News : 2013 : November

New Publication on Alternative Method to Generate Reference Intervals

In human and veterinary medicine, reference intervals for hematologic and biochemical testing are an essential part in test interpretation, the generation of a diagnosis, and prognostication. Superior reference intervals are generated from a large sample size of clinically normal individuals. This is sample set is often greater than 120 data points. In veterinary medicine, and perhaps especially so in avian medicine, it is particularly difficult to obtain specimens from that number of individuals or even a lesser number of 40 individuals for which a different method of statistical analysis can be used. Time and cost factors are major considerations as well as access to the appropriate number of birds that have a complete health history. In our current study, we have examined the use of indirect sampling method for the generation of reference intervals in avian species.

The indirect method allows for the use of a large database where birds have unknown health status. After the examination of a data set with known clinically normal status, “fences” can be set in the larger database together with a set of statistical calculations to indirectly generate the intervals. These fences will omit the part of the data set with abnormally high or low values. In the data presented in this paper, seven species were examined where data sets varied from 346 to 2358 data points. For white blood cell counts, the indirect method resulted in wider reference intervals than those traditionally expected for these species while for many biochemical determinations, the method resulted in either ranges that were more narrow or approximated the expected ranges. In total, our results indicate that while the indirect method may not be suitable for all determinations, it may provide rough estimates for some analyses until more traditionally generated reference intervals can be produced. This might be especially important for the less common avian species which may be represented well (over time) in clinical laboratory databases.

For more information, see Tang, Messinger, and Cray, Journal of Avian Medicine and Surgery 27(3):194-203, 2013.