It’s time to talk about a new venture called EQRx. This has made quite a splash in the last few days at the JP Morgan investor conference, and it’s been launched by Alexis Borisy (involved with founding and/or helping run CombinatoRx, Foundation, Blueprint, WarpDrive Bio, Editas, Relay and others) with the aim of generating cheaper medicines. How, you ask, are they going to do that? Are they aiming for cheaper generics? The state of California has been talking about that, but it’s a tough place to work in. The high-volume generic drugs are already competing on price, and the low-volume ones, where some bizarre pricing does indeed take place, don’t have much of an impact (being low-volume and all).
No, EQRx is more ambitious. Their plan is apparently to come up with new competing pharmaceuticals in existing categories and sell them for lower prices than the patented ones. Which immediate makes one ask again, how are they going to do that? I can do no better than to quote Borisy, via this C&E News writeup from Lisa Jarvis:
“The world has fundamentally changed. . .Today, you can do a virtual screen of a billion compounds, do on-demand synthesis of all of those, and you can do it overnight in the cloud”
My challenge here is to write about this in something approaching a measured tone. The temptation is strong to spend the next couple of paragraphs in full Daffy Duck mode, because that quote alone seems to have been designed to send me over the edge. So to buy time, let me refer to this Stat writeup by Matthew Herper, where Borisy assures the world that “This is not fantasy world. This is something that can be done” Here’s exactly what he has in mind, by the numbers:
Over the next 10 years, Borisy said, he’d like for EQRx to start developing somewhere in the ballpark of 50 different experimental medicines. He wants the company to come out with its first medicine in five years, and to have 10 drugs within a decade.
Must. Not. Lose. It. Let’s buy some more time. I note that Herper’s article focuses on the downstream business uncertainties of this plan. As he says, the idea of a “fast follower” drug is not a new one in the business, but neither have such compounds ever competed on price. The business question is whether physicians and insurance companies are ready to make that part of their decision-making process in the prescription drug world because frankly, they don’t do so much of that now. EQRx is a bet that they’re ready to. Insurance companies are very interested in generic drugs when available, and will often require patients to exhaust the existing lower-priced therapeutic options before they’ll pay for a more expensive one as well, but we really haven’t had that many examples of prescription medicines fighting it out on price alone.
But EQRx is also, of course, a bet that they can discover all these medicines in the first place. And that’s where I have comment to offer. It has not escaped my attention that they are picking up funding from some of the venture capital shops that have traditionally stayed out of biopharma, such as Andreessen Horowitz. Here’s their take, which tiptoes up to this question:
But how? What secret do they have that the multi-billion dollar biopharma industry doesn’t? Often the biggest innovations in technology address the process, not the product. Take Amazon, for example; by building a different retailer from the ground up and infusing tech throughout, they were able to deliver products faster, at a lower price and a higher margin. This simply can’t be done by revamping an existing company, as many retailers and other companies have tried. In much the same way, EQRX is now reimagining how medicines are created, tested, and commercialized by re-engineering the system itself.
The hallmark of good engineering is where the solution is simple, but not easy; the answer elegant, but not overwrought. EQRx shifts science risk to execution risk by focusing on known biology, where drug targets have been validated and proven valuable for treating a disease. They remove friction within the system through early and deep partnerships across the healthcare value chain. And most importantly, they integrate data science and technology end-to-end throughout the discovery and development process. Engineering a better product is good, but engineering a better process is best.
So we come down to “They’re gonna be like Amazon” and “They have lotsa data science”. That’s unfair of me, but not as unfair as it should be. Let’s get down to it: if a16z and the other backers of EQRx think that they have shifted the science risk here, they are mistaken. You cannot dodge it in this business; science risk is stalking you all the way through the execution of your business plan. It would indeed be dreamy if that were not the case – it’s pretty to think so. But a company like Amazon is a story of pure, world-class execution, coming up with a smoother, cheaper, more convenient way to sell things that you didn’t invent or optimize by yourself, that you don’t manufacture yourself, and that you already know that people are buying constantly (and what’s more, buying with ferocious price discrimination). Coming up with ten new drugs in ten years ain’t none of those things, guys.
Why not? Let’s look at the existing fast-follower “me-too” landscape. As mentioned, the folks developing such compounds are not doing so with an eye to competing on price. They have large profits in mind, certainly. If they were able to come up with ten such drugs in ten years, why wouldn’t they? There is no incentive to move slower – the burst of speed that EQRx is talking about and its desire for lower-priced drugs are two separate things and they are not related. Why hasn’t anyone done the speed part while keeping the high prices part, you wonder?
Is that because, as the Silicon Valley types would have it, the existing drug development system is too stuck in its old ways of thinking and operating and is ready for some disruption? What disruption is EQRx bringing to the R&D process? When I read that Borsiy quote from the C&E News piece, I don’t know whether to laugh or shiver. No, in case you were wondering, you cannot virtually screen a billion compounds overnight. Allow me to refer you to Pat Walters, who knows what he’s talking about in this area. Virtual screening can be quite useful, but the larger and more useful it gets, the less trivial it becomes. A specific problem is the generation of false positives, which can eat you alive. This is the point to note that no, you cannot do “on-demand synthesis of all of those”, either. I know that Borisy is not talking about synthesizing a billion compounds, of course (although who knows what some of the people reading his statements think), but compounds of any real complexity are still kind of a pain to make in quantity.
And let’s remember, EQRx wants patent-breakers. That’s what the company is, at bottom: an attempt at speedy patent-busting, with the disruption-on-price stuff as a separate menu item. The fastest way to come up with a fast-follower is to hew as closely to the existing chemical matter as you can and still get IP. The further away you venture, the greater your chances of surprises – PK surprises, selectivity surprises, toxicology surprises. Mind you, you’re in for those no matter how closely you stick to the known compounds, but you’ll definitely uncover more of them as you move into new chemical space.
Now, it’s true that the existing fast follower drugs generally come along with some advantage over the older compounds – more convenient dosing, fewer side effects, better efficacy and so on, because they’re competing on these and not on price. Not all of these turn out to be real in clinical practice, but they were aimed towards these things during development, because coming to market with “Here’s something that’s pretty much the same as Drug XYZ, which will be going generic soon, and here’s your opportunity to still pay a higher price” is not a winning proposition. But this brings up two other factors. The first is that even in fields where there are several drugs all hitting the same mechanism, it’s usually the case that they behave differently in different people. A true copycat drug that acts exactly the same is very rare – partly because that’s not what people were aiming for during development, but even more so because of the variety of physiology and metabolism found in the human population.
The second factor is timing. If you’re going to introduce a fast follower drug, it kind of does have to be fast. Otherwise, the competition will have gone generic, or will be getting close, by the time you’re hitting the market, and then (as mentioned) you’d really better have a reason for people to pay for your compound. If EQRx is trying to lead a price revolution, it’s got to be in the prescription side of the business, because generic drugs in this country are already pretty cheap. So their ten-drugs-in-ten-years pace, while still frankly nuts, is the sort of thing that they’ll actually have to meet if they’re going to come with less costly alternatives while the expensive stuff is still on patent.
Let’s get real: no matter how spiffy your data science integration is, discovery and development of a new drug – even one that copies an existing one as closely as possible but is still patentable – just takes longer than that. Finding the compound. Setting up and validating the assays. The counterassays. Screening. Analog synthesis and analog screening. Scale-up. Pharmacokinetics. Toxicology. More toxicology. Formulations. GMP scale-up. Design of a trial. Recruiting patients. Running the trial. Sending it all to the FDA. These things take irreducible amounts of time, and not-so-irreducible amounts of cash. I know that I sound like a crusty codger who’s all salty ’cause he’s about to get Amazoned, sure. There are people in the Valley who live for that kind of story and to see the looks on the faces of people once the Mighty Disruption hits them. But let me put my cards out on the table:
I will put up $500 dollars on LongBets.com against the proposition that EQRx will produce ten drugs in the next ten years. Money on these propositions goes to charity – if someone associated with this venture would like to take the other side of this, I will gladly name an organization that my money will go to if I lose. I am easy to contact, by email or Twitter. I am registered on LongBets, and will be happy to work out more detailed terms of the wager in advance on request. Failing that, I will write my own statement and post it there, for all comers to challenge and for the bet to be eventually mediated and recorded. Updates to follow.
Why am I being such a curmudgeon about it? I certainly have nothing against lowering drug costs. I definitely have nothing against speeding up drug development. What rubs me the wrong way is hype. I freely admit that part of this is having devoted the last thirty years of my life to a difficult and complex process, which makes it difficult to be take with equanimity being told about how it could all be done so much faster and easier. But if it can, perhaps such a public bet will spur on the people trying it that much more. If EQRx really can deliver so many new drugs so quickly, I will happily and publicly eat crow. Takers?