Pages

October 30, 2007

Confound it!

Rudy Giuliani makes a mistake on his stats for surviving prostate cancer in the UK. Quotes 44%. In fact, the US is doing better, but not by that much; 82% vs. 74.4%.

Why don't politicians get it right the first time? Their staff has to be extremely specialized and well-informed. In this case, it might not be the people but rather the wide array of answers they get for one simple question.

In health care, its extremely difficult to control for confounding variables. Essentially, to get rid of underlying variables that may influence the answer to the question you're asking. The best example is the placebo effect. You want to figure out how to get rid of a cold. You take some medication and your cold goes away. Its very possible though that because you thought the medication would help you, you psyched yourself out of a cold. Its been proven time and again just by giving patients sugar pills and finding out a couple of days later that they feel great!

Anyway, comparing prostate cancer survival rates of two countries can be extremely confounding. Different people, different doctors, different education, different testing, different technology, the list goes on. I bet the survival rates in the south and north of US alone are different! Depending on what you confound, you're likely to get a different answer to the same question.

Bottomline:
Countries are hard to compare. Maybe we should start smaller.