Search

In three performance evaluation surveys, panels that consisted of human T-lymphotropic virus type I or type II (HTLV-I/II) antibody-positive and -negative plasma samples were mailed to laboratories that voluntarily participated in the Centers for Disease Control Model Performance Evaluation Program. Donor samples were identical among surveys. In each survey, more than 98% of the laboratories reported enzyme immunoassay (EIA) test results; about 11% also reported results of Western blot (WB) testing. Variation in analytic sensitivity (96.7% to 99.4%) and specificity (98.3% to 99.5%) of EIA tests was noted in the three surveys. For WB testing, no nonreactive interpretations were reported for HTLV-I/II antibody-positive samples in any survey; however, indeterminate interpretations were reported for 35.2% to 40.7% of the WB tests that were performed on HTLV-I/II antibody-positive samples. More than 95% of these indeterminate WB test interpretations were reported for HTLV-II antibody-positive samples. Although HTLV-I/II antibody tests are generally sensitive and specific, their accuracy could be further improved by increasing the specificity of EIA tests and the sensitivity of WB tests.

The performance of laboratories enrolled in three of the Centers for Disease Control microbiology performance evaluation programs was studied in relation to the number of actual patient specimens tested by the laboratories. Laboratories were grouped according to the number of specimens tested weekly to compare performance among the groups. Results of the study showed that laboratories, as a group, that tested large numbers of specimens performed better than laboratories that tested small numbers. The infrequency of performing certain tests or identifying certain microbial species can be an important factor in laboratory error. One strategy for laboratories that cannot resolve internal problems associated with low testing volumes would be to limit their testing to procedures that are done well.