How much compliance is enough? Examining the effect of different definitions of compliance on estimates of treatment efficacy in randomized controlled trials.

A recent solution to estimating treatment efficacy in studies with non-compliance has been the development of complier averaged causal effects (CACE) estimates. Based on principal stratification, these models classify subjects who receive an adequate amount of the treatment as potential compliers and compares them to control subjects who have an equal probability of being classified as compliers if they had been randomized to treatment. No studies have systematically examined how sensitive CACE estimates are to different definitions of compliance. This study hypothesizes that incorrect definitions of compliance can bias CACE estimates and seeks to determine under what circumstances bias can occur.The standard CACE framework is extended to a partial compliance framework where there can be multiple principal strata of partial potential compliance and there is a true minimum partial potential compliance principal stratum where subjects receive the minimum treatment exposure necessary to have a relevant outcome effect. In this framework, subjects can be incorrectly classified as non-compliers and compliers. Mathematical investigations and numeric analysis suggest that when non-compliers are incorrectly classified as compliers, CACE estimates are minimally affected. On the other hand, when compliers are incorrectly classified as non-compliers, CACE estimates can be grossly inflated. These results remain when CACE estimates were calculated using the exclusion restriction or a covariate, when the exclusion restriction is true and when is false. Missing data, a common occurrence in research that is often related to noncompliance, was found to somewhat attenuate the amount of bias observed.These findings suggest that misclassifying true non-compliers as compliers might introduce a small amount of bias into CACE estimates, but that misclassifying true compliers as non-compliers may introduce a substantial amount of bias into CACE estimates. This divergence below or above the true partial compliance principal strata may provide researchers with a method of identifying the true partial compliance principal strata using sensitivity analysis. This approach was tested using data from a large cluster randomized field trial, and appeared to be able to provide an estimate of the true partial compliance minimum, but the derived estimate did not obtain statistical significance, making it of questionable value.

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

doctoral

Discipline thesis:degree_discipline

Epidemiology and Biostatistics

Grantor dc:publisher

Case Western Reserve University School of Graduate Studies

Year dc:date

2013

Author dc:creator

• Grey, Scott F.

Contributors dc:contributor

• Neuhauser, Duncan

• Biostatistics
• Causality
• Data Interpretation, Statistical
• Models, Statistical
• Randomized Controlled Trials as Topic/statistics & numerical data

Statement dc:rights

• unrestricted
• This thesis or dissertation is protected by copyright: some rights reserved. It is licensed for use under a Creative Commons license. Specific terms and permissions are available from this document's record in the OhioLINK ETD Center.

Language dc:language

English

Repository record dc:identifier

http://rave.ohiolink.edu/etdc/view?acc_num=case1363172687

OAI identifier oai:identifier

oai:etd.ohiolink.edu:case1363172687