When Bad Policy = Bad Business Models = Bad Public Health

0:00:05 Hi, and welcome to the A16Z podcast. I’m Hannah. In some ways, the coronavirus feels
0:00:08 like something that came out of nowhere, a sort of black swan event. But at the same
0:00:13 time, it’s also been exposing a lot of the fundamental cracks and flaws in our healthcare
0:00:17 system that feel like a perfect storm when they all come together and have hurt our ability
0:00:22 to address a problem that many say we should actually have seen coming. So we’re here today
0:00:27 to talk with A16Z General Partners Jorge Conde and Julie Yu about some of those big forces
0:00:32 and dynamics in the healthcare system that led us to this moment at the intersection of
0:00:37 business, policy, and public health, and how in healthcare in particular, broken policy
0:00:42 can lead to broken business models that in the wrong circumstances lead to major failures
0:00:46 in public health like the one we’re seeing today. We talk about where we’ve seen this
0:00:51 before in the markets of vaccines, antibiotics, and diagnostics, and what should be different
0:00:56 next time so that when a new pandemic hits, we aren’t facing another perfect storm.
0:01:00 Since we have talked quite a bit about some of the cracks on the healthcare system side,
0:01:04 let’s start with the business or the market side. What was the underlying problem on the
0:01:06 market side that got us here?
0:01:12 Much like a virus attacks the weaknesses in the human body, this pandemic spreads by effectively
0:01:16 attacking or exploiting the weaknesses across the healthcare system writ large in the United
0:01:22 States, healthcare is a fairly unique case in that this is one industry in one area
0:01:30 where policy sets business models and where you have bad policy, you have bad business
0:01:34 models which can lead to market failures, which can lead to public health failures. We’ve
0:01:40 seen it happen in the vaccines industry. We’ve seen it happen in diagnostics. We’ve seen
0:01:45 it happen in anti-infectives and antibiotics more broadly. And so these aren’t isolated
0:01:47 examples where the system fails.
0:01:52 As a system, historically, we’ve made very little investment into prevention in general.
0:01:57 Really where the money has been made is in the treatment of patients who get sick. And
0:02:01 so there’s been really just an orientation around the incentives being aligned with waiting
0:02:08 until those patients do get sick to then provide treatments and therapies and procedures that
0:02:13 generate more revenue, unfortunately, and also higher margins for physicians and hospitals.
0:02:16 And you’re seeing a version of that here where, again, there’s been very little investment
0:02:20 in preparedness for these kinds of pandemic disasters.
0:02:24 So let’s pull apart those threads in those three different areas. So because vaccines
0:02:29 are so top of mind, let’s dive into vaccines. Why does that not a successful market?
0:02:33 Well, I think vaccines in general has been a difficult industry for a couple of reasons.
0:02:39 Number one, if you’re developing a vaccine for something that already affects humankind
0:02:45 broadly, those are considered pretty commodity products today, right? Your mom’s vaccines,
0:02:49 your, you know, any of the number of vaccines that children get in their regular stables
0:02:53 are relatively commodity product. They’re not differentiated. You can’t charge a lot
0:02:58 for them because these are almost basic staples of public health. And therefore, you know,
0:03:01 relatively speaking, we all need them. Yeah, we all need them. They’re well covered and
0:03:05 they’re widely available. The trick comes when you have something that emerges quickly
0:03:09 and has the potential to spread rapidly. COVID-19 is not the first example of this that we’ve
0:03:15 seen. We’ve seen this happen numerous times, whether it’s with SARS or H1N1 or West Nile
0:03:20 or Zika. Obviously, Ebola was one that was a wake up call and made a lot of people nervous.
0:03:25 And I think if you look at a lot of those historical examples, generally speaking, what
0:03:33 we saw happen was the companies that did have active vaccine programs were asked essentially
0:03:38 to stop doing everything they were doing to develop programs against whatever a specific
0:03:43 emerging threat was. And they responded to that call for action. And what happens with
0:03:48 vaccines is when the threat goes away, you know, the urgency tends to go as way as well.
0:03:52 And once the urgency has gone, the market goes with that, along with that, these companies
0:03:57 that were developing vaccines for these specific newly emerging threats. In many cases, they
0:04:01 were left holding the bag where, you know, in some cases, orders that governments had
0:04:06 put in for them kind of went away when the threat went away. And so the business model
0:04:09 of saying, well, ramp up production for an emerging threat, where if you are successful
0:04:13 the market for what you’re developing goes away. And so the public health efforts are
0:04:16 successful in containing the disease and you don’t need the vaccine. And therefore there’s
0:04:21 no market for it. And therefore you essentially wasted the effort. The other thing that’s true,
0:04:25 that has been true in previous pandemics is there’s generally a call on the industry by
0:04:31 governments to make sure that the vaccines become available at cost or at low cost. And
0:04:35 it’s very hard to build a business around that, where if you’re building for, you know,
0:04:40 an event like you called it a black swan event. But at the moment of that black swan event,
0:04:43 you’re not able to recoup profits for some valid reasons that you obviously have to make
0:04:47 this widely available. It’s a very difficult business model unless you’re in a situation
0:04:51 where it’s the entire globe, and then it’s all of a sudden. Well, yeah, but then you
0:04:55 need the ability to actually ramp up and be able to produce vaccines at that scale and
0:05:00 at that speed. Something that’s a silver lining in all of this is how rapidly some of these
0:05:05 novel platforms have been at looking to develop vaccine candidates. But for that to work as
0:05:10 a business model, you need to either be able to produce a vaccine for every oncoming pandemic,
0:05:15 and you need to be able to have a business model that will enable you to essentially recoup
0:05:19 and make a profit from the investments you’ve made in developing that platform. And that
0:05:20 goes back to policy.
0:05:22 Yeah. I mean, one interesting thing that comes to mind, Horace, you’re talking about this
0:05:28 is there’s a lot of characteristics of vaccines that are somewhat similar to some of the more
0:05:32 novel sort of gene therapies and cell therapies that we’ve talked about, where you essentially
0:05:36 need the vaccine once or maybe a handful of times when it comes to things like booster
0:05:41 shots for your entire life. And therefore, the opportunity to monetize that particular
0:05:46 intervention is very rare in the context of any one patient. You just have to wonder
0:05:51 whether some of the dialogue that’s happening around value-based payments for different
0:05:55 types of treatments, how that would be applied here, because to take the traditional fee-for-service
0:06:01 way of thinking about getting paid a commodity price for a one-time intervention just doesn’t
0:06:05 seem to match well with the paradigm of how vaccines are actually administered.
0:06:08 So I think the analogy is valid. And by the way, I think a lot of the proof is in the
0:06:15 pudding because that kind of incentive structure that you just described hasn’t existed historically.
0:06:20 There have been several calls for better policy put in place. Some of the key areas that have
0:06:27 pioneered this have been sort of the one and done therapies for very rare diseases. And
0:06:32 I think the reason why there was room to have that discussion was, number one, because the
0:06:36 prevalence of those diseases is fairly well understood, even though they’re very rare.
0:06:41 And so you can essentially run the actuarial calculation to say, “Okay, well, if we charge
0:06:47 $2 million to treat somatic muscular atrophy, we’re still benefiting the system a significant
0:06:53 amount by extending life and reducing the need for supportive care and all of those things.”
0:06:58 And so you can actually model those out. And it’s a rare event that with a high value,
0:07:03 you can put a high price on. Here, it’s a little bit trickier because it’s hard to model
0:07:08 out the actuarials on this because a pandemic is, generally speaking, an unknown event.
0:07:13 So unless you have a policy for pandemics broadly, where you essentially provide some
0:07:18 sort of incentive, whether they’re block grants or success fees, et cetera, that would have
0:07:23 to be really large for any institution that comes up with a effective vaccine against
0:07:29 a newly emerging threat, you’re sort of trying to solve for an unknown number. But I think
0:07:34 we could learn a lot from what we’re seeing with these new modalities like gene therapies
0:07:39 and cell therapies that have in many ways trailblazed novel business models to make them viable.
0:07:44 I understand how for vaccines that model falls apart when sort of a massive event like this
0:07:48 happens really quickly. But how about with antibiotics? That feels like something that
0:07:53 everybody needs, that we know there’s an increasing demand for the right kind of antibiotic. Where
0:07:58 does the market policy, health care, public health failure come together there?
0:08:03 Well, so there it’s, you know, similar situation with different circumstances, right? So in
0:08:08 the case of antibiotics, public health agencies and the medical practice broadly have been
0:08:15 very focused on how you prescribe antibiotics and in what order, in order to prevent the
0:08:21 emergence of resistance. And so we have broad spectrum antibiotics, and then we have narrow
0:08:27 spectrum antibiotics, and we have ones that are more potent than others. And once bacteria
0:08:32 become resistance to all things, and we have the threat of superbugs, where you basically
0:08:38 have no last line of treatment against a bacteria that has become resistant. And so the conventional
0:08:43 thinking around antibiotics has been to think of it as first line and second line and third
0:08:49 line. And so if you develop a novel antibiotic, you might be addressing a very important unmet
0:08:57 need. But by definition, physicians are going to use these new antibiotics as sort of last
0:09:03 line of therapy. I develop a new antibiotic, and it’s good, it’s only going to be used
0:09:09 sparingly. And if I’m charging per treatment or per use, that’s obviously not going to
0:09:14 be a very effective model for me either. Right. You don’t want to be having to use these unless
0:09:18 everything else has failed. That’s right. And by the way, we’ve seen the real world
0:09:23 examples of this as well. Large pharmaceutical companies have also exited, like with vaccines,
0:09:29 they’ve also exited the antibiotic space. Novartis has exited, Sanofi has exited. It’s
0:09:33 very hard for them to find a way to make a profitable business for all the reasons we’ve
0:09:37 been talking about. There was a startup called the Kaogen that was actually successful in
0:09:43 developing a novel antibiotic against a nasty bug. So they had a public health success.
0:09:48 But what they found was that they didn’t have a business model. And so they went bankrupt.
0:09:51 This is a really strange industry where you have something that’s so successful that you
0:09:55 cannot build a business on it. Scott Gottlieb, the former head of the FDA, who’s of course
0:10:01 been so vocal and helpful throughout this COVID-19 pandemic. He had floated the idea
0:10:06 a year ago of similarly creating some sort of incentive structure so people could develop
0:10:12 new business models around novel antibiotics. And these included things like, similar to
0:10:18 vaccines, success fees for developing a novel therapy. He even floated an idea on something
0:10:23 like subscriptions where, you know, institutions would subscribe to get access to antibiotics
0:10:28 and they wouldn’t pay for use. They would pay for access. So it’s almost like an all-you-can-treat
0:10:33 model versus, you know, a paper pill model. Why would that need to come from policy? Why
0:10:37 wouldn’t that just make good market sense for the market to respond in that way? Well,
0:10:40 I think you’re going to see a combination of those things. So some of it needs to come
0:10:46 from policy in the sense that you can require institutions to have a broad toolkit of antibiotics
0:10:50 and require them to carry all of them. And then if you require that, the companies are
0:10:53 developing novel antibiotics, come up with new business models to make it viable for
0:10:58 them and to make it viable for the hospital systems. But I mean, I think you do need policy
0:11:04 at some level to define what is required for standard of care. And if you can do that,
0:11:09 then you at least create some room for companies to develop novel business models. A great
0:11:15 example is in self-therapy and gene therapy. Like unless you change the way CMS pays for
0:11:20 therapies and how they calculate cost of therapy, you can’t have installment models in medicine.
0:11:22 So there had to be a policy change there.
0:11:26 People do look to CMS to be kind of the tip of the spear with regards to these things.
0:11:33 CMS reimbursement is the ultimate policy change that generally is what triggers this type of
0:11:38 business model innovation and adoption of these types of things in the market. And commercial
0:11:44 payers tend to follow suit when CMS makes those kinds of policies. CMS did put into place
0:11:49 a what’s called a qualified infectious disease products program in partnership with the FDA
0:11:56 around antibiotics to essentially up the reimbursement rate for the use of antibiotics in hospital
0:12:01 based settings under certain circumstances. So there is some precedent that shows that
0:12:05 there is a willingness to do this sort of from the top down. There was other legislation
0:12:11 proposed that has not yet been passed around the concept of essentially mandate that the
0:12:17 hospital report on the utilization of antibiotics and sort of demonstrate that they are abiding
0:12:21 by standard of care. And it’s sort of a quid pro quo where in exchange for doing that and
0:12:26 sort of demonstrating compliance with that kind of program, the proposal is that there
0:12:28 would be higher reimbursement rates for that.
0:12:32 The other interesting thing about antibiotics that I find is I’ve had a number of experiences
0:12:36 myself where let’s say you go to an urgent care clinic, you think you have strep, they’ll
0:12:40 do a test, but the test results aren’t going to come back for a couple of days and therefore
0:12:46 you are put sort of prophylactically on an antibiotic without yet knowing what the pathogen
0:12:51 is or whether or not you even have that pathogen. And there’s studies that show that it’s upwards
0:12:57 of a third of antibiotic prescription could be unnecessary, but given the delay in the
0:13:00 information needed to determine whether or not it’s correct, that’s just sort of the
0:13:04 general practice of how it plays out. So it’s interesting in that this sort of intersects
0:13:10 with how having better real time diagnostic capabilities could actually impact and just
0:13:13 reduce the inefficiency amount antibiotic use as well.
0:13:18 So it’s interesting not just the sort of market policy public health failure starts to come
0:13:22 together, those actual three buckets start coming together as well. But let’s talk about
0:13:29 diagnostics now and pull again apart these sort of different failures of market policy
0:13:35 public health. Why is the diagnostic market not as successful? Why was that broken?
0:13:39 Diagnostics historically have been a challenging area for a couple of reasons. The first one
0:13:46 is it takes a while to develop a diagnostic. It’s time consuming and expensive and it wasn’t
0:13:51 easy to really recoup that investment because reimbursement rates historically haven’t been
0:13:52 very high.
0:13:56 But why would it not always be better to know if you have some, I’m sort of like not clear
0:14:00 on the very question of why is a diagnostic not always valuable?
0:14:03 You know, look, I think diagnostics are a form of prevention. And I think if one thing
0:14:08 we’ve seen generally across the healthcare system is prevention is not always the most
0:14:12 effective business model. Whether or not it’s valuable isn’t the debate. It’s how it’s
0:14:16 reimbursed. There’s plenty of examples where reimbursement rates don’t always track to
0:14:21 value, especially in the system where a lot of things have been around fee for service
0:14:24 versus value based care.
0:14:30 Diagnostics as an industry is a fairly tricky one because reimbursement levels for diagnostics
0:14:36 have over time have trended down. If you look at things like diagnostics that could be used
0:14:42 to determine which patients should be a candidate for a therapy, for any therapy, for any number
0:14:47 of conditions, those have historically been underutilized. One of the reasons that people
0:14:51 speculate is because physicians are going to prescribe the treatment that they think
0:14:55 their patients are going to benefit from. And unless the diagnostic is meaningfully going
0:14:58 to change that decision, they probably won’t use it. They’ll use other factors like risk
0:15:04 factors or family history. In many ways, the diagnostics industry has been viewed, I think
0:15:12 unfairly, as a commodity industry. There is an element of competing on price and trying
0:15:15 to essentially make it up with volume. And that’s why you have very large players in
0:15:20 the industry that have these very, very large menus of tests that they provide and they
0:15:24 do it at scale because the amount you’re going to reimburse on a per test basis is relatively
0:15:30 small. And so this whole area of diagnostics as being viewed as a commodity has led to
0:15:31 generally underinvestment.
0:15:37 It’s not that it’s not perceived as useful. It’s more that the price that you get paid
0:15:42 for that component of the patient journey is not where the money’s made. Like a cholesterol
0:15:47 test, a very basic test that’s pretty routine. The value of the test itself is probably a
0:15:52 few tens of dollars in terms of how much you get reimbursed for that test. But the outcome
0:15:56 of that test, if you are in fact diagnosed as having high cholesterol, the result of
0:15:59 that is that you’re put on a statin really for the rest of your life. And that’s really
0:16:03 the blockbuster category of drugs. And so that’s just an example of where the value
0:16:10 distribution is very skewed. CMS very recently just published their list price sheet for
0:16:16 how they intend to reimburse for the COVID test. You do a double take when you see it.
0:16:22 It’s literally in the $35 to $50 range per test that would be reimbursed by Medicare.
0:16:26 And when you think about the value of the information being provided by that test, especially
0:16:31 right now when that information could literally mean life or death, it’s kind of astounding
0:16:36 to think that it’s so little compared to these disaster stories that you hear about patients
0:16:41 in the ICU on ventilators, having to undergo anesthesia to get those ventilators in. There’s
0:16:46 just a huge skew in terms of where the dollars flow relative to where the perception of the
0:16:47 value might be.
0:16:50 I think what we’ve seen in diagnostics in terms of how they have been deployed across
0:16:55 the country is a great example of, unless you have very clear policy on who can get
0:17:01 tested and what it will cost, it makes it very hard for this to be deployed very widely.
0:17:06 And we can put aside for a second any issues that actually took place in the logistics
0:17:11 of rolling out tests. The incentives just weren’t clearly in place early enough.
0:17:15 I mean, what we’re talking about is all these broken models that show up when something,
0:17:20 do we need to wait for these moments of really kind of horrifying failure when there isn’t
0:17:25 a vaccine available, when we don’t have the right antibiotics or the right tests in order
0:17:28 to change some of these ingredients to prevent it from happening again?
0:17:33 Well, I think the specifics will matter. But I think generally speaking, when the next
0:17:36 pandemic comes, there are a couple of key things that I think would be helpful to have
0:17:41 in place beforehand. One, from a policy standpoint, it should be clear that, number one, there
0:17:47 won’t be any cost for any member of the public to get tested, assuming tests are available.
0:17:52 Number two, for those companies that are providing tests that need to in many cases massively
0:17:57 scale up capacity in order to meet demand, that they will be reimbursed in a meaningful
0:18:03 and timely manner. Number three, I think it would be helpful to have from a policy standpoint
0:18:07 coverage of treatments associated with whatever the pandemic is.
0:18:12 And that way, you don’t have the situation where somebody that’s underinsured or uninsured
0:18:17 is avoiding the healthcare system out of fear of cost repercussions. I mean, that’s obviously
0:18:21 a challenging thing to do in a normal environment. It can be a catastrophic thing to do in a
0:18:25 rapidly emerging pandemic environment. And so I think there’s certain things you can
0:18:30 do from a policy standpoint to make sure that both the public is responsive, but also industry
0:18:34 and those that have technology to address the situation are also responsive.
0:18:39 But how about from a market perspective? I mean, are there things that you can do to
0:18:45 fix these basic market dynamics of dealing with rare emerging diseases and last line
0:18:48 of defense antibiotics, things like that?
0:18:53 So if the policy were to be anyone that can provide a validated test against the pandemic
0:18:59 is going to be reimbursed X amount, and these are going to be the reimbursement criteria,
0:19:05 then you have incentive to very quickly scale up production to do that. If the policy is
0:19:10 that anyone who wants and seeks out testing, regardless of whatever that individual circumstances
0:19:14 would be, they can get access to that test and they won’t be charged at all for that
0:19:20 test. You’ve also created demand. In effect, what’s necessary in a pandemic is to ensure
0:19:24 that you have a guaranteed buyer, whether it’s, you know, you’re developing vaccines
0:19:28 that are going to be deployed broadly, whether you’re developing tests that need to be distributed
0:19:33 around the country or around the world, you know, whatever the situation may be, if you
0:19:39 need to scale something up for a one time event, you need to be sure from the supply
0:19:45 side that someone is going to buy that product. And really the only institutions that have
0:19:49 that kind of scale when you’re talking about pandemic level events are nation states or
0:19:54 countries. And so I think another key aspect of the right policy in place is essentially
0:19:59 have the backstop guarantee that the government is going to step in and buy product if you
0:20:03 develop it to address an unexpected event like a pandemic.
0:20:08 One of the big failures that we’re experiencing right now is that we who are told to stay at
0:20:13 home and avoid the hospital don’t have a means to get tested at home. And, you know, there
0:20:18 were, I think there was a set of unintended consequences that occurred when the FDA intervened
0:20:23 and put into place this emergency approval process for new tests that were being developed
0:20:27 for COVID did actually, and I think this was done unintentionally, but is one of the major
0:20:31 reasons why there’s been such a bottleneck in getting new tests developed unintentionally
0:20:36 created a need or requirement that the FDA does approve under these emergency provisions
0:20:41 every test that is being developed. And so we saw a bunch of innovative companies, startups,
0:20:46 et cetera, spin up home based tests that would work for patients in their home setting, which
0:20:51 would not require them to go to a lab or go to a clinic to get the testing done. But those
0:20:57 were unfortunately shut down because they did not comply with these emergency rules.
0:21:01 And I think things like that where there are sort of intersecting policies that may conflict
0:21:05 with each other between different government agencies have to be looked at differently
0:21:10 such that the next time this happens, we’re able to respond and have the market, you know,
0:21:14 sort of do its thing and be able to get new product out there in a way that doesn’t get
0:21:15 hung up in bureaucracy.
0:21:18 Yeah, it’s interesting because I feel like a lot of what we’re talking about is this
0:21:24 delicate balance of regulation, right, in a heavily regulated space where regulation
0:21:28 can push forward innovation in some ways, but then time passes and you don’t want that
0:21:33 to then become kind of a weight that you’re dragging along when events change or things
0:21:39 unfold. Like, how do you generally think about that balance between the heavy regulation
0:21:45 that we need in place for helping to sort of perhaps create market incentives or the
0:21:49 relaxation we need in others like in testing therapies? How do you think about how those
0:21:52 evolve over time and making sure that we keep that balance?
0:21:56 Yeah, I mean, we talked a little bit about this with regards to telehealth and the fact
0:22:02 that many of the laws both at the state and the federal level that have prevented or at
0:22:07 least created friction for telehealth to be very broadly adopted. Our laws that were put
0:22:12 into place like decades ago before anyone contemplated telehealth. And so, you know,
0:22:16 even though it’s unfortunate that this kind of event and pandemic is the forcing function
0:22:20 for us to revisit some of those things and actually undo them because the world has
0:22:24 changed so much because technology has changed so much because the way that we deliver care
0:22:28 is changing so much where the laws that have been put into place are, you know, maybe very
0:22:32 outdated. I think that’s just a general thing that this whole pandemic is shedding a light
0:22:36 on is the fact that so many of those rules that have been put into place, you know, likely
0:22:39 don’t make sense anymore in this day and age.
0:22:44 Look, I think the idea that we need to ensure that, you know, we do no harm, that things
0:22:49 that we develop to be an effective intervention against how we treat or manage disease are
0:22:54 in fact effective to ensure that we have the processes in place to make sure that we are
0:22:59 in fact not blind to issues around safety or efficacy. I think it’s very important.
0:23:03 But as we’ve talked about before, the world has changed in terms of, you know, what medicines
0:23:08 look like. You know, now we have gene therapies and cell therapies and digital therapies. And
0:23:14 I think the regulation has to evolve so we can properly evaluate those opportunities to
0:23:19 change how we treat disease. And we’re also seeing a lot of innovation that’s coming down
0:23:24 the pipe right now in terms of how we think about how do we monitor patients, right? Historically,
0:23:29 that was always done, you know, at the point of healthcare at a hospital or at an institution.
0:23:33 But it’s, you know, things like, you know, monitoring of patients with wearables emerge.
0:23:37 That’s going to change a lot of paradigm. So I think the reality is that, you know,
0:23:41 it is true that regulation is always lagging behind innovation. I think that will always
0:23:42 be true.
0:23:47 But I think what we need to do is find ways to close the gap between the innovation that
0:23:52 emerge and the regulations that help us govern those innovations. And to the extent that we
0:23:56 can further close that gap, we’ll obviously be in a better position, you know, for future
0:23:58 emerging health threats.
0:24:00 Thank you guys so much for joining us on the A16Z podcast.