Comparison of methods to determine analytical sensitivity

Abstract

The three methods to determine analytical sensitivity were compared on their advantages and disadvantages. The statistical test for the difference of the two population means with t-distribution is not appropriate as a evaluation method to get analytical sensitivity of a measurement method. Because the sensitivity value determined by this test is too small to examine routine data with this value. It is desirable that the maximum of the results of real specimens including no object substance should be considered as analytical sensitivity, if such specimens are available. I suppose followings on account of both difficulties to get such specimens and importance that anybody can check the sensitivity value. 1. Analyze obvious zero-concentration sample such as diluent in replication of fifty times. 2. Measure parameter such as light absorbance or luminescent intensity, and calculate both mean (M) and standard deviation (SD) of fifty results. 3. Convert the parameter value M + 3SD into substance concentration, and take this concentration as analytical sensitivity. 4. Consider three to five times analytical sensitivity value as practical sensitivity.