Sunday, January 12, 2014

Megafund versus Megalosaurus: Funding Drug Development

This new 10-minute TEDMED talk is getting quite a bit of attention:

 (if embedded video does not work, try the TED site itself.)

In it, Roger Stein claims to have created an approach to advancing drugs through clinical trials that will "fundamentally change the way research for cancer and lots of other things gets done".

Because the costs of bringing a drug to market are so high, time from discovery to marketing is so long, and the chances of success of any individual drug are so grim, betting on any individual drug is foolish, according to Stein. Instead, risks for a large number of potential assets should be pooled, with the eventual winners paying for the losers.

To do this, Stein proposes what he calls a "megafund" - a large collection of assets (candidate therapies). Through some modeling and simulations, Stein suggests some of the qualities of an ideal megafund: it would need in the neighborhood of $3-15 billion to acquire and manage 80-150 drugs. A fund of this size and with these assets would be able to provide an equity yield of about 12%, which would be "right in the investment sweet spot of pension funds and 401(k) plans".

Here's what I find striking about those numbers: let's compare Stein's Megafund to everyone's favorite Megalosaurus, the old-fashioned Big Pharma dinosaur sometimes known as Pfizer:

$3-15 billion
$9 billion estimated 2013 R&D spend
81 (in pipeline, plus many more in preclinical)
Return on Equity
12% (estimated)
9.2% (last 10 years) to 13.2% (last 5)
Since Pfizer's a dinosaur, it can't possibly compete with
the sleek, modern Megafund, right? Right?

These numbers look remarkably similar. Pfizer - and a number of its peers - are spending Megafund-sized budget each year to shepherd through a Megafund-sized number of compounds. (Note many of Pfizer's peers have substantially fewer drugs in their published pipelines, but they own many times more compounds - the pipeline is just the drugs what they've elected to file an IND on.)

What am I missing here? I understand that a fund is not a company, and there may be some benefits to decoupling asset management decisions from actual operations, but this won't be a tremendous gain, and would presumably be at least partially offset by increased transaction costs (Megafund has to source, contract, manage, and audit vendors to design and run all its trials, after all, and I don't know why I'd think it could do that any more cheaply than Big Pharma can). And having a giant drug pipeline's go/no go decisions made by "financial engineers" rather than pharma industry folks would seem like a scenario that's only really seen as an upgrade by the financial engineers themselves.

A tweet from V.S. Schulz pointed me to a post on Derek Lowe's In the Pipeline blog. which lead to a link to this paper by Stein and 2 others in Nature Biotechnology from a year and a half ago. The authors spend most of their time differentiating themselves from other structures in the technical, financial details rather than explaining why megafund would work better at finding new drugs. However, they definitely think this is qualitatively different from existing pharma companies, and offer a couple reasons. First,
[D]ebt financing can be structured to be more “patient” than private or public equity by specifying longer maturities; 10- to 20-year maturities are not atypical for corporate bonds. ... Such long horizons contrast sharply with the considerably shorter horizons of venture capitalists, and the even shorter quarterly earnings cycle and intra-daily price fluctuations faced by public companies.
I'm not sure where this line of though is coming from. Certainly all big pharma companies' plans extend decades into the future - there may be quarterly earnings reports to file, but that's a force exerted far more on sales and marketing teams than on drug development. The financing of pharmaceutical development is already extremely long term.

Even in the venture-backed world, Stein and team are wrong if they believe there is pervasive pressure to magically deliver drugs in record time. Investors and biotech management are both keenly aware of the tradeoffs between speed and regulatory success. Even this week's came-from-nowhere Cinderella story, Intercept Pharmaceuticals, was founded with venture money over a decade ago - these "longer maturities" are standard issue in biotech. We aren't making iPhone apps here, guys.

Although big pharma companies are central to the later stages of drug development and the marketing and distributing of approved drugs, they do not currently play as active a role at the riskier preclinical and early stages of development
Again, I'm unsure why this is supposed to be so. Of Pfizer's 81 pipeline compounds, 55 are in Phase 1 or 2 - a ratio that's pretty heavy on early, risky project, and that's not too different from industry as a whole. Pfizer does not publish data on the number of compounds it currently has undergoing preclinical testing, but there's no clear reason I can think of to assume it's a small number.

So, is Megafund truly a revolutionary idea, or is it basically a mathematical deck-chair-rearrangement for the "efficiencies of scale" behemoths we've already got?

[Image: the world's first known dino, Megalosaurus, via Wikipedia.]


CBT said...

I applaud the attempt by Roger et al in terms of exploring innovative approaches to address clinical trial backlogs. The model relies heavily on the approval of blockbusters, and although I am sure it is possible to identify 150-200 such candidates, I highly doubt you could find that many in cancer. Of the top 10 cancer selling drugs in the United States, only Rituxan and Avastin have revenues in excess of $2 billion per year. Further to your comparison with Pfizer, its oncology division only accounts for 3% of revenue. This is to be expected given our increasingly personalized understanding of cancer treatment


Paul Ivsin said...

Thanks for the comment! It's interesting that the model requires blockbusters - do you have any sense of what percentage of the portfolio would need to hit it big?

I did notice that the megafund apparently assumes short-term revenue by selling off some of its early successes. That would be a good topic of discussion - I wonder if the authors have any idea how hard it is to get any value from early stage compounds?

CBT said...

My comment re: blockbusters was predicated on the following publication

In Section I, the authors hypothesize a drug costing $200 mil to develop with a 5% chance of approval in 10 years, generating $2 billion in income over the next ten years. So, from a pool of 150 candidates (i.e. $3 bil fund), about 8 would have to be approved