On the calculation of a "reference change" for comparing two consecutive measurements

Abstract

We describe a statistical method for calculating a "reference change," defined as that difference between two consecutive test results in an individual that is statistically significant in a given proportion of all similar persons. By allowing for variation in within-person variances, this procedure computes a reference change that is more specific (i.e., less prone to false positives) than that obtained directly from the distribution of observed differences between measurements. Moreover, the method may easily be extended to a test for trend in three successive measurements. The method has been applied to semi-annual measurements of serum calcium and alkaline phosphatase in 698 men and women enrolled in a large health-maintenance program. We believe that these ideas may also be usefully applied to successive laboratory tests in carefully defined patient populations--but this introduces special problems, which are discussed briefly.