To evaluate the interlaboratory and intralaboratory agreement in the performance of Lyme disease serological testing, we sent serum specimens from 132 outdoor workers in New Jersey to as many as four independent laboratories. These included one state department of health laboratory, one large commercial laboratory, and two research laboratories. The measurement of agreement employed, the kappa statistic, ranged from .45 to .53 among the four laboratories and from .50 to .54 within the commercial laboratory. These values represent low levels of agreement. The data suggest that Lyme disease serological testing procedures should be standardized so that Lyme disease test results are more comparable between laboratories.
JAMA. 1989 Dec 22-29;262(24):3431-4