Proficiency test performance as a predictor of accuracy of routine patient testing for theophylline.

Proficiency testing (PT) is pivotal in assessing laboratory qualifications for certification and licensure. PT is expected to typify routine assay performance and determine whether the laboratory is producing clinically useful test results. Conventional schemes use mail-distributed test specimens and are often criticized as measuring the best possible laboratory performance, principally because of special practices associated with processing PT specimens. We used on-site proficiency tests and split samples to evaluate the ability of conventional PT schemes to accurately characterize routine laboratory performance. Using 412 assays of theophylline, performed routinely by 200 laboratories and subsequently in a reference laboratory, we found that the predictive value of PT performance in assessing quality of routine testing was high (100% for predicting substandard reliability of routine patient testing and 94% for excluding substandard reliability of patient testing). The imprecision of interlaboratory PT results was equivalent whether testing was observed (hand-carried specimens) or unobserved (mail-distributed specimens). Many methods used for determining theophylline concentration in serum were highly automated, closed, and precise analytical systems. The performance characteristics of these analytical systems are not easily manipulated by the analyst for purposes of improving PT outcome, and PT by use of mail-distributed test specimens is effective for assessing intralaboratory performance.