Prizes in public health -- the new kaggle?

This morning I saw a short piece by James Surowiecki on the absence of a vaccination for the Ebola virus. Surowiecki points out that the incentive structure for pharmaceutical companies rewards work on drugs that are likely to be taken by i) a large number of people, ii) Westerners, and iii) are likely to be taken over a long period of time. This means, Surowiecki argues, that under the present incentive scheme we are unlikely to develop cures for things like Ebola, which is essentially confined to the developing world, and up until recently did not affect many people.

As a possible solution, Surowiecki offers 'prizes' -- sponsored by governments -- which compensate firms in exchange for the right to manufacture the resulting pharmaceuticals. Put differently, the government can intervene on the market value of development in these areas by paying 'more than the going rate.'

I think Surowiecki is surely right about the potential benefits of this sort of approach -- paying companies more will definitely give them an incentive to change their research priorities. Thinking about all of this, though, reminded me of the kaggle-style data competitions that are growing increasingly popular. Here, a problem is posted (sometimes with a large prize for the best solution), and data scientists of all stripe work on solutions. The competition that really vaulted these into the mainstream was the Netflix Prize -- offering $1 Million for the best improvement in predictive film ratings. I remember going to a talk a few years ago by the 'BellKor' team that ended up winning the competition, where the winners remarked that while the prize seemed big, the tremendous time the team put into the challenge meant that they were actually working at a really low wage-rate.. and they were the one's who actually won!

I bring all this up because it seems that, at least in the case of data competitions, it's not really the money that's driving entry (or work) on these problems. The competition, collaboration, and social value of doing well seem like much more important causes. Now, data science is quite different from pharmaceutical research -- the start up costs are MUCH lower, and it can be a much more individual activity -- but after reading Surowiecki's article, I'm especially curious to see whether some of the non-monitary incentives that we're seeing at work in the data science world might emerge if public health adopts a similar incentive structure around sponsored prizes.