I'm of two minds on this New York Times article on Alzheimer's research. It details some recent progress on biomarkers for the disease, and that work does look to be useful. A lot of people have proposed diagnostics and markers for Alzheimer's and its progression over the years, but none of them have really panned out. If these do, that's something we haven't had before.

But my first problem is something we were talking about here the other day. Biomarkers are not necessarily going to help you in drug development, not unless they're very well validated indeed. We really do need them in Alzheimer's research, because the disease progression is so slow. And this effort is really the only way to find such things - a good-sized patient sample, followed over many years. But unfortunately, 800 people (divided out into different patient populations) may or may not be enough, statistically. We're now going to have to take the potential assays and markers that this work has brought up and see how well they work on larger populations - that's the only way that they'll be solid enough to commit a clinical trial to them. Both the companies developing drugs and the regulatory agencies will have to see convincing numbers.

That general biomarker problem is something we really can't do anything about; the only cures are time, effort, money, and statistical power. So it's not a problem peculiar to Alzheimer's (although that's a tough proving ground), or to this collaborative effort. But now we come to the collaborative effort part. . .overall, I think that these sorts of things are good. (This gets back to the discussions about open-source drug discovery we've been having here). Bigger problems need sheer manpower, and smaller ones can always benefit from other sets of eyes on them.

The way that this Alzheimer's work puts all the data out into the open actually helps with that latter effect. All sorts of people can dig through the data set, try out their hypotheses, and see what they get. But I think it's important to realize that this is where the benefit comes from. What I don't want is for people to come away thinking that the answer is that we need One Big Centralized Effort to solve these things.

My problem with the OBCE model, if I can give it an acronym, is that it tends to cut back on the number of ideas and hypotheses advanced. Big teams under one management structure don't tend to work out well when they're split up all over the place. There's managerial (and psychological) pressure, from all directions, to get everyone on the same idea, to really get in and push that one forward with all the resources. This is why I worry about all the consolidation in the drug industry: fewer different approaches get an airing when it's all under the roof of one big company.

So this Alzheimer's work is just the sort of collaboration I can admire: working on a big problem, sharing the data, and leaving things open so that everyone with an idea can have a crack at it. I just hope that people don't get the wrong idea.