Developments in Artificial Intelligence and Machine Learning (AI/ML) are rapidly creating competitive advantages for established and emerging companies across many industries. The healthcare sector is no exception, as health services, biopharmaceuticals and medical device firms are enjoying the benefits and experiencing the disruption of AI/ML. In fact, Statista projects that the market for AI in healthcare will climb from just over one billion in 2017 to more than $28 billion U.S. dollars by 2025.

Our clients rely on IMS and our extensive network of best-in-class experts to provide them the foremost experts to consult, opine and often testify regarding emerging technologies, policies, and areas of disruption in critical strategic projects and complex commercial litigation. Stu Lipoff, a sought-after expert in healthcare data, cybersecurity and data privacy has been relied upon by IMS clients for many high-profile cases, notes that the introduction of Software as a Medical Device (SaMD) and AI/ML in the space are leading to complex questions.

“AI-based systems are learning devices,” states Lipoff. He explains that in initial product releases, such systems, “are often not as smart as they need to be until they have been in use for some time and then ‘learn’ from their mistakes.” This allows AI/ML devices to adapt learnings into new AI rules. In a typical architecture supporting AI devices, all devices in the field can upload individual experiences into the cloud, and the original equipment manufacturer (OEM) would analyze the upload and then download a new set of AI expert system rules.

Lipoff anticipates numerous challenges and disputes as the U.S. Food and Drug Administration’s guidance and industry standards around AI/ML applications in the space evolve. “Should an end user who suffers a medical issue have cause of action against an OEM because their early generation AI based device lacks the benefit of future learnings?” If so, Lipoff notes, OEMs and manufacturers would have complex considerations for when products could or should be released into production and distribution.

Dr. W. G. “Bill” Bysinger, another prominent expert and published author depended upon by IMS clients, observes that the integration of AI/ML technology into healthcare and medical devices invites many questions, including: “When an AI algorithm ‘learns’ during use with patients, is it reasonable to require the manufacturer to clear (through FDA procedures) each incremental adaptation in the algorithm?”

As a recognized industry authority in areas including healthcare software patents, healthcare security and HIPAA, and healthcare and medical technology, Bysinger also sees other challenges ahead. “The issue with AI and Machine learning is not the technology implementation, but the application of the proposed model for making decisions based on good analysis versus existing decisions supported without AI and ML implementation.”

Following the FDA’s launch of a 2019 pilot and release of its discussion paper, Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device, the agency received more than one hundred comments from industry associations, biopharmaceutical and medical device manufacturers, and other interested parties around SaMD.

Bysinger notes that the barriers to market entry will increase in the space, as will the cost of devices deploying AI/ML capabilities. “This oversight and regulatory approach by the FDA regarding AI/ML will impact small innovative developers from entering the market,” observes Bysinger.

Some of the current FDA guidance around AI/ML SaMDs, Bysinger states, could be subjective and create barriers to entry for smaller innovative companies, especially related to the assessment of developers for the presence of a “high-quality culture.” Meeting these criteria in the FDA’s streamlined review process could result in approval for marketing and make the product subject to ongoing real-world performance evaluations.

“This can be a very subjective measure,” adds Bysinger. “AI and Machine learning development is a highly technical area involving engineering and social disciplines differing from software and hardware development.” As a result, Bysinger believes that the space could be difficult to culturally evaluate.

As consumers and companies become increasingly reliant on AI and Machine Learning across industries and applications and the FDA’s 2019 pilot unfolds, the need to help clients navigate complex challenges will evolve. IMS ExpertServices is committed to empowering our clients’ efforts to stay abreast of new innovations, developments, and challenges, and honored to be trusted as the consultative expert services firm for the world’s most influential litigators.