Analyzing Longitudinal Data from Glaucoma Trials

In glaucoma trials, patients often come in for multiple visits, and at each visit, intraocular pressures are taken at multiple timepoints throughout the day. A mixed model for repeated measures is often used to model the data from such trials. One modeling approach is to jointly model all visits and timepoints together. Another modeling approach is to model data across all visits at each timepoint separately. I conducted a simulation study to compare the two modeling approaches. The results suggest that modeling each timepoint separately across all visits is superior, especially when missing data are low. This approach is recommended over jointly modeling all visits and timepoints together.