Monday, March 21, 2022

ARPA-H Needs to Think Bigger

Everyone loves DARPA, the Defense Advanced Research Projects Agency that is credited with such hits as the internet and GPS, but is also responsible for things like the Boston Dynamics back-flipping robots and even Siri.  DARPA’s mission is to make “pivotal investments in breakthrough technologies for national security,” but, as the previous examples illustrate, we can’t always tell how those breakthrough technologies are going to get used. 


Healthcare is, at long last, getting its own DARPA, with ARPA-H (Advanced Research Project Agency for Health).  It’s been discussed for years, but just last week was finally funded; a billion dollars over three years.  But I fear it is already off on the wrong foot, even ignoring the fact that President Biden had requested $6.5b. 

Let’s start with the problem that it not only doesn’t have a leader yet, but it doesn’t even have a home.  By law, the HHS Secretary now has 30 days whether to make it part of NIH or establish it as a separate part of HHS.  Respected experts like Dr. Francis Collins, former head of NIH, advocate making it part of NIH, to take advantage of NIH’s resources, while others want it to be independent.  Congress is weighing bills to both effects. 


More worrisome, though, is that, according to President Biden, “ARPA-H will have a singular purpose: to drive breakthroughs in biomedicine.”   The mission wasn’t originally that limited, but priorities like the “cancer moonshot” tend to focus emphasis.  There are certainly huge breakthroughs ahead in biomedicine, and that’s both exciting and needed, but biomedicine isn’t going to solve all of our health and healthcare problems.

Take, for example, the developing fiasco that is the replacement VA EHR.  Decades ago the VA developed VistA, one of the first attempts at an EHR  (its origin story, as reported by Politico in 2017, is fascinating), but during the Trump Administration it was decided to move to a commercial EHR.  Cerner won the bid.  According to three reports issued by the VA Inspector General last week about the first implementation, it is not going well, to say the least. 

Politicians like Senator Patty Murray and Senator Jon Tester have expressed their concerns, not for the first time.  Senator Murray was clear: “Here is my message to VA: stop the rollout before there is another catastrophic failure.”  Apparently we’re already past the point of avoiding a catastrophic failure; now we’re trying to avoid “another.” 

Our bar is way too low.

Not to mention, the costs have already soared to over $16b, way over original projections.  It’s worth noting that not only is VistA open source, but it’s free: “VistA is public domain and freely available through the US Freedom of Information Act (FOIA).”  Just saying…

Now, I’m not intending to pick on Cerner – its DoD EHR rollout seems to be doing better – or advocating for VistA. I don’t want ARPH-A to help invent a new and improved EHR.  I want them to help invent the technology that is to EHRs as EHRs are to paper records.  I want a big leap.

I want the next generation of health information technology, allowing people to store, share, exchange, and use key health information, making all that look easy – like magic (as Arthur C. Clarke posited, “any sufficiently advanced technology is indistinguishable from magic).   I want it to be as useful for the people whose health it is as for the clinicians who help them with it.  I want technology that may not become mainstream until 2050, but which would still be useful in 2100.

I don’t know what technologies will be important to that. Maybe holograms, maybe digital twins, maybe DNA storage, maybe blockchain/Web3, maybe AI, maybe quantum computing.  Maybe all of those, and others.  The important thing is, think big enough.  About this problem, and others.

Biomedicine is doing amazing things (thank you, COVID vaccines!), and opening all kinds of doors for new treatments, but all those doors seem to be ones that are more expensive and more rooted in a medical approach to health. 

We should be looking at breakthrough technologies that get at health in our everyday lives.  How do we track it, how do we foster it, how do we improve it where we live?  Smart toilets, self-monitoring, AI assistants, 3D printing of prescription drugs and even devices (hello, nanobots!) – those are already 2022 technologies that can be used.  There are undoubtedly breakthrough technologies that will make those look primitive.  Those are the kinds of thing ARPA-H needs to help develop.

DARPA’s projects can take 20-25 years to reach commercial viability – if they ever do.  We can’t afford to think too short-term (e.g., anything less than 10-15 years) or too narrowly (e.g., only biomedicine). When we think about ARPA-H projects, we should thus be thinking about what we want to our health and our healthcare system to be in 2050. 

I hope the VA EHR recovers from its stumbles.  I hope it coordinates seamlessly with the DoD Cerner EHR.  I wish all Cerner EHR coordinated seamlessly with all other Cerner EHRs, that all Epic EHRs coordinated seamlessly with other EHRs, and, really, that it shouldn’t matter which EHR vendor information started in.  But wishing and hoping isn’t getting us where we need to be; we need breakthroughs. 

As Lisa Jarvis wrote in Bloomberg: “ARPA-H needs to foster a culture in which it’s possible to quickly test ideas that fall outside the norm. It needs to cultivate wildly promising lines of thinking that might lead to a breakthrough — or not.”

Based on DARPA’s experience, it’s not important that the leader be a visionary.  We don’t need a Steve Jobs or Elon Musk.  We need a leader who can separate the crazy from the wild, who welomes that wild, and who knows how to get out of the way of innovators trying to make the wild plausible.   

If what we’re debating is whether ARPA-H should be in or out of NIH, we’ve missed the point.  If we’re focusing ARPA-H on biomedicine, we’re missing the opportunities.  If we’re just trying to avoid more catastrophic failures, we’re having one.

Think bigger.  

No comments:

Post a Comment