Serum cholesterol analysis in the C.B. Heart Project. Intra-and inter-laboratory continuous quality control

Standardization and quality control of measurements are of crucial importance in large medical surveys. In the C.B. Heart Project serum cholesterol analyses in the central laboratory were extensively controlled with regard to precision as well as accuracy. All measurements are carried out in side-by-side duplicate, the maximum difference allowed being 0.3 mmol/l. Bench control, aiming at maintaining linearity and precision with control limits, requires insertion of standard solutions and several serum controls into every run. Control values must fall within 2 S.D. limits from the established mean values, e.g. 4.1 +/1 0.25 mmol/l and 8.2 +/- 0.25 mmol/l. A very reliable check on the precision was obtained by blind introduction of duplicate patient specimens. Repeatability (within-run precision) as calculated each month has S.D. values of approx. 0.10--0.20 mmol/l, whereas reproducibility (long-term precision) S.D. values are 0.15--0.25 mmol/l. Accuracy is checked in the Cooperative Standardization Programme of the WHO Regional Lipid Reference Centre in Prague. This makes the C.B. Heart Progject results comparable with other standardized studies. Our bias approximates +3% and usually remains within WHO "narrow limits".