Are routine clinical chemistry analyses 'in control' according to Shewhart's technique?

The clinical laboratory routinely uses quality control (QC) procedures. The Shewhart technique of charting control data, introduced in 1950, is usually applied in a much simplified manner. This reduces its power to distinguish between random errors and errors due to assignable causes. In a comparison of one of our routine charting procedures with a correct Shewhart procedure we checked 23 chemical tests. We found that using the latter 'a stable system of chance causes' did not exist for 6 tests. Both control procedures agreed well on the evaluation of reproducibility. For three of the tests, albumin, urea and uric acid, the instability was not detected by the routine procedure. Although charting of group means (mean i) and ranges (Ri) more readily detected instability, it did not lead to a negation of the practical utility of our patient data.