Funding Open Source Drug Discovery
I’ve been speaking to many people about funding open source drug discovery. If there is no secrecy, and no patents, then Who Will Pay? It’s the obvious, central question behind the whole enterprise, particularly around Phase 3 trials. I’ve been asked this probably 20 times in the last 2 weeks of conferences and meetings.
The intellectual arguments in favor of openness are clear – openness means efficiency, inclusiveness, etc etc. It’s a done deal. But what about the money? How do you marry open source and a Big, Sustainable Financing Model?
An answer is to say “Pharma can’t do Open, so the money needs to come from the public sector. Or philanthropy.”
I guess that is true, but when I say this to people, their spark of interest is lost. This just sounds like another drain on the public purse. “These are times of austerity” people will say “where will you find the money for developing new drugs, for clinical trials.” I think people are hoping for something striking and new.
Funding for open source drug discovery will come from many places. But I am left wondering whether there are really new mechanisms we could consider. I’ve previously wondered aloud about retrospective patents and data exclusivity.
Major financing mechanisms have been proposed, such as the Health Impact Fund (“take a drug to market and you’ll be rewarded if it works, then it goes generic”) or Advance Market Commitments (“You can invest since there will be a market”) backed by large financing facilities – essentially permitting GAVI. There are more recent calls for up-front funds e.g. for antibiotics, which is mimicking what has already been achieved with the Global Fund. There are much-maligned Priority Review Vouchers (“develop a medicine for malaria and you’ll save millions developing a new drug for cancer”). There are also prizes and combinations of mechanisms.
(Notice how some of these are meant to work: governments issuing guarantees. We’ll come back to that.)
Question: What unifies these mechanisms? Answer: None has been combined with open source discovery (I think). Why? Because there have not yet been any open source projects that have taken a medicine to clinical trials. This is needed. What happens to these financial tools if you bolt on a commitment to complete transparency and no IP all the way from discovery to market?
I’ve been thinking about this hybrid:
On Day 1 of a drug discovery project you commit to open source principles, and you sell shares. People can buy these and sell them. Initially all shares could be owned by a public body, but could be sold to investors. Perhaps the entity selling the shares could be an organization with a portfolio of projects. Donations could be made, and competitive grants could be awarded, independently of such investor buy-in.
The projects run. Some fail. One hits. When the drug is successfully taken to market, the price is set at a level that ensures investors (including those who contributed key resources) get their money back. This would be achieved by legally-enforcible data exclusivity. The medicine would be made by the generics industry, but the price is initially enforced. Once the money has been recouped, the medicine goes fully generic. Everyone will know how much money needs to be recouped, because the development has been transparent.
So this is rather like the Health Impact Fund: payment for performance – in this case the period of exclusivity ends once the money has been recouped through patient use. A more successful medicine will probably have a shorter period of exclusivity. But here we have the ability to buy and sell shares, which adds the motivational aspects of the market. And with a commitment to zero secrecy.
Obviously also the period of price exclusivity could be shortened by cash donations – a zero-risk investment in global health.
I need economists to please point out the problems with this idea. I don’t know what happens, for example, if the managing organisation is valued, through shares, at $1Bn and Bill Gates gives a cash contribution to the research of $1Bn. If Gates wants money back he’d need to buy the shares, correct? Or he could just donate $1Bn and that would not affect the value of the other shares? Similarly governments could pay for the research in the usual way (“Track 1”: grant funding schemes, pooled funds) that would not need to be repaid. The repayment part (“Track 2”) is an option for people providing private funds?
Does that work?
(Notes: i) I’m posting this in advance of reading this important-but-5-year-old-book on the Health Impact Fund and other mechanisms. ii) I’m grateful for interesting preliminary conversations about this idea in the last 2 weeks with Steve Ward, Tom Higley, Steve Williams, Dave Busha and others (which is not to say they endorse))
Jack Scannell, 
mattoddchem, and 2 others are discussing. 
Intermolecular 
Thomas Arildsen 7:01 pm on May 29, 2015 Permalink |
I am not an economist or health professional for that matter, but I cannot help but wonder: what prevents any other generic drug producer from using the openly available documentation of the medicine to produce and sell it at a lower price with no obligation to return any of their earnings to the project that developed it?
mattoddchem 9:55 am on May 30, 2015 Permalink |
Nothing, but, as I understand it, the data exclusivity condition is that if you want to sell a (non-generic) drug, you need to run your own trial in order to generate the clinical data for registration. So: given the available data, you’d be quite confident that the trial would work. But would you really plough money into a trial (that would last years) just to compete with a medicine that is already going to be low cost, and which you know will go generic anyway in a few years? I don’t see this as a strong market incentive.
Chris Sampson 5:51 pm on June 2, 2015 Permalink |
I am an economist, but no expert on this! Hopefully someone in my network will pick this up when I share the link. Economists are generally known for being bad at predicting things, but actually microeconomists are pretty good at predicting how businesses will behave. So it’s good that economists play a central role in HIF et al. For me, the challenge will be getting the regulation right. And the main challenge will be the international context. But pharma will respond to their biggest markets, so if we can get US and EU regulators to work together we’re in with a chance.
Nick Dragojlovic 7:10 am on June 3, 2015 Permalink |
Very interesting post. A few thoughts:
1) The key issue would seem to be what the permitted ROI for investors is before it goes generic…
Investors typically aren’t looking just to get their money back, they’re looking to make the best return they can get. If returns in this sector are sub-par, most money will flow to other investments (tech, real estate, whatever) that will yield a higher rate of return.
2) Buying or selling shares (between investors) won’t really provide an incentive unless the returns are comparable to or better than other asset classes, which probably requires pricing to maximum willingness-to-pay (i.e., current pharma strategy).
3) The one exception to this is the hypothetical “impact investor” that is motivated primarily by the goal of developing a treatment for a specific disease (say, because their child is affected), and is unwilling to simply donate the money (because they don’t like the idea of donating to a corporation or because they think for-profit entities are more likely to bring a drug to market).
The problem: no one knows how big this pool of potential investors is and what sorts of resources they command and are willing to invest in this manner. Could be that there is enough impact capital to fund only 1 project in any year, or it could be enough to fund 100. The viability of this type of drug development model crucially depends on that.
4) In effect, you’re describing equity crowdfunding by impact investors in a “venture philanthropy” model. Again, the viability of this model depends entirely on the number and preferences of impact investors with substantial assets (i.e., accredited investors). Many equity crowdfunding sites focused on healthcare have emerged, and may be good partners in any such venture. For a list, see:
http://fundedscience.com/resources/portals-for-research
mattoddchem 10:13 pm on June 10, 2015 Permalink |
Hi, Nick, thanks for the comments.
1) Correct about the typical expectations of investors, but there is more to investing than money – e.g. Corporate Social Responsibility, ethical investments etc.
2) Ditto – the idea would not necessarily be to maximise returns but to provide modest returns as a feature, i.e. exactly not to accumulate wealth during the course of generating important medicines.
3) Just as you say – this is the idea. And I agree with you over the level of uncertainty.
4) To some extent. I would not be prescriptive about the source of the money, if that’s what you mean. I would want to design this thing in a way that money could flow from government grants, philanthropists and simple private investment.
On reflection I wonder whether I am in effect defining something like a Social Impact Bond?
http://en.wikipedia.org/wiki/Social_impact_bond
Nick Dragojlovic 7:03 am on June 11, 2015 Permalink |
Hmmm. Seems like a Social Impact Bond in an R&D context to which anyone could contribute would basically be something like a crowdfunded X-prize, no? I’m all for prize-based incentives, though it’s worth keeping in mind that the participants must still find a way to fund the R&D to actually win the prize (or meet the terms of the Social Impact Bond), which requires either fundraising (status quo) or a large enough return on the SIB to justify early investors (same issue of lots of public money going to corporations, just in this case via SIB returns, rather than via high prices charged to government payors – at least in jurisdictions with pharmacare).
Agree that there is more to investing than money, but the key question is: for whom and how much are they worth? Total impact investment capital as of 2014 is a tiny proportion of total capital ($12.8B globally, by this estimate – http://www.sauder.ubc.ca/Faculty/Research_Centres/Centre_for_Social_Innovation_and_Impact_Investing/Research/~/media/Files/ISIS/Reports/Social%20Economy%20Reports/Demystifying_Impact__Investing.ashx).
Is there a lot of untapped potential there? Hard to say without more data.
In terms of philanthropy, that’s probably the better opportunity at the moment for impact-focused capital to flow into open drug development. Again, the question is: would the ability to invest in a social venture with limited financial ROI increase the total capital put into the sector, or would it simply crowd out philanthropy? Again, hard to say without more data.
In my opinion, those are the key questions that need to be answered before betting on impact investment as a route to open drug dev instead of or in addition to doubling down on philanthropy.
Jack Scannell 4:00 pm on June 16, 2015 Permalink |
Just so readers know my biases, I have worked in drug and biotech equity investing, drug discovery in a small “bleeding edge” private sector biotech, and in academic research. I now do some work at a translational medicine group at Oxford (www.casmi.org.uk) and a life science policy group at Edinburgh (www.innogen.ac.uk).
I have absolutely zero experience in open source drug discovery (or anything else open source), so apologies in advance if I have misunderstood things or my comments are simply going over old ground. However, with that proviso, here are my reactions.
(1) Clarify the benefits of open science in drug R&D
“The intellectual arguments in favour of openness are clear….” They may NOT be clear to a lot of the people who control large sums of private or public sector capital to deploy in drug R&D. Most would probably doubt that any general approach is going to be good for everything.
Therefore, I think that it would help to educate public and private sector funders (and people like me) if there were some clear statements of where and why open source drug R&D is likely to beat the conventional approach by a long way. I would frame this in terms of “discovery efficiency” before worrying about IP. If you can convincingly argue the approach is more efficient / productive, and for what specific kinds of activities or objectives it is more efficient / productive, then people with money will become more interested in how you overcome some of the obvious challenges (e.g., IP). If you can’t convince people it will be much more efficient/productive, then there is less point doing the work necessary to overcome the IP / funding problem.
To give a concrete example, I would guess that interactions with the drug regulator would be difficult to do in open source mode. If I read the Second Law of the open source principles to which the blog links, I see that “Anyone can take part at any level.” It would be hard to deal with the FDA or the EMEA on that basis. The FDA wants a named individual who it can send to jail if data are deliberately fabricated. Similarly, there will be manufacturing-related activities during the R&D process; preparing drug for clinical trials where open science would probably not help much.
Maybe the crisp and clear “open drug R&D – a guide for dinosaurs” already exists. If so, apologies. If not, write one.
(2) Disease / market choice
Following a couple of conversations with Mary Moran regarding new antibiotics and the AMR problem, I have become much more sensitive to the distinction between (A) diseases where a good product would, in principle at least, have a commercial market, and (B) diseases where a good product would still never sell very well. because the patients or their governments are too poor, or because the disease is so rare.
In class (A), charities and governments can support basic research, and may help manage the market, but one can often rely on private sector investment for a lot of the process. In class (B), it seems to be more trouble that it is worth to try to get the private sector involved. Therefore, if you are really doing (B), then there is no point creating complex schemes to pull in private sector/for-profit money. Conversely, if you are really doing (A), you need to be more careful about inadvertently disrupting for-profit efforts (see below).
(3) Unintended consequences
As an aside, I worked for a small private sector bio-informatics oriented biotech company in Oxford. We were trying to do what was effectively informatics-driven drug repurposing (or sometimes using re-purposing opportunities as “hits” from which to generate novel chemistry). In some ways, this appears to be a nice, low-cost, high-efficiency business model. We were using large publicly available chemo-proteomic data sets and protein-protein interaction data sets.
The experience made me sensitive to problems that current patent / IP law presents. The bioinformatics tools might identify something that looked interesting. One would then do a big Pubmed search, and find that a couple of lousy review articles written years ago had speculated (on the basis of other information) that drug class X might be good for disease Y. These lousy review articles would make us think very hard about whether or not we would be able to secure any IP because the patent office might regard them as “Prior Art”. This would tend to be a real disincentive for us to invest any further in the result.
In a similar vein, drug companies will sometimes publish stuff (either academically or in patent form), not because they think it will be useful to themselves, but because they know it will destroy the IP that their competitors are working on and will reduce competition. This is called “salting the field”.
One thing that I think people doing open access drug discovery need to think about is the fact that they may be salting the field some of the time, but that the collateral damage will be invisible to them. This is why I think Disease/market choice (above) is important. If you are working in a therapy area where there is unlikely to be much private sector interest, the risk is low. However, once you are working in an area where the public sector is interested as well, then the risk of inadvertently “salting the field” increases.
One way this can be managed, of course, is by managing the scope of the open science. If, for example, you produce “chemical probes” that might help people find Leads, but which are not Leads, then the risk is probably very low even if you are working in an area that has a lot of commercial interest (as per the Structural Genomics Consortium).
As you can probably tell, I am not an IP/patent expert, but I think this is a general area that requires real expert thought. Maybe I think it is more important than it really is, given the rather peculiar commercial niche that I was in (bioinformatics-oriented drug re-purposing)? None the less, I have heard other people in the drug industry talk about this problem; where an academic group publishes stuff and simultaneously destroys the IP that would have been required to bring that very stuff to market.
(4) Selling the idea to for-profit private sector investors
I have tried to think about selling the idea that Mat sets out to some of the cautious, clever, and critical fund managers that I used to deal with when I worked in investment. If I came to them with the idea as it is presented in the blog, I think I would struggle to part them with their money for the following reasons. Perhaps some of the questions / concerns that the fund managers would have can be answered in the next iteration?
Here are the kind of things that one or two of the more prickly fund managers would have asked me:
“Why the hell have you brought me this? You know I can’t invest in this kind of thing?”
This will be a very niche investment product. You need investors who are healthcare specialists, who can invest in illiquid assets, and who may have to be very flexible about the length of their investment horizon. These kind of people are probably identifiable, but they will be few and far between. Conventional venture capital, for example, typically has predefined holding periods and will need to exit (i.e., sell). It might be better to go after the “commercially-oriented philanthropists” (e.g., the Wellcome Trust translational outfit) and forget about the purely commercial investors.
“Aren’t you going to promise me a lot more than the prospect of simply getting my money back?”
Drug R&D has the economics of a lottery. No one buys lottery tickets hoping to “get their money back” because there is a good chance they will get nothing at all. The vast majority of drug R&D projects fail. 9/10 candidates entering clinical trials fail. Around half of the entire economic value of drug industry profits come from the top 10% of marketed drugs (i.e., the top 1% of drugs entering clinical trials) with half of all marketed drugs generating profits that are worth less than the average drug’s fully-allocated R&D cost (i.e., allocating the cost of failures). Therefore, if you cap per-project profits to a “reasonable” level, you destroy – in some peoples’ minds at least – the economics of the industry.
In my view, for what it is worth, most of the drug industry today is losing money on R&D and some Wall Street analysts agree with me on this. Therefore, if you go to commercial investors promising returns that are at face value going to be lower than the returns from an industry that already has poor returns on R&D, you will struggle to make a sale. If you are promising lower returns, you also need to promise much lower risk. You could only do this if you have a big slug of philanthropic or government money from the start, and the philanthropists or government promise to bail out the shareholders to a large extent if things go wrong. However, this starts to require a very complex investment vehicle.
“Can you come back and talk to me when you have lined up the money to pay me if it all works?”
I think the “critical path” that you would need to follow to pull in private sector investors looks long. So, for example, you would first have to get an agreement with whoever was going to cough up the cash at the end of the R&D process. Their commitment would have to be utterly convincing. Only when there was an obvious way of getting paid at the end of the process, would you get any kind of financial pull that would interest commercial investors.
Given the cost of late stage trials and typical failure rates, you need to get hundreds of millions of dollars committed as an absolute minimum. Not many single agencies will commit that kind of cash, which means you need to build some kind of buyers’ consortium, which will take years to do. Then, of course, R&D projects take 8+ years in clinical development, and 20+ years if you include the basic science. I think you would struggle to build the buyers’ consortium and even if you did build the consortium, private sector investors would be worried that priorities would change and by the time the drug was approved, the cash had been spent on something else.
“Can I get my money out early if it all looks hopeless?”
Mixing private sector money that is chasing financial returns with philanthropic money that is seeking to make the world a better place may cause governance problems. What if, half way through the project you find out that your drug is bad for malaria (a lousy commercial market) but good for, say, male pattern baldness (a great commercial market). Do you ditch the malaria plan and go hell for leather for baldness? But maybe you can’t because if it is open source so there is no IP. What if another great malaria drug beats you to market, and you are worried that governments will no longer honour the agreement on data exclusivity and cost recovery because they no longer need your drug as much. The private sector shareholders may want to wind up the project and get their remaining cash out before it is spent, but the philanthropists would want to push on.
“Who is in charge if we find something that will be useful for some big commercial market?”
A similar but slightly different point is that a standard commercial R&D project has all sorts of option value, because you can chop and change the objectives to match the emerging science or changing market conditions. An open innovation project with some kind of long term commitment from governments or public sector bodies around endpoints / regulatory approval for malaria or another disease has none of this kind of option value, because you can’t easily repurpose the science if there looks to be a better opportunity to address. This is true even if there turns out to be science that could have been turned into IP for a commercial market.
This kind of on-the-fly repurposing is very common. Many of the world’s most successful drugs or drug classes were invented for something else (e.g. the anti-TNFs, Viagra, SFUs for diabetes, MAO inhibitors for depression, the anti-VEGF drugs used in AMD, etc. etc.). Because you have tied payment to a specific outcome, many years in advance, while at the same time foregoing opportunities to re-purpose the project if other outcomes look better, the “option value” is much lower than a typical R&D project.