The Study of Candidate Genes in Drug Trials
ROBERT ELSTON
Epidemiology and BioStat, Case Western Reserve University
With the increasing number of candidate genes being discovered that may
affect levels of drug metabolizing enzymes, the design of drug trials that
incorporate their study is becoming relevant. One can determine the
necessary sample size for such studies given the number of tests to be
performed, or, alternatively, the number of tests that can be performed
given the available sample size. However, in many cases a uniformly most
powerful test does not exist and/or normal approximations are not accurate
enough to determine sample size. These problems, and solutions to them,
are discussed.