Date of Award

2016

Embargo Period

8-1-2024

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Public Health Sciences

College

College of Graduate Studies

First Advisor

Valerie Durkalski

Second Advisor

Wenle Zhao

Third Advisor

Sharon Yeatts

Fourth Advisor

Jody Ciolino

Fifth Advisor

Keith Borg

Sixth Advisor

Patrick Mauldin

Abstract

There has been little work to date regarding the proper use of covariate information in non-inferiority trials. Too often knowledge obtained in the superiority setting is applied directly to the non-inferiority setting. However, due to the reversal of the hypotheses and the consequent reversal of the implication of error probabilities, this is a dangerous practice. The current work demonstrates that in both superiority and non-inferiority, failure to adjust for important covariates results in estimates of treatment effect that are biased towards zero with standard errors that are deflated. However, as no treatment difference is approached under the null hypothesis in superiority and under the alternative in non-inferiority, this results in decreased power and nominal or conservative (deflated) type I error in the context of superiority, but inflated power and type I error under non-inferiority. This occurs regardless of adjustment at randomization. Generally, it is advised that covariates requiring adjustment be specified before the start of the trial. However, important prognostic factors are not always known in advance. Thus, a joint statistic for the identification of important covariates based on the simultaneous assessment of influence on outcome and disparity across treatment groups is developed for the non-inferiority setting. This statistic, when calculated for all available covariates in a trial, can be used to rank them according to importance. This ranking can be used to identify the subset that will optimize the tradeoff between the change in the point estimate of the treatment effect and its precision while preserving type I error. This method is applied to the Rapid Anticonvulsant Medication Prior to Arrival Trial (RAMPART) and its pediatric cohort.

Rights

All rights reserved. Copyright is held by the author.

Download

Share

COinS