Analysis of partial variance (APV) as a statistical approach to control day to day variation in immune assays

The considerable day to day variability in some immunologic assays can obscure relationships among measures of interest and thereby complicate interpretation of findings when effect sizes are modest. Analysis of partial variance (APV) is a technique that utilizes data obtained from controls run in the laboratory on the day of each assay to control for day to day variance statistically. The application of this technique to permit detection of otherwise obscured effects for both larger and smaller samples is exemplified utilizing data from two studies in which subjects were assessed on multiple different dates. These included a study of the association of depression with in vitro immune measures in 296 adolescents and a longitudinal study of 20 spouses of patients admitted to a cardiac intensive care unit studied together with 20 matched controls. APV was most useful in analyses of mitogen-induced lymphocyte proliferation assays, which had considerable day to day variability. In general, multivariate studies with functional immune measures may benefit most from application of this technique.