Addressing Challenges in Estimating Treatment Effects in Oncology Clinical Trials
Access status:
USyd Access
Type
ThesisThesis type
Doctor of PhilosophyAuthor/s
Soon, Yu YangAbstract
Accurate estimation of treatment effects in randomised controlled trials (RCTs) is crucial for guiding clinical care, informing regulatory decisions, and optimising healthcare resource allocation. In oncology RCTs, where benefits are often modest but meaningful, small estimation ...
See moreAccurate estimation of treatment effects in randomised controlled trials (RCTs) is crucial for guiding clinical care, informing regulatory decisions, and optimising healthcare resource allocation. In oncology RCTs, where benefits are often modest but meaningful, small estimation errors can substantially impact clinical and policy decisions. This thesis examines four scenarios that challenge accurate treatment effect estimation: (1) non-randomised comparisons of concomitant therapies, (2) treatment switching after disease progression, (3) pooling of phase 2 and 3 data, and (4) early declaration of superiority at interim analysis. Chapter 3 used target trial emulation to assess non-randomised comparisons in metastatic prostate cancer trials, finding no clear survival benefit from adding chemotherapy to enzalutamide. It provides a framework for such analyses but cautions against overreliance without randomised confirmation. Chapter 4 examined treatment switching in two Lu-177 PSMA trials. Advanced statistical adjustments showed switching had minimal impact on effect estimates; survival differences arose from comparator choices. Careful comparator selection and statistical adjustments are crucial when switching occurs. Chapter 5 investigated pooling phase 2/3 data in INTEGRATE trials. Adjusting for treatment switching and winner’s curse preserved consistency between phases. Prospective pooling under closed testing is feasible if misalignment risks are evaluated. Chapter 6 analyzed interim superiority declarations in 71 RCTs, finding overestimation bias, especially early. Penalised methods mitigated bias. Case studies (ENZAMET, KEYNOTE-426) highlighted how trial dynamics affect interpretation. Interim timing and estimator choice are critical for robust decisions. This thesis integrates advanced statistics into trial design and analysis, thereby improving the estimation of treatment effects for more reliable clinical and policy decisions.
See less
See moreAccurate estimation of treatment effects in randomised controlled trials (RCTs) is crucial for guiding clinical care, informing regulatory decisions, and optimising healthcare resource allocation. In oncology RCTs, where benefits are often modest but meaningful, small estimation errors can substantially impact clinical and policy decisions. This thesis examines four scenarios that challenge accurate treatment effect estimation: (1) non-randomised comparisons of concomitant therapies, (2) treatment switching after disease progression, (3) pooling of phase 2 and 3 data, and (4) early declaration of superiority at interim analysis. Chapter 3 used target trial emulation to assess non-randomised comparisons in metastatic prostate cancer trials, finding no clear survival benefit from adding chemotherapy to enzalutamide. It provides a framework for such analyses but cautions against overreliance without randomised confirmation. Chapter 4 examined treatment switching in two Lu-177 PSMA trials. Advanced statistical adjustments showed switching had minimal impact on effect estimates; survival differences arose from comparator choices. Careful comparator selection and statistical adjustments are crucial when switching occurs. Chapter 5 investigated pooling phase 2/3 data in INTEGRATE trials. Adjusting for treatment switching and winner’s curse preserved consistency between phases. Prospective pooling under closed testing is feasible if misalignment risks are evaluated. Chapter 6 analyzed interim superiority declarations in 71 RCTs, finding overestimation bias, especially early. Penalised methods mitigated bias. Case studies (ENZAMET, KEYNOTE-426) highlighted how trial dynamics affect interpretation. Interim timing and estimator choice are critical for robust decisions. This thesis integrates advanced statistics into trial design and analysis, thereby improving the estimation of treatment effects for more reliable clinical and policy decisions.
See less
Date
2025Rights 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 HealthDepartment, Discipline or Centre
NHMRC Clinical Trials CentreAwarding institution
The University of SydneyShare