How can variability be controlled

A conventional response to issues of variability in bioassays is to construct Shewhart Control Charts based on the results achieved in repeat tests within a laboratory using a reference toxicant. This effectively describes the range of results typically found within the laboratory and hence can be used to define limits within which the laboratory normally expects to operate. However, there is a flaw in such internal quality control because the more variable a laboratory s reference toxicant test results are, the wider the limits of acceptability will be. Indeed, it can serve merely to reinforce high variability or bias. 

Such an approach was trialled in the late 1990s as part of an initiative by the Environment Agency in the UK. It led to proposals for a quality control scheme for use within the Agency s Direct Toxicity Assessment programme for effluent assessment and control. The trial entailed an interlaboratory ring-test of four test methods using two toxicants (3,4-dichloroaniline and zinc sulphate) from which accuracy and precision criteria were estimated, as described below. 

---