Ikea furniture is not the best furniture in the world. Indeed, the quality is considerably lower than that of the kind of Nordic modernist stuff that it’s imitating. But rise of Ikea has been a good thing for human quality of life. Sometimes furniture that’s almost as good and much cheaper is exactly what you want. On the health care front, however, I often ask myself if that’s something that we, as a society, would be willing to accept. Suppose someone invented a device that, when used properly at home, could diagnose a range of illness about 90 percent as well as an average general practitioner at a much lower cost. Would a device like that be legal? Would it be brought to market? Or would regulations and fear of lawsuits shut it down. After all, it’s not safe to use the DiagnoMachine rather than see a doctor. Or, at least, it’s not as safe especially when you consider that untrained civilians might use it wrong.
But by the same token, if you can’t bring a device like that to market, it’s going to be hard to generate the kind of competition that over time will raise the accuracy and user-friendliness of DiagnoMachines. The earliest MP3 players were pretty crappy, after all, but their existence helped bring forth the original iPod, which itself was a pretty sorry device compared to its descendants.
This is what I was reminded of when I read Michael Mandel’s piece (PDF) on MelaFind, “a handheld computer vision device intended to help dermatologists decide which suspicious skin lesions should be biopsied for potential melanoma.” The FDA wouldn’t approve the device because:
— The device did not do better than the experienced dermatologists in the study (“the FDA review team does not believe this is a clinically significant difference between MelaFind and the examining dermatologist.”)
— The device was tested on lesions identified by experienced dermatologists, not on the broader set of lesions that might be identified by “physicians less experienced than these dermatologists.”
— The device did not find every melanoma in the sample (“Since the device is not 100% sensitive, if use based on the device’s diagnostic performance reduces the number of biopsies taken, harm could ensue
in the form of missed melanomas.”)
— The device was not demonstrated to make inexperienced physicians the equal of experienced dermatologists (“Currently, formal training is offered to physicians to become board certified dermatologist and thus be able to diagnose clinically atypical lesions. The FDA review team would have to compare this board certification training to that offered by the sponsor to those physicians operating MelaFind to determine if it is found adequate.”)
To me the important thing about this is that the objections are basically cogent. It seems like you really are better off with the professional judgment of an experienced dermatologist than you would be with a less-trained doctor trying to see patients quickly by using MelaFind as a speedy low-cost screening device. That said, if you care about cost — about the capacity of a community health center to serve as many clients as possible within a fixed budget constraint — then none of these anti-MelaFind findings seem very damning. And as Mandel says, things tend to improve over time. This is sort of “the equivalent of rejecting the first cell phone on the grounds that callers might mishear important messages.”
[UPDATE] Another way of getting at the issue here is to imagine you’re with some friends and proudly announcing “we took our son to the new discount pediatrician over the weekend, so dinner’s on us!”