ATTENTION: The software behind KU ScholarWorks is being upgraded to a new version. Starting July 15th, users will not be able to log in to the system, add items, nor make any changes until the new version is in place at the end of July. Searching for articles and opening files will continue to work while the system is being updated.
If you have any questions, please contact Marianne Reed at mreed@ku.edu .
Statistical Evaluation of Drug Safety in Clinical Trials
dc.contributor.advisor | Wick, Jo A. | |
dc.contributor.author | Duan, Jiawei | |
dc.date.accessioned | 2020-03-25T18:14:42Z | |
dc.date.available | 2020-03-25T18:14:42Z | |
dc.date.issued | 2019-12-31 | |
dc.date.submitted | 2019 | |
dc.identifier.other | http://dissertations.umi.com/ku:16879 | |
dc.identifier.uri | http://hdl.handle.net/1808/30167 | |
dc.description.abstract | The evaluation of drug safety is critically important in clinical trials. The first part of this dissertation explores new statistical methods for drug safety signal detection in two-arm clinical trials. Current statistical methods for safety signal detection in two-arm clinical trials are typically based on comparing only the incidence rates of adverse events (AEs) using frequentist p values or Bayesian posterior probabilities, regardless of AE severity. To enhance the safety signal detection, chapter 2 of this dissertation describes a frequentist test for evaluating both the AE incidence rate and AE severity in two-arm clinical trials. The frequentist test is based on the Fisher's exact test for AE incidence rate and a proposed conditional test for AE severity that adjusts for potential selection bias. Moreover, in chapter 3 of this dissertation, from the Bayesian perspective, we further proposed a Bayesian three-level hierarchical non-proportional odds version of the cumulative logit model for detecting safety signal with respect to both the incidence rate and severity when all the AEs reported from a two-arm clinical trial are classified into different body system.The three-level hierarchical prior structure takes advantage of the classification of AEs and adjusts for multiplicity because information is borrowed across AEs, especially across the AEs within the same body system. The second part of this dissertation explores statistical applications for safety monitoring in two-arm clinical trials. A few statistical methods for blinded safety monitoring have been proposed. The complex nature of these methods makes the applications challenging. In chapter 4 of this dissertation, we developed two user-friendly R Shiny interactive tools to accelerate, facilitate and improve the process of blinded safety monitoring and reporting in two-arm clinical trials. The interactive tools are based on two blinded safety monitoring methods proposed by Gould & Wang (2017) and Ball (2011) respectively. The dissertation concludes with summary and future studies in chapter 5. | |
dc.format.extent | 111 pages | |
dc.language.iso | en | |
dc.publisher | University of Kansas | |
dc.rights | Copyright held by the author. | |
dc.subject | Biostatistics | |
dc.subject | Adverse events | |
dc.subject | Bayesian hierarchical model | |
dc.subject | Blinded safety monitoring | |
dc.subject | Causal inference | |
dc.subject | R-Shiny interactive tool | |
dc.subject | Safety signal detection | |
dc.title | Statistical Evaluation of Drug Safety in Clinical Trials | |
dc.type | Dissertation | |
dc.contributor.cmtemember | Gajewski, Byron J. | |
dc.contributor.cmtemember | Mahnken, Jonathan D. | |
dc.contributor.cmtemember | Mayo, Matthew S. | |
dc.contributor.cmtemember | Weir, Scott | |
dc.thesis.degreeDiscipline | Biostatistics | |
dc.thesis.degreeLevel | Ph.D. | |
dc.identifier.orcid | ||
dc.rights.accessrights | openAccess |
Files in this item
This item appears in the following Collection(s)
-
Dissertations [4889]
-
KU Med Center Dissertations and Theses [464]