Dissertation
Principal stratification and Bayesian testing in clinical trials with intercurrent events
Doctor of Philosophy (Ph.D.), Drexel University
Dec 2023
DOI:
https://doi.org/10.17918/00001945
Abstract
Patients in clinical trials may discontinue their randomized study treatments due to intercurrent events such as adverse side-effects. Accordingly, clinical trials for drug approval must evaluate treatment effects while accounting for intermediate outcomes, to prevent biased inferences that could arise from confounding of latent variables. Existing statistical methodologies for clinical trials such as intention- to - treat typically ignore treatment discontinuation and focus on treatment assignment. Principal Stratification as identified by the ICH E9 (R1) addendum is a strategy to deal with intercurrent events in clinical trials. In this dissertation we (1) introduce a Bayesian testing methodology that can account for the existence of principal strata in clinical trials with intercurrent events, (2) present a Bayesian multiple testing strategy that can account for the existence of principal stratum and multiplicity, and (3) extend methodology for sensitivity analysis to assess assumptions of the use principal stratum methods. The potential utility of these methodologies will be illustrated on simulated clinical trials from a novel data generating model that accounts for multiple intercurrent events under the potential outcomes framework.
Metrics
50 File views/ downloads
30 Record Views
Details
- Title
- Principal stratification and Bayesian testing in clinical trials with intercurrent events
- Creators
- Dominique Antionette McDaniel
- Contributors
- Leslie McClure (Advisor)
- Awarding Institution
- Drexel University
- Degree Awarded
- Doctor of Philosophy (Ph.D.)
- Publisher
- Drexel University; Philadelphia, Pennsylvania
- Number of pages
- vii, 110 pages
- Resource Type
- Dissertation
- Language
- English
- Academic Unit
- Dana and David Dornsife School of Public Health; Epidemiology and Biostatistics; Drexel University
- Other Identifier
- 991021819114604721