A0177
Title: Grouped GEE for heterogeneous longitudinal data
Authors: Tsubasa Ito - Hokkaido University (Japan) [presenting]
Shonosuke Sugasawa - Keio University (Japan)
Abstract: A generalized estimating equation is widely adopted for regression modelling for longitudinal data, taking account of potential correlations within the same subjects. However, since the standard GEE assumes common regression coefficients among all the subjects, such an assumption is not reasonable when there are potential heterogeneities in regression coefficients among subjects. Then, the method called grouped GEE analysis, which is more flexible and interpretable, is proposed to model longitudinal data by allowing heterogeneity in regression coefficients. The proposed method assumes that the subjects are divided into a finite number of groups and that subjects within the same group share the same regression coefficient. A simple algorithm for grouping subjects and estimating the regression coefficients simultaneously is proposed, and the asymptotic properties of the proposed estimator are shown. Finally, the finite sample performances of the proposed methods are demonstrated through simulation studies and real data analysis using health data.
