Crowdsourcing

Interesting post today by Matthew Holt over at THCB. He interviewed Alexandra Carmichael, CEO of CureTogether.com, who claims to be at the cutting edge of something called Open Source Health Research.

Her company is trying to leverage the collective knowledge of disease communities to discover new connections between symptom patterns and underlying diagnoses. For example, after the jump you can see a graph generated by CureTogether.com's user surveys that illustrates the most frequent co-morbidities reported with vulvodynia.

Could information like this one day be useful in clinical decision-making? Perhaps, though it would be terribly difficult to interpret. How's a doctor supposed to tease apart co-morbidities that are spuriously linked from those that represent real clues as to common etiology? Is it ethical to make clinical decisions on the basis of uncorroborated Internet surveys?

It's my hope that generic user surveys like this get tied to more robust analysis, including an attempt to tease out confounders (it sounds like CureTogether.com is trying to do this). Currently the population sizes in question are probably far too small to drive any significant outcomes analysis, but expect to see more sites like this spring up in the coming years...

The real value may be in the sense of 'working towards a cure' that it gives to patients. Here's a money passage:

Me: Which kinds of takes us back to your prior life in genomics and ideas about GeneTwist. Care to speculate about the actual ways that CureTogether's data and community will actually lead to cures, together?

Alexandra: Yes, the grander vision for down the road is to integrate all kinds of data - genome scans, biomarker tests, streaming data from wifi health tracking gadgets. Collaboration with the larger research community as well, integrating our data with traditional clinical trial data. We already have researchers knocking on our doors wanting to start analyzing the data.

Me: So now "all" you have to do is get more people and more data into the site. How are you going to do that?

Alexandra: That's a good question! By talking to people like you, for a start. Health and condition-specific bloggers seem happy to help spread the word. Disease foundations have also approached us, eager to help out. And media have all been coming to us so far. I just started the process of seriously getting out there in the past couple of weeks - I'm even on Twitter @accarmichael. It's still mostly a grass-roots effort though. If and when we get funding, that could change. But it's very rewarding work. At the end of the day, I know I've spent my time helping people and working to make a difference.

Reading her comments, it strikes me that we face a similar challenge at ClinicalTrialsWiki.org in trying to develop a core community of users who are motivated to populate our site with discussion and commentary.