Upon reading the story in MobiHealthNews this week about CrowdMed, the startup that relies on crowdsourcing to suggest possible diagnoses to patients with rare or mysterious conditions, I had two thoughts: clinical decision support and Dr. Gregory House.
Dr. House, the fictional diagnostic genius from the dearly departed Fox show, "House M.D.," crowdsourced medical decision-making to his team of interns, underlings and foils for his practical jokes, ostensibly to teach them to be better diagnosticians. It made for great television for most of its eight years on the air, but it's really a terrible way to practice medicine, essentially making educated guesses to develop differential diagnoses.
Though I loved the show for its humor, drama and complex characters – not to mention House's obvious parallels to Sherlock Holmes – as a health IT reporter, I knew there was an Achilles' heel. Give House a clinical decision support system and he solves his cases in a matter of minutes. No more guessing, no more unnecessary testing, no more questionable treatments that cause adverse reactions and make ailing patients even sicker. And you would have a boring show.
Unfortunately, the guesswork and overtreatment are all too common not only in a TV medical drama, but in real hospitals all across the industrialized world. I fear CrowdMed is just going to perpetuate the problem, the financial backing by Y Combinator and presence at TEDMED notwithstanding. The same goes for Grand Roundtable, a crowdsourcing platform for medical treatment ideas, part of the inaugural class of health incubator DreamIt Health.
The "wisdom of crowds," as CrowdMed describes it, might work when searching for doctors, as Blueprint Health-supported MyNewMD offers, or for analyzing lots of data, but if I'm the one on the exam table, I don't want a bunch of docs guessing over the internet. I want the right answer.
Clinical decision support systems sift through massive medical databases to suggest courses of treatment. Some focus specifically on diagnostic decision support, helping physicians match symptoms to create differential diagnoses. Isabel Healthcare comes to mind here, as does Dr. Larry Weed, who has spent the last six decades or so advocating computer-aided matching of patient problems to medical knowledge that isn't always apparent to a physician used to seeing a small set of common diseases.
Sure, CDS has its drawbacks. Systems don't always fit physician workflows. Doctors get overwhelmed with "alert fatigue" if the computer keeps sending repetitive warnings, and they start tuning the alerts out. And CDS, particularly for diagnostic purposes, needs a thorough patient history to be effective.
If I am the one being diagnosed, I'd rather spend the time filling out a complete history than having a bunch of doctors play guessing games. It doesn't matter if it's at House's Princeton-Plainsboro Teaching Hospital or over the internet.
