|
Abstract:
|
The Trial Of Preventing Hypertension study is a randomized trial conducted to examine whether early pharmacological treatment in subjects with high blood pressure prevent or delay the development of hypertension. The primary study endpoint is the development of clinical hypertension over a 4-year period. During the study, some of the subjects dropped out prematurely. Commonly, analysis for such type of data is based on the intention-to-treat (ITT) principle where assumptions about the dropout mechanism are made. These assumptions are usually not testable using the observed data, and may introduce bias. This paper presents a method for performing sensitivity analysis under a Bayesian approach for different scenarios of the dropout mechanism. The model introduces an intuitive, easy-to-understand parameterization for performing sensitivity analysis, which is applied to the randomized trial.
|
