2019/12/01 by Giorgos Bakoyannis, Bakoyannis, Giorgos
Mathematics · Social Sciences · #Advanced Causal Inference Techniques #FOS: Computer and information sciences #Insurance, Mortality, Demography, Risk Management #Methodology (stat.ME) #Statistical Methods and Bayesian Inference #Statistical Methods and Inference
paper · pdf · doi:10.48550/arxiv.1912.00487
openalex publication_date 2019/12/01 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Frequently, clinical trials and observational studies involve complex event\nhistory data with multiple events. When the observations are independent, the\nanalysis of such studies can be based on standard methods for multi-state\nmodels. However, the independence assumption is often violated, such as in\nmulticenter studies, which makes the use of standard methods improper. In this\nwork we address the issue of nonparametric estimation and two-sample testing\nfor the population-averaged transition and state occupation probabilities under\ngeneral multi-state models based on right-censored, left-truncated, and\nclustered observations. The proposed methods do not impose assumptions\nregarding the within-cluster dependence, allow for informative cluster size,\nand are applicable to both Markov and non-Markov processes. Using empirical\nprocess theory, the estimators are shown to be uniformly consistent and to\nconverge weakly to tight Gaussian processes. Closed-form variance estimators\nare derived, rigorous methodology for the calculation of simultaneous\nconfidence bands is proposed, and the asymptotic properties of the\nnonparametric tests are established. Furthermore, we provide theoretical\narguments for the validity of the nonparametric cluster bootstrap, which can be\nreadily implemented in practice regardless of how complex the underlying\nmulti-state model is. Simulation studies show that the performance of the\nproposed methods is good, and that methods that ignore the within-cluster\ndependence can lead to invalid inferences. Finally, the methods are applied to\ndata from a multicenter randomized controlled trial.\n