Back to search

Boston University

Sample size re-estimation in active controlled non-inferiority clinical trials using a frequentist approach

Abstract

dc:description.abstract

In active controlled clinical trials a possible objective is to test a non-inferiority hypothesis that the experimental treatment is therapeutically not inferior to the active control within a pre-defined margin. At the design stage, the misspecification of any design parameters (e.g., the variance or treatment difference for continuous endpoints, control event rate or non-inferiority margin for binary endpoints) can lead to study power below the desired level. Sample size re-estimation (SSR) procedures protect study power by allowing sample size re-estimation based on an interim analysis using revised estimates of the design parameters. For continuous endpoints, current approaches to SSR for non-inferiority trials focus on updating the sample size based solely on the estimated variance (blinded or unblinded) at the interim. The SSR using both sample variance and the observed treatment difference at interim in conditional power calculations is used in superiority trials. We have extended the methodology to non-inferiority trials, quantified the effect on the type I error rate, and proposed controlling it by modifying the critical value and/or stopping the trial at the interim for futility. For binary endpoints, current approaches to SSR for non-inferiority trials focus on estimating the event rates (blinded or unblinded) at the interim and update the sample size solely on the estimated event rates at the interim without updating the non-inferiority margin. A procedure that adapts both the absolute non-inferiority margin, and sample size based on the underlying interim observed pooled (blinded) event rate, and updates non-inferiority margin again at the final analysis based on the observed estimate of the event rate in control group at the end of the study is proposed. Our simulation results show the proposed adaptive procedures for extending a study by adding sample size, if necessary, preserve the overall type I error rate and maintain desired power. Combining sample size re-estimation methods with early stopping rules for continuous endpoints and adapting non-inferiority margins for binary endpoints could increase study flexibility, scope, and efficiency of non-inferiority trials. The proposed methodologies can be used for designing efficient two-stage non-inferiority trials with sample size re-estimation in active controlled non-inferiority clinical trials.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Biostatistics
Grantor
Boston University
Year dc:date.issued
2013

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Guo, Wei
Advisor dc:contributor.advisor
  • Massaro, Joeseph

Subjects

dc:subject × 4

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/2144/49300

Chain of custody

source
Harvested from
Boston University
Base URL
open.bu.edu/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Guo, Wei. Sample size re-estimation in active controlled non-inferiority clinical trials using a frequentist approach. doctoral thesis, Boston University, 2013. https://hdl.handle.net/2144/49300