Six Sigma

Six sigma provides a convenient way to monitor the performance capability of a testing system. During the 1980s, Motorola set out to improve their manufacturing process so virtually no defective product would be produced. Motorola defined this as having six sigmas (standard deviations) of process variation fit within the product tolerances.

Assuming a normal (Gaussian) distribution, the following table shows the effect of product specifications (expressed as standard deviations) on the defect rate and defects per million.

Note: The advantage of controlling a process to six sigma is that the process can tolerate small shifts without significantly increasing the defect rate. In an ideal world, all processes would be six sigma and could be monitored with very simple QC.

Unfortunately, not all processes are six sigma and as process capability decreases, the choice of QC procedures becomes increasingly important in detecting significant errors. In fact, some processes may have such low process capacity (that is, a high total error), that they cannot be controlled to a defined level of quality. This condition would trigger a maximum QC condition in the optional Westgard Advisor™ online.

For clinical laboratory tests, the total error (TE) for a test provides an indication of the process capability of the test because the TE combines bias and imprecision. The following formula is used to calculate sigma:

Note: The optional Westgard Advisor™ online calculates and displays sigma based on the data for a test, the selected TEa, and the consensus group.

It is possible to correlate sigma with the TEa as shown in the following table. (The table assumes bias is zero.)

Note: When the bias is not zero, the sigma classification of a process decreases as its bias increases.

The information in this chapter was abstracted from the Six Sigma Quality Management and Desirable Laboratory Precision lesson available on http://www.westgard.com. Dr. Westgard provides an interesting correlation between CLIA performance requirements and the performance requirements for a five or six sigma process. As Dr. Westgard concludes, "Six sigma quality management sets demanding standards of performance for laboratory testing processes."