Crowdsourcing the Diagnosis

crowdsourcingRegular readers of the New York Times Magazine are familiar with Diagnosis, Dr. Lisa Sanders’ long-running column of medical mysteries. With false leads, twists, and surprise endings, these hospital narratives are typically as suspenseful as a Raymond Carver story, or perhaps more appropriately, an episode of House. The subject matter of these mysteries are usually what doctors call “fascinomas,” unusual and odd maladies that one rarely, if ever, encounters in the clinic. But in this past Sunday’s installment, which featured the Medical Center’s John Henning Schumann, how a particular fascinoma was diagnosed was just as interesting as the diagnosis itself.

The story, in a nutshell: a 40-year-old friend of Schumann’s was suffering nightly fevers of unknown origin, and doctors were stumped. A multitude of tests found nothing out of sorts, save a non-cancerous mass in the patient’s liver that appeared to be a hemangioma – a cluster of blood vessels. Hemangiomas don’t typically cause fevers, so the patient’s doctors were unsure whether to perform the major surgery required to remove the mass from the patient.

And then: the internet! Amid his doctors’ uncertainty, that patient (an internet expert) and his significant other decided to start a blog about the medical mystery in which he found himself the central character.

“They had a huge social network of very over-educated people, many doctors, many not, but people connected to doctors,” said Schumann, an assistant professor of medicine. “Their attitude was let’s put it out there and see what they think.”

Schumann, a concerned observer from afar, posted a link to his friend’s blog on his own blog, Glass Hospital, and the story was picked up by Kevin Pho, founder of the popular medical blog KevinMD. Pho’s post drove traffic to the patient’s blog, where doctors left comments with diagnosis suggestions and links to published obscure case reports similar to his ordeal.

The commenters’ consensus – that the hemangioma was the likely cause of the symptoms and should be removed – agreed with the eventual, independent assessment of the patient’s doctors. But Schumann said in an interview with ScienceLife that the process was still helpful to his friend, if mostly in a psychological sense.

“I don’t think it necessarily solved the case, but I think it might have expedited it slightly, and it provided a ray of hope,” Schumann said. “For a guy who is web-savvy, it gave him a level of comfort that I don’t think all the medical specialists in the world could provide.”

It also provided the happy ending that the Diagnosis column needed, as the patient fully recovered and returned to normal after having part of his liver removed. But constrained by her column’s format, Sanders doesn’t spend much time talking about the interesting implications of the patient’s unusual internet experiment – could this kind of crowd-sourced diagnosis be a tool for other fascinomas encountered in the clinic?

Schumann said he sees both the potential of the approach and the drawbacks. Hospitals have long used “tumor boards” – meetings of oncologists, surgeons, radiologists, and pathologists – to determine the best treatment course for individual cancer patients. In a way, the blog experiment was a digital version of a “tumor board” thrown open to a worldwide community of participants.

“It’s not a new idea, it’s just the exponential power of bringing in people from around the world and crossing institutional boundaries,” Schumann said. “He had what amounted to a fairly rare anecdotal thing with limited data, and an internist in Pennsylvania happened to see a case of this, pulled an article from a Korean medical journal, and there we were.”

But a patient’s own research and a second opinion – never mind a third, fourth, or tenth opinion – can complicate a patient’s care. The most exotic diagnosis is only rarely the correct diagnosis, and test results on a blog aren’t a perfect substitute for an in-person evaluation. On the KevinMD post, commenter tgottsdo sums up how this can cause problems for those directly caring for the patient: “I think its an interesting experiment although I hope it doesn’t catch on. I would hate to have to walk into every patient’s room to confront that days opinions from around the world.”

Still, as electronic health records become reality, sharing such information (within the bounds of privacy considerations, of course) across institutional and geographical borders could become easier and more commonplace. Who needs Dr. House when you have the hivemind?

“I think it demonstrated the vast potential of crowd-sourcing and social networking,” Schumann said. “There’s something to be said for the wisdom of crowds and this is a great example.”

About Rob Mitchum (526 Articles)

Rob Mitchum is communications manager at the Computation Institute, a joint initiative between The University of Chicago and Argonne National Laboratory.

%d bloggers like this: