Optimising analysis of trial data for translational gain in a randomised controlled trial of patient-led surveillance for subsequent melanoma (MEL-SELF)
Access status:
USyd Access
Type
ThesisThesis type
Doctor of PhilosophyAuthor/s
Medcalf, EllieAbstract
Randomised controlled trials (RCTs) are regarded as the gold standard in study design to understand the causal relationship between an intervention and an outcome. Thus, it is critical that the use and translation of trial evidence is maximised. However, post-randomisation events, ...
See moreRandomised controlled trials (RCTs) are regarded as the gold standard in study design to understand the causal relationship between an intervention and an outcome. Thus, it is critical that the use and translation of trial evidence is maximised. However, post-randomisation events, such as missing outcome data and nonadherence to treatment, can compromise the validity and usefulness of trial evidence. This thesis investigates and evaluates statistical and epidemiological approaches to address missing outcome data and nonadherence at the analysis stage of RCTs. It aims to define and specify estimands and estimators that will assist in generating the most informative evidence about the MELanoma SELF surveillance trial (MEL-SELF), an RCT assessing a new model of melanoma surveillance (patient-led surveillance) versus standard care for patients with early-stage melanoma. This thesis includes 1) two methodological scoping reviews summarising: a) estimators to handle missing outcome data when estimating the intention-to-treat (ITT) estimand and b) non-ITT estimands that estimate the effect of adhering to treatment (e.g., per-protocol (PP) estimand) and estimators used to estimate these estimands; 2) description of the baseline characteristics of the MEL-SELF trial population, which defines the clinical population to which estimators will be applied, and to which trial evidence will be relevant; 3) simulation study evaluating six missing outcome data estimators and 4) update to the MEL-SELF statistical analysis plan, describing estimators that will be used to address missing outcome data (for ITT estimation) and nonadherence (for PP estimation). This thesis highlights that there is a significant gap between the development of new methodological approaches for handling post-randomisation events and their practical application in RCTs. It bridges this gap by providing clear, practical guidance for trialists on how best to address these issues at the analysis stage of a trial.
See less
See moreRandomised controlled trials (RCTs) are regarded as the gold standard in study design to understand the causal relationship between an intervention and an outcome. Thus, it is critical that the use and translation of trial evidence is maximised. However, post-randomisation events, such as missing outcome data and nonadherence to treatment, can compromise the validity and usefulness of trial evidence. This thesis investigates and evaluates statistical and epidemiological approaches to address missing outcome data and nonadherence at the analysis stage of RCTs. It aims to define and specify estimands and estimators that will assist in generating the most informative evidence about the MELanoma SELF surveillance trial (MEL-SELF), an RCT assessing a new model of melanoma surveillance (patient-led surveillance) versus standard care for patients with early-stage melanoma. This thesis includes 1) two methodological scoping reviews summarising: a) estimators to handle missing outcome data when estimating the intention-to-treat (ITT) estimand and b) non-ITT estimands that estimate the effect of adhering to treatment (e.g., per-protocol (PP) estimand) and estimators used to estimate these estimands; 2) description of the baseline characteristics of the MEL-SELF trial population, which defines the clinical population to which estimators will be applied, and to which trial evidence will be relevant; 3) simulation study evaluating six missing outcome data estimators and 4) update to the MEL-SELF statistical analysis plan, describing estimators that will be used to address missing outcome data (for ITT estimation) and nonadherence (for PP estimation). This thesis highlights that there is a significant gap between the development of new methodological approaches for handling post-randomisation events and their practical application in RCTs. It bridges this gap by providing clear, practical guidance for trialists on how best to address these issues at the analysis stage of a trial.
See less
Date
2026Rights statement
The author retains copyright of this thesis. It may only be used for the purposes of research and study. It must not be used for any other purposes and may not be transmitted or shared with others without prior permission.Faculty/School
Faculty of Medicine and Health, The University of Sydney School of Public HealthAwarding institution
The University of SydneyShare