By Alyx Arnett
Software is no longer just a tool in healthcare—it’s becoming a medical device in its own right. Known as software as a medical device, or SaMD, this technology is playing a growing role in diagnosing and treating conditions, with more companies, including those in sleep medicine, earning US Food and Drug Administration (FDA) clearances.
But what exactly is SaMD?
Simply put, it’s software designed to perform medical functions independently—without being part of a physical medical device.
The International Medical Device Regulators Forum (IMDRF) defined the term in 2013 after regulators recognized the need for a unified framework and principles to ensure innovation and patient safety in the growing SaMD space.1
Chaired by the FDA, the SaMD Working Group agreed upon key definitions for SaMD, a framework for risk categorization, a quality management system, and guidance for clinical evaluation. The FDA continues to review and refine these documents “to ensure ongoing consistency, predictability, transparency, and quality of premarket regulatory programs and criteria for assessing premarket technical documentation,” according to a spokesperson with the FDA.
The Path to FDA Clearance
Marketing a SaMD is similar to marketing any other medical device, as far as providing clinical trial evidence that shows safety and efficacy. However, Jenna Carl, PhD, chief medical officer at BigHealth, maker of SleepioRx, SaMD for insomnia treatment, notes that it also requires meeting specific quality management standards.
“These are both things on the software quality side to make sure that it is well-tested and doing what it says it’s doing, but it’s also cybersecurity and other privacy aspects to make sure that patients using the software are fully protected,” Carl says.
The IMDRF quality management system document spells out clinical and technical considerations.2 “Good software quality and engineering practices need to be incorporated into the device’s quality management system to ensure medical device quality management principles are met,” says a spokesperson with the FDA.
510(k) clearance is the most common pathway for marketing authorizations among SaMDs, followed by De Novos and some Premarket Approval applications. “The breadth of applications has continually and rapidly increased in the last few years and is not predicted to decelerate in the near future,” says the FDA spokesperson.
SaMD for Sleep Analysis
Several companies have received FDA clearances for SaMDs for sleep analysis. One of the earliest examples is Natus Medical’s SleepWorks PSG software, cleared for marketing by the FDA in 2011. SleepWorks works with amplifiers used in polysomnography (PSG) to analyze physiological signals to help diagnose sleep disorders.
SleepWorks PSG SaMD, cleared in 2011, allows recording, displaying, analysis, printing, and storage of physiological signals to assist in diagnosing sleep disorders. Photo credit: Natus Medical
SleepWorks requires hardware to collect patient data, as SaMDs do for sleep analysis. Anthony Palone, district sales manager at Natus, explains that the FDA clearance, however, is around how the data is reviewed. It’s classified under FDA product code OLZ for the software’s ability to automatically detect events during PSG. The hardware simply captures patient data before the software analyzes it.
Another FDA-cleared SaMD falling under this code is EnsoData’s EnsoSleep, which uses artificial intelligence (AI) to automatically score sleep study results to aid in diagnosing sleep disorders.
EnsoSleep, an AI-assisted sleep scoring and analysis solution, automates the scoring of studies. Photo credit: EnsoData
Sleep software manufacturers have also received clearance for SaMDs in the product category of breathing frequency monitors, consistent with hardware sleep diagnostic equipment.
These include EnsoData’s EnsoSleep PPG—which received clearance earlier this year to automatically analyze photoplethysmogram data collected during sleep using FDA-cleared pulse oximeters—and SleepImage’s SleepImage System, cleared in 2019 to evaluate sleep disorders, as well as for diagnosis and management of sleep-disordered breathing in children and adults using PPG data from pulse oximeters.
SaMD for Insomnia Treatment and Sleep Apnea Detection
SaMDs also exist for insomnia treatment and sleep apnea detection. Two solutions have received FDA clearance for the digital treatment of insomnia: Nox Medical’s Somryst, cleared in March 2020, and BigHealth’s SleepioRx, cleared in August 2024.
Both Somryst and SleepioRx deliver cognitive behavioral therapy for insomnia (CBT-I) through prescription-only software-based mobile apps downloaded onto a smartphone.
Somryst, a prescription-only SaMD, provides CBT-I to patients 22 years of age and older with chronic insomnia. Photo credit: Nox Medical
Additionally, the FDA recently cleared Apple’s sleep apnea notification feature, an over-the-counter software-only mobile medical application that analyzes Apple Watch sensor data to identify patterns of breathing disturbances suggestive of moderate to severe sleep apnea and notify the user.
Previously, a De Novo was granted to Samsung for its sleep apnea feature.
Benefits of SaMD in Sleep Medicine
SaMDs in sleep medicine can expand diagnostic flexibility and accuracy, reduce costs, and improve access.
Bogi Palsson, CEO of SleepImage, highlights that the SleepImage System is also indicated to establish sleep quality. This enables licensed healthcare providers to offer a sleep test to patients with a range of sleep-related complaints allowing for conditions like insomnia, for instance, to be evaluated. “And I think that the SaMD clearance makes that multi-functionality of the medical device possible,” Palsson says.
SleepImage SaMD establishes sleep quality and analyzes electrocardiogram or plethysmogram data to aid in evaluating sleep disorders. Photo credit: SleepImage
Another advantage of SaMD is the ability to conduct repeated studies at a lower cost than traditional hardware-based medical devices with many consumables, says Palsson, making more affordable longitudinal monitoring possible.
Serina Fischer, product strategy lead of orexin at biopharmaceutical company Takeda, which is focused on addressing rare sleep disorders such as narcolepsy and idiopathic hypersomnia, says SaMDs that can score PSG recordings and differentiate narcolepsy type 1, say, from other sleep-wake disorders play a critical role supporting patients in receiving a more accurate, timely diagnosis.
Such technology also helps patients get treated faster, according to Chris Fernandez, chief research officer and co-founder of EnsoData. Infrastructure pieces like AI accuracy, he says, contribute to saving clinician time across diagnostic and treatment workflows.
EnsoSleep PPG analyzes, displays, and summarizes photoplethysmogram data collected during sleep using FDA-cleared pulse oximeters. Photo credit: EnsoData
Patient access was a major driver behind the development of Somryst and SleepioRx. There is a recognized shortage of practitioners trained in delivering CBT-I which has limited access to this firstline therapy for insomnia, notes Heidi Riney, MD, chief medical officer at Nox Health.
“Digital CBT-I solutions effectively bridge this gap by offering accessible, scalable, and cost-effective alternatives,” says Riney.
SaMD Challenges: Compatibility, Cybersecurity, and Continuous Updates
SaMD also introduces unique challenges. For one, SaMD must work across different hardware platforms.
“Being a software-only product in a field that relies heavily on physical medical devices and sensor data requires software to integrate with a very large number of tools and solutions,” says EnsoData’s Fernandez. “So establishing a level of integration and interoperability such that we could operate as a software-only tool and be utilized with the full spectrum of diagnostic devices was a big technical hurdle.”
Also, users are responsible for installing updates, which can result in different versions existing on the market. It can also open the software up to cyberattacks, says Nidhi Gani, medical device cybersecurity fellow at Archimedes Center for Health Care and medical device cybersecurity adjunct professor at Northeastern University. “If these software are not updated on time, they are vulnerable to attacks, which is a direct impact to patient safety,” she says. “Patches and updates are critical to the maintenance of the software.”
As cyberattacks become increasingly sophisticated, managing data security and patient privacy is becoming more complex and challenging, says Sveinbjorn Hoskuldsson, chief technology officer and co-founder of Nox Medical. “It is essential for customers choosing clinical software solutions to not only confirm that the product is a regulated SaMD but also ensure it is operated by a provider with a strong cybersecurity framework,” Hoskuldsson says.
Unlike traditional medical devices, SaMD can also change over time as manufacturers introduce continual updates. This presents further challenges for manufacturers, requiring them to ensure the safety and reliability of their products throughout the software’s lifecycle.
The IMDRF’s risk categorization guidance advises that when manufacturers modify SaMD in a way that alters its definition statement that it should be reevaluated.3
AI’s Growing Role in SaMD in Sleep Medicine
As the role of SaMD in sleep medicine grows, experts agree that AI and machine learning (ML) will be at the forefront of future advancements.
BigHealth’s Carl believes AI will eventually reshape how insomnia is managed. While it’s possible for AI to deliver CBT-I, she notes that it first needs to be proven safe, and clinicians must reach a level of comfort with using it.
SleepioRX is a prescription SaMD delivering CBT-I via a smartphone app. Photo credit: BigHealth
Natus’ Palone envisions AI-powered tools that will be able to pull in data from multiple sources, not just amplifiers or home sleep tests, and aggregate that information to create a more complete picture of the patient.
“AI is definitely going to shape all of the software-based devices from the ones that do assessment to the ones do treatment,” says Carl. “I think it’s just a question of what that ends up looking like. What are the guardrails? How does FDA ensure safety in that respect? And what ends up being the actual role of the AI? That’s clearly going to take years to get sorted out.”
The FDA recognizes that its “traditional paradigm of medical device regulation was not designed for adaptive AI/ML,” according to a spokesperson with the FDA. “The application of AI-based technology into many different clinical areas, combined with the unique nature of clinical medical data—for example, low prevalence of disease and lack of or difficulty in obtaining truth data—produces many challenges in developing robust evaluation methods and understanding the effects of AI-based technology in real-world settings,” says the spokesperson.
Additionally, the ability of AI-enabled medical devices to continuously learn, update, and adapt based on the availability of more data or to respond to changes in the data also presents unique regulatory challenges.
The FDA published draft guidance on marketing submission recommendations for AI/ML-enabled in March 2023. The FDA spokesperson notes that the agency has been considering the needs of such devices to ensure they “will remain reasonably safe and effective over their use life and to promote high-quality innovations.”
Further Advancements
As far as how SaMD may evolve, EnsoData’s Fernandez expects increased connectivity between diagnostic and treatment tools. This may come in the form of SaMDs that can provide more data on what treatments or modalities are likely the most effective for a patient.
He believes this type of connectivity will involve the broader ecosystem of SaMD and non-SaMD tools that can work together to integrate new kinds of applications. “We continue to see innovative sensors and wearable devices come out that enable more longitudinal patient care and management,” Fernandez says. “A lot of sleep testing and sleep medicine today is still based on closer to a night of data than a longitudinal view on the health trajectory, and I think the increased availability of SaMDs and devices together will result in this whole new perspective of longitudinal data be much more widespread and utilized than it is today.”
Takeda’s Fischer sees “immense potential” in SaMD, particularly in improving diagnosis for conditions like narcolepsy. She highlights the potential of at-home sleep tests to overcome current limitations and reduce the burden on patients, caregivers, and providers. “Software-based solutions in the form of digital companions that offer personalized support could assist patients and their providers in disease management and facilitating data-based treatment decision-making, which is very exciting,” she says.
For SleepImage’s Palsson, he expects SaMD will become regarded as less of a differentiator from the patient and clinician’s perspective. “I think they will expect a good user experience from SaMD, but most importantly, they will require accurate and actionable results from the test,” Palsson says.
References
- US Food and Drug Administration. International Medical Device Regulators Forum (IMDRF). Available at https://www.fda.gov/medical-devices/cdrh-international-affairs/international-medical-device-regulators-forum-imdrf.
- IMDRF SaMD Working Group. Software as a Medical Device (SaMD): Application of Quality Management System. 2015 Oct 2. Available at https://www.imdrf.org/sites/default/files/docs/imdrf/final/technical/imdrf-tech-151002-samd-qms.pdf.
- IMDRF SaMD Working Group. “Software as a Medical Device”: Possible Framework for Risk Categorization and Corresponding Considerations. 2014 Sept 18. Available at https://www.imdrf.org/sites/default/files/docs/imdrf/final/technical/imdrf-tech-140918-samd-framework-risk-categorization-141013.pdf.
ID 330181056 © AminaDesign | Dreamstime.com
