Thanks to Michael Millenson, TDWI writer for sending us this post from Slate.com
By Darshak Sanghavi
(published on slate.com on Oct. 6, 2010, at 3:27 PM ET)
Sometime in 1995, an e-mail from China arrived in my inbox with a desperate request for medical advice. I was a naïve medical student at Johns Hopkins University and an early adopter of the modem; the e-mail’s author was identified only as “Peking University.” In broken English, the message described a 21-year-old woman who had felt sick to her stomach and within days lost all her hair. This problem went away, but a few months later, “She Began to facial paralysis, central muscle of eye’s paralysis, self-controlled respiration disappeared,” and needed to be put on a ventilator. “This is the first time that Chinese try to find help from Internet,” the message explained. “Please send back e-mail to us.” With immature confidence I consulted some texts and replied that maybe she had a weird form of lupus. I never heard back and figured it was a prank.
The following year at the supermarket, I was browsing the August issue of Reader’s Digest and saw a piece titled “Rescue on the Internet.” It turned out that I wasn’t the only one who’d replied to the posting, and the whole thing had not been a hoax. Incredibly, hundreds of doctors had seen the brief message and correctly determined that the patient was being poisoned by a tasteless, odorless heavy metal called thallium. Soon after, Chinese doctors were able to give an antidote to save the woman’s life. (She did end up permanently disabled.)
In 2006, Wired magazine coined the term “crowdsourcing,” to describe the process of seeking a problem’s solution from a wide community, often online. Such collaboration certainly didn’t come naturally to doctors. For millennia, they’d worked as solo practitioners who jealously guarded their secrets. But the Chinese e-mail episode shows how large groups of doctors might come together to solve a problem. More than 1,000 trained medical professionals independently guessed at the cause of the woman’s illness, and while many were wrong, almost one-third suspected thallium poisoning. That was enough to get her doctors in China to consider the possibility and then confirm it.
In many ways, such crowdsourcing resembles an expert poll—sort of like Trident’s claim that “four out of five dentists” recommend sugarless gum, minus the commercial bias. And it works. One example: the New England Journal of Medicine hosts a weekly, online “image challenge,” which shows mysterious x-rays or biopsies and asks for a diagnosis via multiple-choice. I don’t often pick the right answer, but if you look at the aggregate choices from tens of thousands of other doctors around the world, the plurality invariably hits the mark.
Doctors in the United States have found help from their peers for real-life situations, too. Earlier this year, an internist (and sometime Slate author) named John Schumann posted the details of a peculiar case to a widely-read medical blog. A friend of his had developed inexplicable weight loss, low blood counts, and a weird, softball-sized mass in his liver. A dozen doctors saw the MRI scans and other test results online, and about half came up with the correct diagnosis: a benign growth of blood vessels.
Debunking the myth of the lone maverick, health researchers suggest that groups of doctors outperform individuals not only in diagnosing problems but also in treating them. In 2007, NEJM began polling its readers for consensus opinions on tough treatment questions: how to handle an abnormal prostate cancer screening test, how to treat an athlete’s skin infection, what to do for hepatitis C infection, and a few others. Responses from almost 20,000 doctors across the globe were tabulated.
Still, patients may be uncomfortable turning over their care to a majority vote of faceless doctors. But the reality doesn’t have to sound so scary and impersonal. In some cases, crowdsourcing works by soliciting many ideas, and then having designated subspecialists vet them. (As on Wikipedia, not all contributors are given equal authority.) At major medical centers, for example, a group of experts called a “tumor board” reviews tough cancer cases as a group, e-mails around the country for advice on the hardest ones, and then discusses the options with the individual patient. In my own field of pediatric cardiology, more than 1,500 doctors worldwide subscribe to an e-mail list for crowd-sourcing tough issues, but the final decisions are left to the cardiologist who knows the patient.
Doctors can draw on group consults when necessary, but nonphysicians rarely have access to the power of crowd-sourced medicine. To be sure, many patients with complex or poorly understood medical problems likeamyotrophic lateral sclerosis congregate in large virtual communities such as PatientsLikeMe, where they share details of their medical treatments and symptoms with each other—and occasionally even launch their own unregulated and informal drug trials. These communities provide some helpful information and support for many people.
But few medical professionals are involved in patient sites; mostly patients just talk amongst themselves. That’s a problem since crowdsourcing works best when expertise is widely distributed. After all, it won’t help you much to have 1,000 people reading your mammogram or selecting your chemotherapy if they’re all well-meaning amateurs. What you’re really looking for is the consensus of 1,000 highly trained specialists. A patient working alone might be able to use the Internet to get a handful of credible second opinions, and nothing more.
It may be hard to get hordes of doctors to give you their personal attention, but you can often do the next best thing: Find out how hordes of doctors would help someone with a problem just like yours. In fact, the federal government and various groups like the American Academy of Pediatrics convene huge panels of experts to address medical issues all the time. These expert panels put out an electronic call for the world’s medical literature on a problem—often including dozens if not hundreds of studies—and then attempt to consolidate the medical wisdom. To date, such entities have crowdsourced 2,527 problems, and produced public reports outlining black-and-white guidelines for diagnosis and treatment.
Want to know what an enormous crowd of doctors thinks about heartburn? It’s here. Has your child been diagnosed with asthma? In 2007, the National Heart, Lung, and Blood Institute collected the opinions of the nation’s best asthma doctors here. What’s the deal with head lice? Get the facts here. Cancer treatments? Check. Diagnosed with a weird genetic condition? The world’s experts have combined their talents and created adatabase for you. Every problem one can imagine—ingrown toenails, hoarseness, sinus infections, weight loss methods, and so on—are addressed.
Having gone through the trouble of crowdsourcing so many questions, doctors have a great resource at their disposal. But they routinely fail to use the help. Some doctors may honestly think the crowd is wrong, butmore likely, they’re unaware of the fact that guidelines exist or they’re wedded to outmoded practices.
What does all this mean for patients? First, check online to see if there are published guidelines for treating your problem. (Many times, your doctor will help you find them.) That will give you at least a general consensus among doctors in the field. Second, if you have a confusing problem, and your doctor is as baffled as you are, ask if there is an expert case conference (an actual crowd of doctors) or a large Internet discussion list (a virtual crowd) that he or she might consult. And if all else fails, maybe you too can send an e-mail blast out to everyone you can find. Someone out there just might have the answer for you.
Darshak Sanghavi is Slate‘s health care columnist. He is chief of pediatric cardiology and assistant professor of pediatrics at the University of Massachusetts Medical School as well as the author of A Map of the Child: A Pediatrician’s Tour of the Body. Follow him on Twitter.
Article URL: http://www.slate.com/id/2270169/