Get access to Protocol
‘Everyone is looking at health care differently’: How medicine will evolve after the pandemic
David Stavens, co-founder of teleradiology startup Nines, on the post-COVID health care landscape.
Medical triage was ripe for technological disruption before coronavirus, but the global pandemic has turned an opportunity into an urgent need virtually overnight. Hospitals suddenly facing a flood of patients with a dizzying array of risk factors and complications from COVID-19 are scrambling to prioritize cases and tailor care, and some have turned to AI and machine learning for help.
Nines, a Silicon Valley teleradiology startup co-founded in 2017 by former Udacity CEO David Stavens and Dr. Alexander Kagen, site chair of radiology at Mount Sinai West and Mount Sinai Morningside Hospitals in New York, has found itself on the front lines.
Nines radiologists read scans for hospitals remotely and deliver the results over the internet — a naturally socially distanced supplement to in-house expertise. On April 21, the company received FDA approval for Nines AI, a technology that helps rapidly identify two critical emergency conditions in CT scans of the brain: mass effect (where tumors or other masses create pressure on the brain) and intracranial hemorrhage. The software can flag those conditions in under a minute, so those scans can be read immediately and patients can be prioritized. It's a potentially life-saving innovation for coronavirus patients (and others) suffering from brain hemorrhages.
Protocol recently spoke with Stavens, an expert in machine learning who co-founded Stanford's seminal self-driving-car team in 2005, about how COVID-19 has changed the terrain and the outlook for health care AI.
This interview has been edited for length and clarity.
Do you remember the moment you realized that the pandemic was going to shift the health care landscape?
March 3, the last day we were in the office normally, or somewhere around then. I was doing an interview, not unlike this, but in person with a video camera. It had been scheduled. And I woke up to a text from Maureen Bradford, our VP of marketing, saying, "Are you still planning to come in for the interview today?" And I thought this was somewhat strange, because why wouldn't I be planning to come in for the interview? And then, as I saw the news, I realized that much of the Bay Area was starting to shut down, and big tech companies were sending tens of thousands of employees home, and we ultimately that day sent all of our employees home. So I remember that as kind of the day that everything changed.
When did you realize the implications for Nines?
I think it took a few weeks to set in. The pandemic has brought health care front and center, but the mission of Nines, to improve radiology and make it more effective and more affordable while maintaining quality, has kind of always been there. So I don't like to frame it either in my own mind, to the company or to our customers as: "This terrible thing is happening in the world, but it's sort of good for us." I think we have built this opportunity, and we're bringing it to market and, thankfully, now is the time where we can be particularly effective.
When you started Nines in 2017, what led you to think it was a good moment for something like this?
It was largely a personal motivation. I'm an only child, my parents are in their late 70s, and although they enjoy good health in many respects, they've had many touch points with the health care system. While they've been fortunate to receive excellent care, it was clear to me as a technologist and tech entrepreneur that technology and health care could be improved. There was more opportunity to do more.
I think also a driving force around that time was that AI was, and still is, going through a bit of a renaissance with the advent of deep neural networks and all of the problems that they could unlock.
Before coronavirus, what were the main obstacles you thought you would have to surmount?
I think whenever you bring a new technology into an industry, whether it be in radiology or in education or in transportation, I think people rightly say, "Hey, yes, our industry could always be better, but we're actually doing quite well now. And how do we know that your technology is actually going to make things better or not?" The burden is on the innovator to show its quality and efficacy. And I think that, particularly in health care, that bar is often very high. I knew we could hit that bar, but there's always risk around being able to hit the bar and then being able to communicate persuasively that you have.
And now, after COVID?
I think that COVID changes everything. Hospitals are thinking about things differently, thinking about their priorities differently. It is the case that hospitals now are expected to do even more than ever, so the founding thesis of Nines — let's make radiologists even higher quality, even more efficient, so they can do more as they are under strain — that's even more true now, and the need is even greater now. But certainly, everyone is looking at health care differently.
I think that COVID has made people more open-minded. I don't think it has lowered standards. But I think that people now say with COVID, "Hey, there's this new technology. Gosh, it really looks safe, it looks really high quality. Maybe we have some inertia where before we wouldn't have adopted it, but now we really have to. Because the world is different now." So I think that has been a mindset shift.
I think also the economic resources of the U.S. health care system are somewhat fixed. As there is more care needed, there has to be a way to move funds to COVID treatment and things like that. And so, making care more efficient without reducing quality — that is also, I think, a driving factor for change.
How do you see the future of technology and health care in the post-COVID world versus how you had envisioned it in January?
I think that there will be a significant portion of the workforce that will work remotely in perpetuity. I think that will drive thinking around other ways interactions can be virtual and still effective, including in health care. So I think you'll see a lot more telehealth, a lot more virtual doctor's visits and seeing what can be done over videoconference. I think it makes health care more efficient for people and for physicians. It makes health care more accessible. It reduces wait time.
A lot of people are seizing this moment to jump in and test out a bunch of different medical AI technologies — what do you see as the opportunity and the risks?
I think it's important, even in these times, to move fast but also move with purpose and move with quality, accuracy and safety. I think there are many ways that harm can happen right now — of course, many hundreds of thousands of people are dying of COVID — that's very harmful. I think another source of harm would be to bring to market poorly tested products, poorly tested vaccines, things that would ultimately have to be pulled, or that ultimately would not be effective, and then would set back innovation in health care or set back trust in the health community. I think we do need to move with speed and with purpose in times like these. But I think that just as the pandemic is causing a lot of harm, bringing shoddy medical products to market will also cause a lot of harm. And that could last even longer than the pandemic.
And how do you think tech companies can move forward speedily without running afoul of those issues?
I think we should still do clinical trials of new medical devices. I think that we can accelerate the review of those clinical trials, so companies and regulators can be working faster and understanding that people are waiting, particularly for COVID-related medical devices and treatments, and can be reviewing things in the fastest possible time. I think that access to clinical data for clinical trials, if it's safe and anonymized, can be made more widely available. During these times, I think that's safe, and I think that's something that is effective.
Next in Protocol's Health Care Manual: Electronic health records haven't delivered on their promise. Could COVID-19 change that?
Correction (3:50 p.m.): An earlier version of this story misstated which hospitals Alexander Kagen was associated with.
Andie Coller is a journalist and content consultant who specializes in framing information for impact. She has been a reporter and editor at POLITICO, deputy editor of National Journal magazine, and national audience editor for McClatchy. She is a past program director of the National Press Club Journalism Institute.