Identical samples sent to different laboratories for testing can sometimes show very different results. This is very disconcerting for all concerned, particularly if the result being contended has regulatory implications. Any such issue, once raised, must be resolved as diplomatically as possible. For that to happen, all concerned must listen to all others. This can sometimes be a problem.
For example, at one time samples came back from laboratory testing showing small but consistent contamination with polychlorinated biphenyls (PCBs). These compounds were once widely used because of their unique electrical properties but were later found to be quite dangerous and so were banned. Because of the chemical stability of these banned compounds, the methods of their final disposal are expensive. As the result of the testing, the company involved was presumed to have a 50,000 gallon tank contaminated with PCBs. The costs of hazardous waste disposal of this quantity of PCB contaminated material would have been ruinous for the company possessing it. Yet, the provenance of this material argued against PCB contamination. The method used by the outside laboratory was a well-known gas-liquid chromatographic method. The literature on PCB analysis clearly showed the possibility of what was termed a "memory effect." This meant that some of the PCBs found in a previous run could find their way, unbidden, into a subsequent run demonstrating PCB contamination that was not really there. This possibility was raised with the outside laboratory, who denied it was the case and stuck to their guns.
Their assertion was put to the test. A sample of reagent-grade water was purchased along with sample containers that were certified clean. One of these containers, filled with reagent-grade water, were labeled as another sample from the same tank and sent to the same laboratory with a request for a PCB analysis. The result that came back alleged PCB contamination of a regent-grade water sample, whereupon the certificates of analysis from both the reagent-grade water and the clean container purporting to have been the "sample" were shared with the laboratory. Only when this ruse was explained to the outside laboratory was the problem finally recognized and resolved. Perhaps the corollary to this tale is that other methods for PCB analysis exist and should have been tried when the situation warranted it.
Another example was microbiological. Samples of the same lot of product were sent to three different laboratories, two of whom declared it free of bacterial contamination while a third claimed a small but significant level of bacterial contamination. Investigation of this discrepancy showed that all three laboratories were using slightly different methodologies. More importantly, it was discovered that the three laboratories all expressed their results using very different levels of analytical sensitivity (10 CFU/ml, 1 CFU/ml and 1 CFU/100 ml). As can be readily seen, the analytical sensitivities used varied by a factor of a thousand overall! As the two labs with the lower analytical sensitivity claimed that the product was clean while the high analytical sensitivity lab showed a positive result, it could be seen that this was not a discrepancy at all, rather just a confused interpretation.
The main point is that different results obtained by different laboratories usually vary for a reason. Discovery of that reason may take some time and technical expertise but that this process should proceed with the utmost diplomacy since, most of the time, all of the disputants are and should be on the same side.
Copyright © 2012 by M. Mychajlonka, Ph. D.