Answers are grounded in this paper and cite their sources.
A variety of estimators have been proposed for modelling how the proportion of patients with a transient condition varies over time. In this paper we compare these estimators qualitatively and quantitatively and introduce a new estimator based on a generalized additive model (GAM). The assumptions on which the various estimators are based are discussed. Using simulations we compare their performances and investigate their robustness to departures from the assumptions. The GAM estimator is the only one which can incorporate covariate information. Even when estimating a single prevalence function without any covariate information the GAM estimator is seen to be preferable as long as the censoring mechanism is random censoring. The GAM prevalence estimator is applied to data on chronic graft-versus-host disease following bone marrow transplantation.
No takes yet. Share an insight, caveat, or question.
Couper et al. (1997) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: