roster wrote:The discussion then took quite a few turns, but to this moment no one has pointed out a single inaccuracy in the original article.
I was interested in this article, which I hadn't seen before, when it was first posted, and I started to go through the list of facts with the intention of reading the sources cited. That can be especially rewarding with material you feel skeptical about, since there's always a chance to learn something new (or unlearn something you thought was true). I was disappointed, though.
I started where Rooster did with "Americans have lower cancer mortality rates than Canadians." The original footnote in the Scott Atlas article he is quoting is this:
U.S. Cancer Statistics, National Program of Cancer Registries, U.S. Centers for Disease Control; Canadian Cancer Society/National Cancer Institute of Canada; also see June O'Neill and Dave M. O'Neill, "Health Status, Health Care and Inequality: Canada vs. the U.S.," National Bureau of Economic Research, Working Paper No. 13429, September 2007. Available at
http://www.nber.org/papers/w13429.
That's pretty frustrating to work with because of the lack of specific information needed to actually check the sources. The first citation is to "U.S. Cancer Statistics, National Program of Cancer Registries, U.S. Centers for Disease Control." No year, no specific publication. I rooted around the web site for the National Program of Cancer Registries,
http://www.cdc.gov/cancer/npcr/, but I couldn't out what data the author was using. It was even more difficult with Canada, since the citation was simply to the names of two organizations. So I pinned my hope on the third source, the O'Neill paper, which can be downloaded free from
http://www.bepress.com/fhep/10/1/3/ if you register as a guest. (It costs $5 from the source the author cites.) Anyway, Table 11 on page 30 has mortality rates and cites to "United States Cancer Statistics, National Program of Cancer Registries, U.S. Centers for Disease Control; Canadian Cancer Society/National Cancer Institute of Canada." Bingo! That looks like it must be the source. But for the life of me I can't see where the author gets his numbers. They don't match up. For example, the article says that the mortality rate for prostate cancer is 184 percent higher in Canada than in the United States. But here is what the O'Neills' table says for prostate cancer:
U.S.
Incidence rate, 162.0
Mortality rate, 27.9
Mortality/Incidence, 0.17
Canada
Incidence rate, 124.7
Mortality rate, 25.2
Mortality/Incidence, 0.20
There is simply no way I can think of to derive from these figures that the mortality rate for prostate cancer is 184 percent higher in Canada. I can't match of the statistics for breast cancer or colon cancer in men either. The O'Neills have a second table, Table 12, on page 31. Those numbers don't support the author's statistics. Moreover, they are derived from only parts of the U.S. and Canada. The U.S. statistics are from what is called the SEER 17 areas: Alaska Natives, Arizona Indians, California, Connecticut, Detroit, Atlanta, Rural Georgia, Hawaii, Iowa, Kentucky, Louisiana, New Jersey, New Mexico, Seattle area, and Utah. That's a weird selection. According to the National Cancer Institute (
http://seer.cancer.gov/about/) this covers 26 percent of the population. The Canadian statistics exclude Quebec. I can't imagine what comparing random parts of the U.S. to Canada with the middle cut out is supposed to prove.
So, to make a long story short, I tried to follow up the source of the initial comparison of U.S. and Canadian cancer mortality statistics, but I couldn't find the actual numbers in the first two vaguely cited sources, and the third source, which claimed to be based on the first two, doesn't support the author's statement. I would love to see someone else take a crack at it. I can say for sure that Table 11 in the O'Neill paper doesn't come close to supporting the high mortality difference claimed for prostate cancer, which was the only startling difference. I'm not very impressed by small differences, since we don't know anything about variances between countries in how data is collected or defined.
In some ways I am more bothered by what the author did with his Fact No. 7. I think someone else has already pointed this out. He says:
People in countries with more government control of health care are highly dissatisfied and believe reform is needed. More than 70 percent of German, Canadian, Australian, New Zealand and British adults say their health system needs either "fundamental change" or "complete rebuilding."
He cites to Cathy Schoen et al., "Toward Higher-Performance Health Systems: Adults' Health Care Experiences In Seven Countries, 2007," Health Affairs, Web Exclusive, Vol. 26, No. 6, October 31, 2007, pages w717-w734. Available at
http://content.healthaffairs.org/cgi/reprint/26/6/w717.
Here is the chart from Schoen's paper. I downloaded it from the link called Chartpack at
http://www.commonwealthfund.org/Content ... -2007.aspx because that was free, but it is from the same author, title, and year, and is clearly what is referred to in the cited paper:
It clearly shows that of all the countries surveyed, the U.S. had the largest number who said that the system should be rebuilt completely and the smallest number who said that only minor changes were needed. The author didn't actually lie. He omitted any mention of the U.S. results. But the whole idea of the article was comparison of the U.S. to other countries, so it was intellectually dishonest to cite the dissatisfaction of the residents of the other countries, while failing to state that the U.S. was also surveyed and that we hate our system even more than other people hate theirs.
Heck, I'm a lawyer. I am paid to take a side and then muster the evidence that supports it. I expect more from academics, though. The author of this article leaves the strong impression that he is an advocate, not a scholar. He seems to have cherry-picked evidence that supports the side he has decided to take and ignored conflicting evidence. This makes me less inclined than I was previously to take seriously anything by Scott Atlas or anything published by the Hoover Institution.
In answer to Rooster's specific challenge to find a single inaccuracy, in one case I can't verify the author's sources and in the other he has left out something that makes what he says seriously misleading.