Projects

Sample Size Re-estimation in Randomised Clinical Trials
(Master's Dissertation)

Nuisance parameter is often unknown for sample size calculations which is thus estimated. To reduce the uncertainty in the estimation, parameter and sample size re-estimation is introduced. Studied the effect of misspecified nuisance parameter on sample size calculation in clinical trials. Investigated blinded and partially unblinded approach using a range of Sample Size Re-estimation (SSR) methods.

GitHub

Highlights

Generated evidence to support the use of different allocation ratios for trial arms while controlling the error rates and power. Estimated impact involves increase in the power of the trials over the non-SSR approach by up to 5-6% on average.
Additionally, the allocation ratio, with obtained evidence, can be adjusted as per financial or trial specific requirements with assurance of error rates being controlled.