When Jared Heyman’s sister, Carly, was diagnosed with a rare illness that affects one in 15,000 women, she could finally get on with her life. For three years, she had been stuck in bed, watching TV and playing Solitaire, while two dozen doctors puzzled over what was wrong with her. At least now she could get the right treatment.
But her brother wondered if there might have been a more efficient road to diagnosis. Carly had spent more than $100,000 while she waited—hardly an endorsement of the traditional medical system.
Heyman’s alternative to the standard way people get diagnosis is to substitute the insight of single physicians for the wisdom of the crowd. People who are struggling for medical answers can post a "case" at CrowdMed, and hope that more than a hundred "medical detectives" can help.
As a test, he recently put up Carly’s symptoms and history, and the crowd cracked the case in three days, and at little cost.
Crowdmed’s detectives range from Stanford medical students to online workers at freelance marketplace Mechanical Turk. It incentivizes participation by awarding points. Detectives get more credit for positive diagnoses, and can award points to other detectives, betting on their theories. From the total submissions, the site works out three most likely diagnoses, which patients can then take to their physician.
To post a case, patients pay $199, plus a small deposit to get them to come back to site to report whether the diagnosis was correct. If they’re not happy, CrowdMed is currently offering a money-back guarantee.
Heyman admits some doctors may resent the idea of crowd-based diagnosis, seeing it as undervaluing years of training. But he reckons many won’t see it that way. "Some doctors say medicine should only be practiced by highly trained professionals. But others say they have patients who know more about their disease than they do."
The site launched in April, at TEDMED. So, it’s too early to say if it will work really. But a recent trial found that 700 participants diagnosed 20 rare conditions among their three guesses.
Heyman says the system minimizes the effect of people who really don’t know what they’re talking about. It gives more weight to participants with higher scores, reinforcing their expertise. And, he notes, people are more likely to try and crack a case if they’re likely to win. If the question is out their area, they’ll probably sit that case out.
He hopes the medical community will embrace CrowdMed as educational, and a bit of fun.
"We designed it to be gamified, engaging, and interactive. We know that a lot of medical students enjoy medical mysteries. They like learning about cases, because there’s an intellectual challenge to it."