179 – Advances in Longitudinal Data Analysis
Application of Doubly Repeated Measures Analyses in Drug Trials on Food Intake
Hongmei Han
Pennington Biomedical Research Center
William D. Johnson
Pennington Biomedical Research Center
Doubly repeated measures designs are commonly used in pharmaceutical and human food intake experiments. This design has strengths in reducing error variance while enabling researchers to study participant behavior over time. Mixed models offer a flexible approach for analyzing such designs by permitting general covariance structures with missing data and unequally spaced assessment times. We illustrate the application of the SAS MIXED procedure in a double-blind randomized trial evaluating the effects of a test drug on food intake and appetite, acutely and after 4 weeks of consumption. Eighty two participants were randomized into two groups, one receiving a test drug and the other receiving a placebo. Visual acuity scoring (VAS) was used to evaluate appetite and satiety at baseline, week 0 and week 4 before and after breakfast, lunch, and dinner. The week of clinical visit was one repeated factor and the time within each visit was the other repeated factor. The changes from baseline to each assessment week on VAS scores were analyzed as a response variable with the baseline score used as a covariate. This paper will discuss the model fitting strategies and the unique covariance structure in this trial.
