Sunshine Coast Clinical School, University of Queensland
Among many basic errors of epidemiology, the authors do not even understand how to calculate Number Needed to Treat. (From p. e767: "Given that the risk of UTI among this population is approximately 1%, the number needed to circumcise to prevent UTI is approximately 100.")
Two of the referenced papers from which this figure was derived, written by investigators who do understand how to calculate NNT, found respectively that 111 and 195 circumcisions were needed to prevent one UTI. These data were reported in the abstracts.
Finally, for the authors to say "Most available data were published before 1995 and consistently show an association between the lack of circumcision and increased risk of UTI" and then <i>arbitrarily exclude all data prior to 1995</i> is an unforgiveable design flaw.
This paper is epidemiologically incompetent and an embarrassment to the AAP.