How to lie with statistics number 283,941: circumcision propaganda
I hate this kind of article where some researchers with a chip on their shoulder get press coverage by overinterpreting regressions and sensationalist wording. (Article is entitled "As circumcision declines, health costs will go up, study projects".
Let's examine some of the claims (quoted from article):
- Declining rates of circumcision among infants will translate into billions of dollars of unnecessary medical costs in the U.S. (Hint: quote the total not the per-person cost, it makes the problem sound HUGE!)
- a team of economists and epidemiologists estimated that every circumcision not performed would lead to significant increases in lifetime medical expenses (Hint: define "significant" as "anything at all")
- But in the last decade, studies have increasingly shown that removing the foreskin of the penis has significant health benefits (Hint: in Sub-Saharan Africa)
Let's examine the claims more carefully: First of all it's clear that they are using a significantly different definition of "significant" than what I would use: from the article, the costs are claimed to be: "$313 in added costs for every circumcision that doesn't happen". Yet how many men in the US who are uncircumcised would be willing to pay $313 to keep some hypothetical authority from cutting their foreskin off? My guess is that almost all of them would be willing to pay MUCH more, therefore economically speaking circumcision would be a net loss. Furthermore considering that $313 over a lifetime works out to around 25 cents a month and average US healthcare insurance costs in the hundreds of dollars a month, I fail to see how a fraction of around 0.001 increase in health care costs is "significant".
Next, let's determine how they got this $313 per lifetime? From the article: "nearly 80% of the additional projected costs were because of medical care associated with HIV infection in men" and "Three randomized trials in Africa have demonstrated that circumcision was associated with a reduced risk of contracting HIV..."
So the researchers have taken randomized trials in Africa, a continent where in some countries HIV rates are in the range of 10 to 25% and where large fractions of the population are likely completely ignorant of what a virus is, and extrapolated it via "computer modeling" to the United States where HIV diagnoses per year have been constant to slightly declining since 1999 even while circumcision rates have declined.
Suppose then that we throw out 80% of the costs as coming from a probably invalid model, we're left with $63 which is not enough to cover the cost of a circumcision.