A slew of recent articles, published mainly in medical journals, suggest that the health-care systems of other countries are superior to ours on all these fronts. Yet the articles are at odds with a substantial economic literature.
Mr. Goodman then goes on to cite various studies and statistics that show that we are often comparing apples to oranges. I don't doubt this at all, as there are countless examples where our diverse population is compared to non-diverse populations (in crime studies, family studies, etc.):
All too often, the heterogeneous population of the United States is compared with the homogeneous populations of European countries. A state such as Utah compares favorably with almost any developed country. Texas, with its high minority population, tends to compare unfavorably.
A call for consistency in data universes is a good thing, but so are rigorous standards in the use of statistics, standards Mr. Goodman did not employ when writing the piece. The following paragraph is an example of what I mean by not complying with standards in using statistics:
Overall, nearly 1.8 million Britons are waiting for hospital or outpatient treatments at any given time. In 2002–2004, dialysis patients waited an average of 16 days for permanent blood-vessel access in the U.S., 20 days in Europe, and 62 days in Canada. In 2000, Norwegian patients waited an average of 133 days for hip replacement, 63 days for cataract surgery, 160 days for a knee replacement, and 46 days for bypass surgery after being approved for treatment. Short waits for cataract surgery produce better outcomes, prompt coronary-artery bypass reduces mortality, and rapid hip replacement reduces disability and death. Studies show that only 5 percent of Americans wait more than four months for surgery, compared with 23 percent of Australians, 26 percent of New Zealanders, 27 percent of Canadians, and 36 percent of Britons.
Something didn't seem Kosher in the above paragraph and I also felt that presenting statistics in written form (rather than charts and graphs) made it difficult to follow or see the trends and comparisons. I attempted to graph the data presented, but couldn't, and that's when I became suspicious that Mr. Goldman was cherry-picking the data.
He begins with: "...nearly 1.8 million Britons are waiting for hospital or outpatient treatments at any given time." There is no comparison to the U.S. (or anywhere else). Without direct comparison to the same summary of data to provide another row in the dataset, it is an interesting factoid, but a meaningless stat in relation to a conclusion of any kind.
He continues with: "...dialysis patients waited an average of 16 days for permanent blood-vessel access in the U.S., 20 days in Europe, and 62 days in Canada." That sentence appears to qualify the first, but it picks a particular procedure, rather than the summary universe of the first sentence.
In the sentence noted directly above, the subset universe was "Europe," but Mr. Goodman changes the universe to a specific country in Europe in the next sentence, "Norwegian patients waited..." There is no comparison to the U.S., nor any other European country. The sentence that follows, "Short waits for cataract surgery..." has nothing to do with the statistics under discussion. It's advice, not data, nor does it show that we do it any better or worse (or that any country does it better or worse).
In the last sentence, "Studies show that only 5 percent of Americans wait..." Mr. Goodman changes the dataset from hospital or outpatient treatments (first sentence) to surgery. Are they the same? I don't know. It could be shoddy use of aggregate row reference, but I can't assume that "hospital or outpatient treatments" is the same as "surgery" in the latter example. He then goes on to change the data subsets from a Europe summary, to selected countries that haven't been introduced in earlier statistical comparisons, and limits to only one country in Europe, Britain.
That is shoddy use of statistics.
I don't doubt the validity of the facts/stats that Mr. Goodman has included, but it makes me suspicious of his conclusions because he appears to have selected only the data subsets that confirm his hypothesis, provide remarkable differences, or sensationalism.
He should do better (especially as "president, and CEO of the National Center for Policy Analysis"). If you don't have comparison statistics when trying to make a comparison point, don't use the stats. Don't be guilty of employing shoddy standards when the subject of your essay is shoddy standards!
Posts
