Reading over this post by David Catron, I was reminded of a recent article in Health Affairs debunking the notion that many Canadians were coming south of the border to seek health care. Titled, “Phantoms In The Snow: Canadians’ Use Of Health Care Services In The United States,” the political left in this country went ga ga over it.

I finally got a chance to read it carefully, and what is remarkable is how limited the data is, and how broad the claims that the authors make given the limits of the data.

Here are the limitations:

1. The data does not go beyond 1998. This seems like a pretty serious limitation, since things may very well have worsened in the nine years since. Of course, the deterioration of Canada’s system might not have been so bad as to precipitate a huge jump in those going south of the border. Nevertheless, any serious increase in the number going south in recent years would be missed since the data ends at 1998.

2. The study limited its data collection to the three most populous provinces in Canada. By my count, that means that nine provinces were excluded.

3. The phone survey the researchers conducted on the U.S. side of the border only surveyed Buffalo, Detroit, and Seattle. Similarly, the researchers also looked at inpatient care for statewide hospitals in the U.S., but limited themselves to data from Michigan, Washington, and New York State. That means they excluded data from the border states of Idaho, Minnesota, North Dakota, Montana, New Hampshire, Vermont, and Maine, not to mention the two states, Ohio and Pennsylvania, that are just a trip across one of the Great Lakes for Canadians.

4. Near the end of the article, there is this rather eye-opening passage:

Periodic formal contracts between provincial payers and U.S. providers have a long history, but a few such contracts have received considerable attention on both sides of the border. Most notable have been contracts for the provision of radiation therapy for cancer patients, in response to backlogs created by shortages of radiation technicians. For example, Quebec contracted with three radiation centers in Vermont and Maine in October 1999 for treatment of patients with breast and prostate cancer; 1,030 patients were treated during the subsequent year. Ontario contracted with three health care organizations in Michigan, New York, and Ohio in March 1999 to provide treatment for patients with breast and prostate cancer, and 1,416 patients had been referred as of 31 October 2000. This is equivalent to approximately 8.5 percent of all prostate and breast cancer patients treated with radiation therapy in Ontario during the same time frame.

8.5%?! That would mean that about 1 in 12 prostate and breast cancer patients from Ontario treated with radiation therapy went south to the U.S. for that treatment!

How can 1 in 12 be described as “Phantoms”?

So not only does the data in this article have some severe limitations, the authors include data that undermines their case.

Perhaps the article should have been titled, “Phantoms In The Snow: Are Their Researchers That Don’t Let Their Biases Color Their Research?”