In 1991, we reported that 55% of laboratories participating in the Wisconsin Proficiency Testing Program could not accurately identify serum samples from Lyme disease patients containing antibody against Borrelia burgdorferi. The purpose of this study was to determine whether the accuracy of Lyme disease test results reported by approximately 500 participants in the Wisconsin State Laboratory of Hygiene/College of American Pathologists Lyme Disease Survey had improved. From 1992 through 1994, 50 serum samples were sent to participants of the survey. Each laboratory received 28 serum samples from individuals with Lyme disease according to the case definition of the Centers for Disease Control and Prevention and 22 serum samples from healthy individuals. Unfortunately, the serodiagnosis of Lyme disease by participants had not improved. The specificity of the Lyme disease assays steadily decreased from approximately 95% to approximately 81% during the 3-year period of the survey. False-positive test results approached 55% with some of the serum samples from healthy donors. A serum sample containing antibody against Treponema pallidum was reported as positive by 70% of the participants. In addition, the sensitivity fluctuated between 93 and 75%, depending upon the conjugate used by the laboratories. These results suggest that stronger criteria must be applied for approving and continuing to approve commercially available kits for the serodiagnosis of Lyme disease.