When Bakul Patel started as a policy advisor in the US Food and Drug Administration in 2008, he could pretty much pinpoint when a product was going to land in front of the reviewers in his division. Back when medical devices were heavy on the hardware—your pacemakers and your IUDs—it would take manufacturers years to get them ready for regulatory approval. FDA reviewers could keep up pretty well.
But as computer code took on more complex tasks, like spotting specious moles and quantifying blood flow, their duties began to accelerate. Software developers needed months, not years, to make it to the market. And there were a lot of them. It got harder to match pace. And then came artificial intelligence.
Today, machine learning powers more and more medical device software. And because it is always learning and improving, it is constantly changing products on the fly. For most regulators, an ever-changing algorithm is their worst nightmare. But Patel is one of those rare Washington bureaucrats who’s also a fervently optimistic futurist. And he’s got big plans to get federal regulators off Washington time and up to Silicon Valley speeds.
To do that, the FDA is creating a new unit dedicated strictly to digital health. Patel will be hiring 13 engineers—software developers, AI experts, cloud computing whizzes—to prepare his agency to regulate a future in which health care is increasingly mediated by machines. (He’s using funds generated by the medical device division’s user fee system, which is FDA’s only other revenue stream besides congressional appropriations.) He’s also got plans to reimagine the path these machines will take to regulatory approval.
For technology giants getting into the health care game, the timing couldn’t be better. Last year Google’s venture capital arm (which manages around $2.4 billion) spent one-third of its investments in the health care space. Its spinoffs Calico and Verily are pursuing ambitious projects like smart contact lenses, Project Baseline—oh, and beating death. Apple, in addition to its wellness through wearables play, is already working closely with the FDA on an app to diagnose Parkinson’s. And IBM is employing its artificial intelligence engine, Watson, to do everything from treating cancers to discovering new drugs.
Over the last year, FDA has put out a number of documents describing the agency’s current thinking on digital health. These guidances help developers understand what FDA does and what it doesn’t regulate as a medical device, and they reflect a largely hands-off approach. The FDA focuses its limited resources mostly on high-risk products, and the most recent of its proposed rules addresses software as a medical device—a category that would include medical apps, which remain largely unregulated.
As Patel, now the associate center director for digital health at FDA, was digging through the guidance’s 1,400 comments, he had a lightbulb moment. “We’ve been trying to translate the current regulation paradigm for digital,” he says. “But what we have today and what we’re going to have tomorrow are not really translatable. We need to take the blinders off, start with a clean sheet of paper.”
Rather than reviewing each line of code or medical device on its own merits for each of its intended uses, Patel wants to flip that framework on its head. Instead, he envisions a model something more like the TSA security line at the airport: New developers or manufacturers with spotty track records would still have to take off their shoes and go through the body scanner. But trusted companies with demonstrated histories of excellence could keep their footwear and stroll through the metal detector. Patel’s not yet sure exactly how it would work, but it’s one of the ideas he’s toying with and running by industry stakeholders. “The idea is to get safe products to market faster, by having people compete on excellence rather than compliance,” he says. The trick is not to get bogged down by stuff you’ve never seen before.
“We’re headed toward a zero code world, where AI writes it for you or you just say what you want and natural language processing takes care of the rest,” says Patel. “The pace will be tremendously faster than what we’re seeing today. The question is, how do we align our regulations to that radically different development timeline?”
Whatever the approval process winds up looking like, it will fall to the new digital health unit to oversee and coordinate it between different offices within the FDA. Right now, those efforts are fragmented by specialty. Experts in the cardiology group would be in charge of a device that uses AI to quantify blood flowing through the heart, while a radiology group would review an AI-powered MRI-reading software. Breaking out of these silos is the goal for the new group, since the same technologies will increasingly apply to products that cut across specialties.
Amidst the HHS-wide hiring freeze, Patel’s been given the go-ahead to start bringing new people into key positions. More jobs will be announced in October, when the new Medical Device User Fee Amendment goes into effect. Every four years FDA renegotiates this agreement with the industry, and the latest one established the formation of Patel’s digital health unit. The user fees will give him the money to make his hires, but in the meantime he’s already begun a recruiting campaign in high-tech hotbeds like Silicon Valley, Seattle, and Boston. His sales pitch should be familiar to tech types: the usual shaping the future, making the world a better place start-up spiel. The only difference being Patel’s start-up is inside the federal government.
The question is whether or not talented people will be willing to leave their corporate gigs (and corporate salaries) for a stint in Washington under the new administration. For the last eight years, talent flowed pretty freely between Obama’s federal government and places like Google, Facebook, and Apple. A suddenly conservative, anti-science cloud cast over Washington has since cooled that jetstream. But if the digital information overlords know what’s good for them, they’ll get someone onto Patel’s team. Because it’s not every day a top regulator decides to wipe the slate clean. And it’s not every day you get to write the rules that will rule what tomorrow’s machines will write.