A Few Do Not Represent the Whole

I am spectacularly impressed with Michael Moore's ability to create such a compelling documentary without presenting any aggregate data whatsoever.

What Moore has sucked his audience into is perhaps the most damaging fallacy in any population-level field of study: he is using a few people (or a few clinics) to represent the situation of entire nations. We are all susceptble to this sort of bias: you have a 2-hour conversation with a cancer patient, and you become far more sympathetic to his plight. You have some connections with senior government officials or some private organizations, and you fight for funding on behalf of cancer treatment programs or advocacy groups. Now, I'm not saying that special interest or minority groups shouldn't have any advocates. But what if you had just happened to talk to an MS patient instead? Or someone with end-stage renal disease? Perhaps that funding you fought so hard for would have ended up elsewhere.

Surely having so many special interest groups all over the place ultimately becomes damaging and framentative instead of beneficial.

But this is exactly what Michael Moore does in his movie. He tells a few, heart-wrenching stories about real US citizens to get us to become sympathetic to their plight. He then uses a few, uplifting examples of people (or services) from other nations to try and convince us that their system is superior to ours.

But this is absolutely the wrong discussion to have. As I've mentioned before in a previous post, there is no nation with a "perfect" healthcare system; there are only nations with imperfect healthcare systems that have accepted what each believe to be the lesser of several evils. Michael Moore showed us a concrete example of a single French family to convince us that the French are better off than we are. He interviewed a single UK physician to convince us that doctors over there are perfectly happy with the NHS. He then went to a single clinic in Canada to dispell the "myth" that Canada has long wait times. He backs his claim with evidence from real patients in a real healthcare setting, so we are inclined to lend credit to what he says.

But zoom out from that single clinic, and this is what the aggregate data shows us:



Clearly, if you look at the nation as a whole, Canada's average wait times are much longer than the US's (and a number of other developed countries). Does this mean that Canada's healthcare system is bad? No. But it does mean that it is not as perfect as Michael Moore leads us to believe.

While humanistic stories from real people are extremely important for capturing or conveying the significance of healthcare problems, I firmly believe that these stories cannot form the basis of policy decisions at the population-level. You can't base healthcare reform that will affect hundreds of millions of people on the stories (however touching) of a few.

The only reliable source of information for national reform is national data. Sadly, Sicko seems to miss out on this point.