Efficiency of General Estimating Equations Estimators of Slopes in Repeated Measurements: Adding Subjects or Adding Measurements? *

Abstract

In controlled clinical trials, subjects are often evaluated at baseline and intervals across a treatment period. In most trials, the treatment period and the number of time points are predetermined by design. When the primary goal is to estimate and compare the rate of change in outcome variables over time, investigators are often confronted with difficult decisions of maintaining a balance between increasing the number of study subjects and increasing the number of measurements for each subject. In this paper, we present a method to evaluate the relative benefit of adding subjects versus adding measurements in terms of the efficiency of the general estimating equation (GEE) estimator of slope coefficients in repeated measurements.