algorithms that can be used in digital health tools. For surgeons and surgical patients, the effect of these changes will depend on the extent to which they can trust the data and algorithms used to establish the knowledge needed to secure optimal care for patients in a safe and cost-effective way.
Digital health and its benefits
Digital health covers a spectrum of digital tools, services, and devices that support clinicians in making clinical decisions and delivering care, and support patients through education and streamlining how they receive care. Digital health technologies use computing platforms, connectivity, software, sensors, wearable medical devices, mobile medical apps, artificial intelligence (AI) and machine learning (ML) algorithms, and more that can be applied in ways that monitor patients’ health status, facilitate the safe exchange of health care data, support medical devices, and aid many other aspects of health care delivery. As modern health care continues to evolve, the advancement of digital health tools and technologies offers substantial possibilities that can improve the accuracy and effectiveness of disease diagnosis and treatment and enhance the delivery of health care services for all patients.
Benefits of using digital health technologies in health care practices include reduced cognitive burden on physicians, fewer inefficiencies, improved access, lower health care costs, increased quality of care, enhanced personalized care for patients, and more efficient communication and information exchange between health care providers. These tools and technologies also allow for a combination of health care data streams to provide a more holistic and longitudinal view of patient health. In addition to the many benefits for health care providers, digital health tools also assist patients by providing increased access to, and greater control over, their health data.
Digital health tools and technologies fall under the FDA’s definition of devices, which are instruments used in the “diagnosis, cure, mitigation, treatment, or prevention of disease, or that affect the structure or function of the body.”1 The FDA has the authority to regulate medical devices, including software that is considered a medical device. This type of software, termed Software as Medical Device (SaMD), is defined as “software intended to be used for one or more medical purposes without being part of a hardware medical device.”2 SaMD can be used in a variety of platforms, such as medical devices, general computing platforms, and virtual networks. A few examples of the many types of SaMD include the following:
- Software that helps radiologists and clinicians find and diagnose a cardiovascular condition by analyzing magnetic resonance imaging scans
- A mobile app that takes input from a blood glucose meter and a patient food log to provide insulin dosage recommendations for diabetes
- Software that analyzes a patient’s medical history and diagnostics data to determine the correct drug to prescribe
- Software that performs image processing to detect cancer
- Software that regulates an installed medical device, such as a pacemaker
Examples of technologies that are not considered SaMD, and therefore are not FDA-regulated, include EHRs, software that provides administrative support to health care facilities, and technologies considered low-risk, for example, tools, such as fitness apps, that encourage a healthy lifestyle.
The FDA organizes medical devices into one of three classes based on level of risk, invasiveness, and impact on the patient’s overall health.3 The classes are as follows:
- Class I: Devices that have minimal contact with patients and the lowest impact on patients’ overall health. Class I devices typically do not come in contact with a patient’s internal organs, the central nervous system, or cardiovascular system. Examples include bandages, surgical gauze, oxygen masks, hospital beds, and electric toothbrushes.
- Class II: Devices that present a higher category of risk because they are more likely to come into sustained contact with a patient. This can include devices that come in contact with a patient’s cardiovascular system or internal organs, as well as diagnostic tools. Examples include catheters, syringes, surgical gloves, absorbable sutures, and blood pressure cuffs.
- Class III: Devices that usually sustain or support life, are implanted, or present a potential unreasonable risk of illness or injury. This classification is generally extended to permanent implants, smart medical devices, and life support systems. Examples include breast implants, pacemakers, defibrillators, implanted prosthetics, and cochlear implants.
In 2020, to better align the FDA with the rapidly changing digital health environment, the agency opened the DHCoE.4 The launch of the DHCoE further solidifies the agency’s dedication to the advancement of digital health technology, such as mobile health devices, SaMD, wearables that are used as medical devices, technologies used to study medical products, AI and ML, and more.
The DHCoE serves various stakeholders—such as patients, developers, health care providers, researchers, industry, payors, other government agencies, international regulatory bodies, and other centers within the FDA—and is tasked with empowering digital health stakeholders to advance health care by fostering responsible, high-quality digital health innovation. The DHCoE will evolve through multiple phases to ensure the center can meet its goal of realizing the full potential of digital health.
The following questions and answers are based on an interview with Mr. Patel, the Director of the DHCoE.
Interview with Bakul Patel, Director, FDA Digital Health Center of Excellence
Physicians can place greater trust in devices using digital technology if these devices have received FDA clearance or approval. The FDA maintains a framework for classification of devices in addition to a well-established process for evaluating products. Device classification depends on the manufacturer’s stated intended use of the device, as well as the indications for use. Device classification also is risk-based. This classification system includes the risk the device poses both to the patient and to the clinician as outlined previously.
As part of the FDA’s clearance or approval process, the manufacturer must demonstrate that the device is safe and effective in its intended use. Varying levels of clinical evidence also are required depending on the level of classification. In addition, all devices regulated by the FDA are subject to good manufacturing practice requirements for registration, labeling, and quality.
Of note, some devices qualify for “exempt” status, meaning there is no need to prove safety and effectiveness, nor a need for clinical trials. But if a device receives FDA clearance or approval, that device has typically undergone a rigorous process by which the FDA deems it safe and effective, which can lead to increased physician trust in the device.
Discuss examples of types of digital health devices or services that could benefit surgeons and surgical patients.
An interesting example of a digital health tool for surgeons is the use of an augmented reality (AR) headset that enhances the surgeon’s vision so it is possible to essentially see through a patient’s body during an operation and increase surgical precision—for example, when placing implants or catheters. Some devices that use this AR technology have been cleared by the FDA.
This “computer vision” gives surgeons a three-dimensional view of patient data, allowing physicians to access multiple data streams at once, including discussion with experts, without having to turn their visual attention away from the patient and with the ability to keep their hands free. Similar headsets can be used for virtual reality (VR) surgical simulations to enhance surgical training. Advantages include a wider range of anatomical viewing angles and the ability to complete a VR training simulation as many times as needed.
Beyond these surgery-specific examples, much of the work that is happening today in the digital health space falls under either early prevention or postcare management. On the early prevention side, implantable devices, sensors, and wearable technology have been developed that could generate real-time data about a patient’s health status. These types of devices take some of the guesswork out of patient care.
One example is a knee implant that includes a built-in sensor to track information on load, activity, and fit of the device. Knee implants wear out over time, so this information, which can be transmitted wirelessly, can provide a surgeon with valuable information about the patient’s status. Another example is the use of “smart” stents. Use of cardiovascular stents is a widely used therapeutic approach to treat coronary artery disease. Stents embedded with sensors can monitor and provide real-time feedback on blood flow and could detect re-narrowing of the artery and alert the patient’s health care team prior to the patient experiencing a heart attack.
Postcare management and postsurgery management digital health tools can take a number of forms, such as mobile apps, web applications, and wearable devices for postoperative monitoring after discharge. This type of management can replace some follow-up visits and can be used to monitor the patient remotely for complications.
The FDA continues to pilot and test its Software Precertification (Pre-Cert) Program. How does the FDA plan to involve stakeholders and physician clinical experts in the testing of the pre-certification framework as the administration continues to build and iterate the program?
By way of background, the Pre-Cert Program is a new regulatory model intended to streamline regulatory oversight of SaMD. Instead of evaluating a product itself through the FDA premarket review process, which is expensive and slow, the Pre-Cert Program first evaluates the digital health technology developer or company for a consistent culture of quality, validation, ongoing maintenance, and product postmarket performance established through real-world performance data. Developers or companies that meet these criteria could bring their products to market without FDA premarket review or after a more efficient FDA premarket review. This program is currently in the testing phase.
In the current iteration of the program, the DHCoE is focusing heavily on developing a working model for how to best assess a company that develops digital health tools. The FDA is connecting information that it has gathered over the years to understand how to best evaluate and establish trust in a company through different metrics that may evolve as the program matures and companies’ product development grows.
There are opportunities for specialty societies and physicians to assist the FDA as the administration develops definitions and terms to guide developers and companies who are creating the intended uses for these technologies, to ensure that products are framed in a way that is broadly understood. Before fully launching the Pre-Cert Program, the FDA will begin simulating the program and its processes to accelerate program development and testing. The simulation will be conducted to identify failure points and ensure that the program processes can be trusted.
What is the FDA’s approach to regulating AI and ML and ensuring that medical devices that are based on these algorithms are reliable, valid, transparent, explainable, and free of bias? How does this regulation remain current as the AI- or ML-based tool learns and iterates?
The regulation of AI- and ML-based devices is ideal for the Pre-Cert Program, which provides more streamlined and efficient regulatory oversight and shifts some responsibility to the manufacturer to demonstrate a culture of quality and a commitment to real-world performance monitoring. This monitoring is important because as devices that use AI and ML evolve, real-world information will be reported back to the FDA regarding the product’s safety, effectiveness, and potential risks. The true power of AI- and ML-based software is that it can improve over time instead of remaining static, and the FDA regulatory efforts aim to leverage this opportunity.
The data used to train AI and ML algorithms are important and should be high-quality, diverse, valid, and representative of the uses for which it will be applied. While data used to train the AI- or ML-based tools are important, it is equally important that up-to-date data are used to retrain such tools so that they remain accurate.
The unintended effect of bias in AI is impossible to completely avoid. This bias can play out in many ways, including those that have a racial, gender, age, or socioeconomic impact. One way to address this problem is for the FDA and manufacturers to examine the effects of AI tools for bias using real-world data monitoring and then correct either the algorithm that the tool uses or point out changes needed in the data on which the tool is trained to eliminate or reduce bias.
How can medical societies collaborate with the FDA? Would it be possible to partner with the FDA in some way as context and contextual clinical experts to verify SaMD or software as a service that includes general surgery clinical knowledge?
Physicians and medical societies can contribute to the FDA’s review of digital health technologies in several ways. The FDA is actively seeking the specialty society and physician perspective on what tools and information would be most useful in driving improvements and advancements in clinical care and the format in which the information should be expressed. Using AI/ML as an example, understanding where physicians see the benefits of this technology in their practices is crucial to help build trust in the capabilities of the technology, leading to broader use.
Feedback about why physicians decide not to use or do not trust some digital health technologies also would be useful as the FDA approves and clears these products. This feedback is crucial when the FDA evaluates the benefits and risks of these nontraditional digital health tools, because they are not as widely accepted and understood by physicians, compared with devices such as implants that have been used for many years. The FDA would use this feedback to continue investigation and research to create more practical and useful digital health tools.
In addition to using feedback mechanisms to collaborate with the FDA, specialty societies should continue to engage and educate their membership about the roles and responsibilities of the FDA and the DHCoE. Education and engagement with the FDA will help drive a greater understanding of how new digital health tools have the potential to improve many parts of care and improve efficiency and care coordination.
Aside from the goals that the FDA has set for the work to be done under the DHCoE, do you have additional personal goals or initiatives you hope to take on while serving as the director?
The DHCoE is a great start in establishing collaborative efforts across the FDA, and it would be very rewarding to see the goals of the DHCoE sustained over a long period of time. Digital health is on the verge of changing patient care in a unique way, and it would be very personally satisfying to see the DHCoE move the needle in the digital health space.
It has been my personal goal to prepare the FDA for the future and to not only regulate from a “‘defensive” stance, but also to shape the digital health landscape based on what the FDA believes the future of digital health tools should look like. This includes continuing to grow the DHCoE staff with professionals with appropriate training and expertise and ensuring that the FDA has the proper processes in place to adapt to ongoing technological advances.
This transition within the FDA is a journey and continuing to develop processes and pre-certifications will help drive the development of better solutions in the future.
- U.S. Food and Drug Administration. The Food, Drugs, and Cosmetic Act. 21 U.S. Code § 321(h). Available at: www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?CFRPart=1&showFR=1. Accessed April 12, 2021.
- Spanou D. International Medical Device Regulators Forum: Software as a Medical Device (SaMD): Key definitions. Available at: www.imdrf.org/docs/imdrf/final/technical/imdrf-tech-131209-samd-key-definitions-140901.pdf. December 9, 2013. Accessed April 7, 2021.
- U.S. Food and Drug Administration. Overview of medical device classification and reclassification. Available at: www.fda.gov/about-fda/cdrh-transparency/overview-medical-device-classification-and-reclassification. Accessed April 7, 2021.
- U.S. Food and Drug Administration. Digital Health Center of Excellence. Available at: www.fda.gov/medical-devices/digital-health-center-excellence. Accessed April 12, 2021.