Performance goals for internal quality control of multichannel haematology analysers.

Performance goals for analytic bias and analytic imprecision are presented for multichannel haematology analysers. These goals represent a compromise between limits that would cause minimal degradation in the information content for clinical decisions and limits that are achievable by current instruments. For midrange measurements the maximal limits for total analytic bias are +/- 14.6% for WBC, +/- 5.3% for RBC, +/- 4.5% for Hb concentration, +/- 4.2% for MCV, and +/- 22% for platelet counts. The limits include error inherent in calibrator value assignment, instrument calibration, instrument drift, and instrument non-linearity as well as the quality control systems necessary to monitor each of these processes. If analytic biases are held within the proposed limits, the clinical specificities of the assays will be decreased by less than 7% for all assays except that platelet counts will have a decrease in specificity of about 15%. Analytic imprecision has minimal direct effect on clinical decisions because the influence of within-person and across-person biological variation dominates. However, analytic imprecision is a major factor in the control of analytic bias and thereby remains an important quality control factor.