The medical industry, and in particular the field of diagnostic devices, has become fertile territory for artificial intelligence (AI)- and machine learning (ML)-based technologies. There is no doubt that the development of these devices has the potential to transform the healthcare field altogether. But making an accurate decision in a complex medical situation is not always so straightforward, thus giving pause to even the most optimistic expert.
In one scenario, set in the not-so-distant future, a lengthy and uncomfortable doctor visit is replaced by a quick glimpse of your own wristwatch. For many, these sorts of high expectations are warranted because of the grand promise of AI/ML-based medical devices. Employing their powerful algorithms, these devices have the potential to continuously adapt to changes in the patient, thus learning how to make accurate medical diagnoses not unlike a licensed physician would.
Indeed, the presence of AI technology can already be seen at work in the medical field. Whether providing patients with advanced prevention indicators, increasing medicine adherence, or delivering more helpful insights into healthcare issues, fast-learning AI has already started leaving its stamp on hospitals around the country.
Despite all this promise, however, the obstacles for full implementation of AI/ML in daily clinical practice are numerous. One of the biggest obstacles will be assuring the "cleanliness" of the data that will essentially be driving these new medical devices.
Speaking to this particular challenge, Bakul Patel, director of the FDA's Division of Digital Health elaborated, "We don't want to set up a system where we figure out, after the product is out in the market, that it is missing a certain type of population, or demographic, or other aspect that we have accidentally not realized."
To help with this issue, Patel stressed the need of maintaining full transparency, while also reaching a delicate balance between empowering innovators and protecting patients from algorithms that will be getting smarter and better trained by the day.
In addition, a crucial element shaping these obstacles is regulation. Currently, the FDA has "so many questions" about what good practices look like for algorithm design, development, training, and testing. But according to Patel, his agency is now considering a total product lifecycle-based approach to regulating medical devices that leverage self-updating algorithms. "We'll probably have a multi-pronged approach come out soon and we're working towards sharing that."
Ultimately, despite the significant challenges to adopting AI/ML-based medical devices, their use promises to revolutionize our healthcare system. And with a shortage of healthcare workers, some might argue that the use of AI-powered devices cannot get here soon enough.
