“The science of doping” revisited: Fallacies of the current anti-doping regime

Abstract

Biomedical doping tests specifically, as well as doping test practises as a whole, have severe shortcomings which are based upon the fact that test procedures are in principle imperfect. With the help of Bayesian logic, it can be shown that the interpretation of doping test results is far from trivial with reference to individual athletes and single doping tests, but that these tests can reveal results which can be fairly dependable when it comes to estimating the prevalence of doping. Even if multiple tests are conducted and the samples are analyzed repeatedly for substance abuse, the need for positive A- and B-sample analysis is a strong argument in assuming that positive-tested athletes are guilty. However, this does not save an anti-doping policy which relies on increasing numbers of doping tests all over the world from the problem of false positive tests, which becomes the more severe the better anti-doping works in terms of lowering doping prevalence through deterrence. The considerations based on test theory and Bayesian logic lead to important ethical questioning of anti-doping policies.