Nature Reviews Drug Discovery has an article on the current state of drug development, looking at what's expected to be launched from 2012 to 2016. There's a lot of interesting information, but this is the sentence that brought me up short: "the global pipeline has stopped growing". The total number of known projects in the drug industry (preclinical to Phase III) now appears to have peaked in 2009, at just over 7700. It's now down to 7400, and the biggest declines are in the early stages, so the trend is going to continue for a while.
But before we all hit the panic button, it looks like this is a somewhat artificial decline, since it was based on an artificial peak. In 2006, the benchmark year for the 2007-2011 cohort of launched drugs, there were only about 6100 projects going. I'm not sure what led to the rise over the next three years after that, but we're still running higher. So while I can't say that it's healthy that the number of projects has been declining, we may be largely looking at some sort of artifact in the data. Worth keeping an eye on.
And the authors go on to say that this larger number of new projects, compared to the previous five-year period, should in fact lead to a slight rise in the number of new drugs approved, even if you assume that the success rates drop off a bit. They're guessing 30 to 35 launches per year, well above the post-2000 average. Peak sales for these new products, though, are probably not going to match the historical highs, so that needs to be taken into account.
More data: the coming cohort of new drugs is expected to be a bit more profitable, and a bit more heavily weighted towards small molecules rather than biologics. Two-thirds of the revenues from this coming group are expected to be from drugs that are already in some sort of partnership arrangement, and you'd have to think that this number will increase further for the later-blooming candidates. The go-it-alone blockbuster compound really does seem to be a relative rarity - the complexity and cost of large clinical trials, and the worldwide regulatory and marketing landscape have seen to that.
As for therapeutic area, oncology has the highest number of compounds in development (26% of them as of 2011). It's to the point that the authors wonder if there's an "oncology bubble" on the way, since there are between 2 and 3 compounds chasing each major oncology target. Personally, I think that these compounds are probably still varied enough to make places for themselves, considering the wildly heterogeneous nature of the market. But it's going to be a messy process, figuring out what compounds are useful for which cases.
So in the near term, overall, it looks like things are going to hold together. Past that five-year mark, though, predictions get fuzzier, and the ten-year situation is impossible to forecast at all. That, in fact, is going to be up to those of us doing early research. The shape we're in by that time will be determined, perhaps, by what we go out into the labs and do today. I have a tool compound to work up, to validate (I hope) an early assay, and another project to pay attention to this afternoon. 2022 is happening now.
Update: here are John LaMattina's thoughts on this analysis, asking about some things that may not have been taken into account.