Fallibility of urine drug screens in monitoring methadone programs.

Urine specimens containing five different drugs, each at three levels of concentration with zero to five drugs in a specimen, were sent to two "approved" laboratories. In only 46.9% and 13.8%, respectively, were all drugs present correctly identified and no false-positive results reported. With some allowances, the results improved to 53.8% and 49.4%. If these tests are to be continued then (1) the fallibility of these tests should be known by all treatment personnel, (2) laboratories should be licensed rather than merely approved, and (3) maintenance of the license should be made contingent on passing "blind" proficiency tests.