Do you think of Arnold Schwarzenegger in the film “Terminator” whenever we talk about “artificial intelligence”?!
It really doesn’t have to be scary. The intersection of artificial intelligence (“AI”) and medical devices has ushered in a new era of healthcare innovation. AI-driven devices incorporate Machine Learning-Decision Support Functions (“ML-DSFs”) to revolutionize diagnostics, treatment, and patient outcomes. (More on this later.)
However, as the landscape evolves, so do the regulatory considerations. After all, we all realize that medical products – even revolutionary ones – can’t be sold without the review and approval by the U.S. Food and Drug Administration and the FDA has been working on this. The FDA’s recent draft of marketing submission recommendations for updates to ML-DSFs hold significant implications for companies in the medical device sector.
The FDA already plays a pivotal role in ensuring the safety and effectiveness of AI-integrated medical devices. The agency evaluates AI-based medical devices based on their intended use and risk classification, ranging from Class I devices with the lowest risk to Class III devices with the highest risk. The FDA’s current oversight process provides a regulatory framework that accommodates the development, testing and marketing of AI-integrated medical devices. It even maintains a list of approved AI/ML-Enabled Medical Devices to highlight FDA’s role in propelling innovation within the industry.
But the FDA also recognizes that AI technology continues to advance, and it has worked with stakeholders to develop draft submission guidelines that balance the benefit of improved software tools with the need to maintain the safety and effectiveness of devices as they undergo modifications.
It’s important to note that these guidelines are currently in draft form and as such are not finalized regulations and are not binding on the FDA or the industry. But it is good and very important that the regulators and the innovators are talking to each other. In the FDA’s words, “guidance describe the Agency’s current thinking on a topic and should be viewed only as recommendations,” but they nevertheless are a good indicator of the direction towards which final regulations will point.
The guidelines lay out the structure of a “Predetermined Change Control Plan” (“PCCP”)(the Government really likes acronyms!), a submission that would enable future planned modifications to be implemented without the need for new review submissions, provided they adhere to the authorized plan. Why is this important, you might ask? Well, this approach encourages innovation by allowing companies to improve their devices over time based on a single approval.
In the conventional case, changes to sophisticated medical devices must be separately tested, validated and then approved by the FDA. Without this change in perspective, no AI-integrated medical device could use machine learning to refine its protocol without a new filing with FDA. That sure doesn’t work.
Naturally, with the flexibility of updates comes the challenge of ensuring continued safety. Deviations from the authorized PCCP could lead to unintended consequences, potentially impacting patient safety. Transparency is essential. Clear documentation of how an AI algorithm is developed, validated, and updated fosters accountability.
Consider a wearable heart rate monitor that utilizes AI to provide real-time cardiac health insights. The device is initially designed to monitor heart rates in the general adult population. However, new research indicates that the device’s algorithm can also be effective in detecting heart arrhythmias in pediatric patients. The manufacturer might decide to update the ML-DSF to include a modification specifically tailored for pediatric patients. The PCCP would detail the data collection, re-training and performance evaluation methods to ensure that the device accurately detects arrhythmias in the new pediatric subpopulation without negatively impacting its utility for the existing adult users. The update procedures would include clear communication to users about the device’s expanded capabilities and any potential changes in usage guidelines.
Medical device companies embracing AI technology must navigate the intricate regulatory landscape while harnessing the potential of innovation. The FDA’s recommendations give companies a means to better ensure the safety, effectiveness and compliance of their AI-integrated devices. Early engagement with the FDA and collaboration with regulatory experts can provide invaluable insights into how to meet the evolving regulatory requirements.
At Crowley Law, we understand the intricate legal nuances surrounding medical device regulations. If you have any questions, concerns, or seek expert legal guidance on matters related to ML-DSFs or the medical device approval process generally, our dedicated team is here to assist you. Reach out to us today at 844-256-5891 or [email protected] to schedule a consultation. Let us help you navigate the path toward innovation with confidence and compliance.