Growth Changes

Abstract

This article presents a method for disentangling (in a study population) the effects of an intervention from changes that are due to growth and regression to the mean. The basic approach is to calculate, for each subject, the expected change over time based on an initial measurement, a set of reference standards, and a set of correlations among measurements over time. The difference between the observed and expected change is attributed to the intervention. Advantages of the methodology are that we take into account both the natural growth process and regression to the mean in calculating the expected change. In addition, the methodology may be applied when no control group is available by letting standards serve as a comparison group, and retrospective measure ments of varying time intervals may be used. The method is illustrated in the context of evaluating the effect of the Special Supplemental Food Programfor Women, Infants, and Children (WIC) on infant growth.