We are delighted to share that the next Royal Statistical Society North Eastern local group event will be a joint talk by Prof James Wason (Newcastle University) and Dr Paul Newcombe (GSK).
Abstract:
In some clinical areas, where it is important to capture multiple axes of disease progression, it is common for trials to use composite responder endpoints that classify participants as responders or non-responders based on several variables, some of which are continuous. Traditionally these endpoints have been analysed as binary, which means a large amount of information is discarded as the continuous variables are dichotomised.
Various methods, referred to as augmented binary approaches (e.g. (1–3)), have been proposed to analyse these endpoints more efficiently by utilising the continuous variables to improve the precision. The efficiency gained from doing this varies, depending on the number of continuous components and which of them divide responders from non-responders. However, the minimum observed efficiency gained is equivalent to reducing the required sample size by 30% - often it is much higher.
This talk will be split into two parts:
- JMSW will cover the original motivation and development of the augmented binary method through to it being used in academic clinical trials to reduce the sample size required.
- PJN will cover how the method is being explored for use in industry, including work on creating a Bayesian version.
The joint talk will provide an interesting case study of the process a new statistical method goes through to be used in practice. It will also highlight the different considerations for use in academic and industry settings.
References:
- Wason, J.M.S. and Seaman, S.R. (2013) “Using continuous data on tumour measurements to improve inference in phase II cancer studies,” Statistics in Medicine. Wiley. Available at: doi:10.1002/sim.5867.
- Wason, J., McMenamin, M. and Dodd, S. (2020) “Analysis of responder-based endpoints: improving power through utilising continuous components,” Trials. Springer Science and Business Media LLC. Available at: doi:10.1186/s13063-020-04353-8.
- McMenamin, M., Barrett, J.K., Berglind, A. and Wason, J.M. (2020) “Employing a latent variable framework to improve efficiency in composite endpoint analysis,” Statistical Methods in Medical Research. SAGE Publications. Available at: doi:10.1177/0962280220970986.
Book now