The role of subgroups and sub-populations in drug development and drug regulation

Garrett, Andrew (2006). The role of subgroups and sub-populations in drug development and drug regulation. PhD thesis The Open University.



This thesis addresses the role of subgroups and sub-populations in drug development and regulation and includes the critical appraisal of regulatory guidance.

Chapter One introduces clinical trial methodology and describes the current regulatory environment.

In Chapter Two, randomisation is reviewed in relation to unbiased estimation of treatment differences and the impact of data exclusion to form subsets is described.

Simpson's paradox (SP) is considered in Chapter Three. Randomisation is shown to protect against SP, while a treatment by factor interaction is not required. Balance is redefined for the odds formulation leading to identical unconditional and conditional parameters. The chances of SP (and less extreme inconsistencies) occurring are quantified using simulation, with varying sample size.

Chapter Four considers treatment by subgroup interactions. Suggestions regarding the magnitude of a clinically relevant interaction are presented while a simple Bayesian approach to evaluate interactions using margins for the interaction parameter is applied to published data.

Chapter Five considers non-inferiority in relation to sub-populations and covariate adjustment for binary outcomes. It is shown that the Per Protocol population is not necessarily conservative and simulation is used to demonstrate the impact on the type I and II errors. Using simulations it is shown that an increase in the type I error occurs if an important covariate is excluded from the logistic model when testing for non-inferiority.

Chapter Six is directed towards the sub-population of children. The impact of off-label treatment is discussed in relation to the ethics of placebo-controlled trials, together with the importance of randomisation in evaluating long-term safety.

In Chapter Seven, a therapeutic area is selected to ilJustrate the challenges raised during the previous chapters and wording changes to current regulatory guidelines are proposed.The thesis closes with personal thoughts regarding the future potential for individualised treatment.

Viewing alternatives

Download history


Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions