As part of a program to develop accreditation guidelines for urine drug testing laboratories, a pilot study for proficiency testing was conducted. Fifty civilian, commercial laboratories were included on a voluntary basis. Drug-free urine specimens were collected and either fortified with commonly abused drugs at concentrations comparable to casual use or submitted unfortified to participating laboratories as blanks. Samples were submitted on both an open and blind basis to the laboratories. Laboratory performance on open proficiency testing was comparable with that reported in existing proficiency testing programs. Blind proficiency testing produced less accurate results in terms of apparent false-negatives, but significant difficulties were evident in carrying out blind testing and in comparing its results with those of open testing. Specific problems have been identified to guide future programs.