Clinical Chemistry, Vol 37, 1432-1436, Copyright © 1991 by American Association for Clinical Chemistry

What can be achieved in an external quality-assessment scheme with a small number of participants: four years of experience with thyroid- related tests in Israel

L Kahana and H Yechieli
Central Endocrine Laboratory, Linn Medical Center, Haifa, Israel.

The Israel External Quality-Assessment scheme (EQAS) provided service from 1985 to 1989, primarily to laboratories in Israel. Participation was voluntary and confidential, and involved 30-35 laboratories performing thyroid-related function tests. Scheme design included 36 human serum specimens distributed to the participants at the beginning of each year. For 12 months, three specimens were analyzed each month and the results were reported to EQAS on pre-agreed dates. Monthly reports sent by EQAS to participants after analysis of the results included the overall consensus mean used as target value, between- laboratory agreement, individual laboratory bias, and recovery data. The overall consensus mean calculated as all-laboratory trimmed mean was validated by reproducibility and recovery studies. Samples with undetectable concentrations of thyrotropin (TSH), obtained from patients proven thyrotoxic, were used to validate measurement of very low TSH concentrations. During the scheme, liquid serum was found superior to lyophilized specimens for distribution to the participants. The scheme helped stimulate major improvements in between-laboratory agreement, especially for low TSH concentrations, with CVs decreasing from 180% in the first year to 20% in the fourth year. Each laboratory's performance also improved as judged by the median bias and variability of bias and by the considerable decrease in the number of laboratories with unacceptable performance for all tests. Method- related differences in performance were observed despite the small number of participants. Better methods, e.g., a radioimmunometric method for TSH, were detected and adopted by the participants, with concomitant improvement in performance. Transfer to analog procedures for free thyroxin, in addition to, but mostly instead of, thyroxin, was also documented, with no gain in analytical performance. Despite this being a small scheme, most of the goals achievable with a large scheme were realized.