Analysis of 2-visit COPDgene data with consideration of missing data
Missing data is a common problem in longitudinal studies. Ignoring the issue and implementing a complete-case analysis can lead to estimates which are biased or confidence intervals with an inaccurate coverage rate. Two analytical techniques which consider both the longitudinal and survival aspects of the data are joint and two-part models. Both of these methods associate a longitudinal component of the model with a survival component through shared random effects. This thesis aims to understand how this association impacts the longitudinal estimates for a dataset with limited longitudinal observations.
(For more, see "View full record.")