Regular reader Qetzal pointed out this analysis of the drug industry's problems, and it's well worth reading. Enoch Huang is a director of molecular informatics for Pfizer, and it's a sign that we're well into the 21st century that his job title doesn't sound odd to me. He was speaking as part of a panel at Harvard Business School, and quoth Huang:

"The number one enemy facing our industry isn't so much Canadian importation or possible regulation on price. It's our own drug candidate attrition. Within Pfizer, and I think it's representative of the industry, the odds of a clinical candidate - that's when you're done making the molecule and you're sending it off to see whether it works and is safe and can make approval - is 1 in 25. That's something like 96 percent failure rate. It's staggering number, [when] coupled with the cost of R&D versus the productivity measured in new chemical entities, which is essentially flat. It's an unsustainable business model for Pfizer and the industry. . ."

Well, I spent yesterday's post beating up on Pfizer, so I'll just note that "unsustainable" is a word that comes often to mind when I think about the company. But Huang is right about the problem and its implications for the industry. Note, though, that his figures are even worse than the historical ones that I cited here back in September, and that he attributes most failures to toxicity, rather than lack of efficacy - I think that over the years these categories have switched places. But all that aside, it's hard to see how we can go on like this, and the failure of our new technologies to reduce these odds is especially galling. (Here's a poll of researchers on these very subjects.)

This takes us back to a point raised in a comment to yesterday's post. There are analysts saying that not only does the drug industry spend too much on marketing, but it spends too much on R&D. Well, as a researcher, my answer to how much a drug company should be spending on research is "as much as they can possibly stand," but that's not as facetious as it sounds. Because of the wasting-asset feature of patent protection, we absolutely have to discover new things all the time, and then try to get them out on the market and sold to someone. Cutting R&D, unless you have a really, really good idea of where you're now wasting the money, would be an act of desperation.

But it's for sure that we're wasting a lot, and I hope we eventually find out where we're doing it. Huang mentions that the reductionist approach to drug discovery itself may have gone too far. That's the standard industry way, and has been for years now: coming up with well-defined molecular targets, screening them in isolated in vitro systems, and working your way up from there. It's intellectually appealing, and has led to some huge successes, but we may have ridden that horse about as far as it can carry us.

The hope for genomics (and the other -omics) has been that they would throw new logs on the reductionist fire. If we just understood things better, the hope has been, then we could pick off relevant drug targets with the systems we already have. The industry has already spend vast amounts on this project, but we may now have to face spending even more to switch to a different sort of system entirely.

What would that look like? This gets back to the discussions of animal models I've posted on recently, and the line of thought is pretty messy. If we're going to get away from reductionism, then the closer to the real living system we are, the better. We'd need better animal models (no easy task), and we'd need to get compounds into them as quickly as we could. And from there, we'd need to get into humans as quickly as possible, too, on the same principle. But to do that, we're going to need to know more about toxicology, because otherwise we're going to to either spend a huge amount running our current tox models in animals, or run unacceptable risks in early clinical trials.

To that point, a recent article in Nature Reviews: Drug Discovery, one of their series on animal models, said that we shouldn't be afraid of using sheer brute force to test our way through these things. It may come to that, but that's going to take some sheer brute cash.