Our lab delivers consistent results with lower false positives and better determinations of true toxicity than other labs nationally. And we have the data to prove it!
EPA DMR-QA Studies Coastal Bioanalysts, Inc. has successfully participated in the Environmental Protection Agency's DMR-QA studies since 1991 (the first year toxicity data was included in the program), submitting 327 toxicity data points to date. Results of the 3 most recent studies are available in the sidebar.
Because acceptance values are based on the 95% probability limits of the mean for all reporting laboratories, including those that are experiencing problems during the mandatory testing period, there is always a 5% chance that the value of an in-control lab will fall out of the acceptance range. Statistically, it is nearly impossible for all of a laboratory�s values to fall within acceptance limits for this many data points. Also, it is well recognized that poor methods (inferior animal health, inadequate water quality, etc.) are more likely to decrease than increase the IC25 or LC50. Thus the absolute true toxicity value is probably somewhat higher than the mean study value . Typically LC50 and IC25 values for our lab are slightly above study values. Of the 14 endpoints that have been out of limits (4.3%), only 2 have been below the study mean.
Reference toxicants are used to evaluate the overall health and sensitivity of the test organisms and the continuing ability of the laboratory to successfully perform the tests. Individual reference test results are compared to control charts to determine acceptability.
The coefficient of variation (CV) for the test LC50s (acute tests) or IC25s (chronic tests) provides a measure of test repeatability or precision; the lower the CV value the less variable the test results and the lower the frequency of false positive and false negative results. For WET testing EPA has compiled frequency distribution tables of intra-laboratory CVs for reference toxicant tests from a national data set. This allows laboratories to compare their test precision with that of other laboratories. Our CV values are consistently well below the national average and most are within the 10th or 25th percentile.
The EPA variability study conducted in 2000 scrutinized laboratory adherence to standard methods and evaluated test performance using a data set of validated results. Our laboratory (number 73) exhibited excellent performance, with test results in close agreement with other in control labs. Additionally, data qualifiers (assigned by EPA) were essentially non-existent for our data compared with other labs.
