What happened in health care technology this week, and why it’s important.
FDA releases ‘guiding principles’ for AI/ML device development
The U.S. Food and Drug Administration released a list of “guiding principles” this week aimed at helping promote the safe and effective development of medical devices that use artificial intelligence and machine learning. As reported by Kat Jercich in HealthcareIT News, the ten guiding principles identify points at which international standards organizations and other collaborative bodies, including the International Medical Device Regulators Forum, could work to advance Good Machine Learning Practices (GMLP).
Why it’s important – The agency says stakeholders can use the principles to tailor and adopt good practices from other sectors to be used in the health tech sector, as well as to create new specific methods. Having these baseline ten principles will help to inform regulatory practices in the future.
User-Centered Design of Companion Robot Pets Involving Care Home Resident-Robot Interactions and Focus Groups With Residents, Staff, and Family: Qualitative Study
Companion robots, such as Paro, demonstrate strong potential for helping reduce this pressure through reported benefits, including reduced agitation, depression, loneliness, care provider burden, and medication use. But there have been few studies undertaken to look at the design of these robots. A recent study published in JMIR aimed to provide user-centered insights into the design of robot pets from critical stakeholders to inform future robot development and the choice of robots for real-world implementation and research.
They found that care home residents, family members, and staff were open and accepting of the use of companion robot pets, with the majority suggesting that they would keep a device for themselves or the residents. The most preferred device was the Joy for All cat, followed by the Joy for All dog. In discussions, the preferred design features included familiar animal embodiment (domestic pet), soft fur, interactivity, big appealing eyes, simulated breathing, and movements. Unfamiliar devices were more often seen as toy-like and suitable for children, producing some negative responses.
Why it’s important – This is the first study to focus on the end-user and their caregivers’ insights into which design elements will support the adoption and regular use of companion robots. The results have implications for future robot designs and the selection of robot pets for both research and real-world implementations.
The unexpected health impacts of wearable tech
We didn’t realize how much doctors would see of our health data and whether or not the information would help treat and manage chronic conditions. But, as Nicole Wetsman reports in The Verge, people were using the devices to monitor their health — tracking their heart rate, steps, and sleep. And gradually, more and more of them started to bring that information along to their doctors’ appointments. The permeation through healthcare is particularly noticeable in three areas: cardiology, sleep medicine, and sports medicine.
Why it’s important – The increase in the use of patient-generated health data (PGHD) has been hotly debated in health care circles for several years now. Depending upon the clinical use case, and as has been described in the article, patients and their care teams can benefit from the real-time monitoring and reporting of data. This accomplishes three things. First, it tells the provider more about chronic disease self-management in everyday life. Second, it holds the patient accountable for that self-management. And finally, it can notify a clinician when a patient’s disease state becomes out of control and spur intervention. PGHD use is still limited, but the data suggest that when patients collect and providers view PGHD, it can positively impact patient health.
A smart knee implant promises to ‘help write the future of orthopedic technology.’ Surgeons aren’t so sure
Mario Aguilar, Health Tech Correspondent at STAT, reviews a new, souped-up knee implant developed by Zimmer Biomet as a way to passively collect data about recovery after one of medicine’s priciest and most common procedures. The implant — cleared by the Food and Drug Administration in August for use in a small subset of knee replacements — contains sensors, a wireless transmitter, and a pacemaker-like battery that could paint a far more precise picture of the recovery process problems that arise. The company has called it “groundbreaking” and claims it will “help write the future of orthopedic technology.” But the surgeons who will need to embrace the implant caution that while the device has potential, insights are likely far off — if the data turns out to be helpful at all.
“Technology has to be proven that it’s going to improve outcomes in order to be used. So, you know, even though this sounds like a cool idea … this isn’t going to improve our outcomes.”Calin Moucha, Chief Joint Replacement Surgeon, Mount Sinai Health System, NY
Why it’s important – The quote above says it all. In joint replacement, novel technology usually isn’t favored over established implants with years of positive results. Some surgeons agree that such information might one day help identify patients whose implants had loosened or who required attention that couldn’t be detected with routine X-rays or changes in symptoms. But surgeons still have concerns about the utility of the data Zimmer Biomet is collecting as it goes to market with the technology. Most agree that mobility data is “interesting from a biomedical science standpoint and understanding the function of these implants,” but are adamant that it’s no substitute for hearing from a patient.
FDA Provides New Draft Guidance on Premarket Submissions for Device Software Functions
More news from the FDA this week. The FDA is making available the draft guidance Content of Premarket Submissions for Device Software Functions intended to provide information regarding the recommended documentation to include in premarket submissions for the FDA to evaluate the safety and effectiveness of device software functions. The proposed recommendations in this draft guidance document pertain to device software functions, including both software in a medical device (SiMD) and software as a medical device (SaMD), and describe a subset of information that would be typically generated and documented during software design, development, verification, and validation.
“As technology continues to advance all facets of health care, software has become an important part of many products and is integrated widely into medical devices. The FDA recognizes this evolving landscape and seeks to provide our latest thinking on regulatory considerations for device software functions that is aligned with current standards and best practices.”Bakul Patel, Director, FDA Digital Health Center of Excellence in the Center for Devices and Radiological Health
Why it’s important – For software developers, this guidance will eliminate any confusion in the submission process and represents a step forward in the regulation of medical software products. The FDA is requesting comments on the draft now. When final, this guidance will replace the FDA’s Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices issued on May 11, 2005. It will update the FDA’s recommendations on the appropriate documentation for the review of device software functions in premarket submissions.
Alphabet has a new drug discovery company building on DeepMind’s AI chops
Katie Palmer in Stat+ (subscription required) reported on the launch of Isomorphic Laboratories. This new Alphabet company aims to leapfrog the success of the protein-folding work to apply deep learning methods to drug discovery. Isomorphic will focus on building predictive or generative models of biological phenomena, using computers to anticipate how drugs will perform and potentially design novel molecules. Rather than developing its own pipeline of drug candidates, the company may aim to sell its models platform as a service.
Why it’s important – DeepMind can leverage the success of its protein folding work in predicting protein structure with its deep learning model, AlphaFold2. The new company could focus upon protein-protein interactions, small molecule design, binding affinity, and toxicity analysis as potential targets for predictive models. Isomorphic’s most immediate task will be in staffing up a multidisciplinary group of deep learning experts, computational biologists, medicinal chemists, biophysicists, and engineers.