Linear Models with Nested Error Structure in Predicting Vision Loss for Patients with Subretinal Neovascular Membranes

Abstract

Age-related macular degeneration (AMD)* and presumed ocular histoplasmosis (POHS) are common causes of macular degeneration. Both are major causes of blindness, with AMD being the leading cause of blindness in people over the age of 65. The major cause of visual loss in both categories is the presence of a subretinal neovascular membrane (NVM) in the macular. Sometimes these conditions can be treated successfully with laser therapy. Our task was to develop a regression model for predicting post-treatment vision as a function of time from treatment and baseline prognostic factors measured at diagnosis. The particular analysis of the model was to examine how patients' post-treatment vision is affected by baseline factors. A nested-error structure was used in a linear model.