Can you crowdsource medical diagnosis?
Crowdsourcing has never been more popular. My last blog showcased the latest NASA Hackathon that saw thousands come together to offer solutions to scientific problems. The wisdom of crowds was also incredibly prominent during the aftermath of the Boston bombing, with people from around the world looking over photos from the event in an attempt to track down the bombers.
Wonderful though these things are, would you want to trust the crowd with your health? New start-up CrowdMed aims to find out.
The site, which was launched last week, aims to apply the wisdom of crowds to our medical problems. The concept is a fascinating one. You don’t have to be medically trained to join the site, with the only criteria being a willingness to investigate cases.
As with many other crowd based services, users will be rated and given points through their use of the site. These points act as a sort of virtual currency that can be bet on the correct diagnosis from a list of suggestions. These bets form the basis of a prediction market, with the suggested diagnosis rising and falling in price, just like stocks on the stock exchange. The site then uses algorithms to calculate the probability that each diagnosis is correct.
Once the crowd have arrived at their consensus, the CrowdMed site shows patients the top three suggestions, which they are then free to explore with their doctor. If the diagnosis proves correct, the doctor is encouraged to feed the results back into the site, with the participants rewarded accordingly.
The results thus far look promising. Participants have ‘diagnosed’ 20 test cases with a good degree of accuracy. As numbers increase from the 700 or so people currently using the site, the owners expect accuracy to increase.
Whilst there will hopefully be mainstream applications for the site eventually, it’s initial aim is to target around 7,000 rare diseases. These alone however have around 30 million victims in Europe alone, with some 40% of these people either going undiagnosed or having a misdiagnosis.
Crowdsourcing meets big data
Could the future see heavy duty computing such as IBM’s Watson teaming up with the crowdsourcing of CrowdMed? It’s something IBM are particularly keen on. Claudia Perlich, who helped develop Watson, imagines it teaming up with CrowdMed to work on tough diagnoses.
“If Watson could get hold of what people submit to CrowdMed, I would love it,” she says. “I absolutely agree with the premise that the big problem of the medical system is that we don’t have sufficient information sharing.”