Suzie Cro

Jul 16, 2024 | onlinelibrary.wiley.com | Suzie Cro |James Roger |James Carpenter

1 Introduction In pivotal clinical trials with longitudinal follow-up, it is now common practice to attempt to collect full outcome data after the withdrawal of randomised treatment up until the nominal end of study. This allows the study to readily adopt a treatment-policy strategy for handling treatment withdrawal, to target the effect of the randomised treatment regardless of any treatment non-adherence [1]. As discussed by Wang et al.

Apr 17, 2024 | onlinelibrary.wiley.com | Suzie Cro |Tim Morris |James Roger |James Carpenter

REFERENCES 1, , , , , . Standard and reference-based conditional mean imputation. Pharm Stat. 2022; 21(6): 1246-1257. doi:10.1002/pst.2234 2, , . Analysis of longitudinal trials with protocol deviation: a framework for relevant, accessible assumptions and inference via multiple imputation. J Biopharm Stat. 2013; 23(6): 1352-1371. 3. Multiple-imputation inferences with uncongenial sources of input. Stat Sci. 1994; 9(4): 538-558. 4, .

Apr 17, 2024 | onlinelibrary.wiley.com | Suzie Cro |Tim Morris |James Roger |James Carpenter

Accurate frequentist performance of a method is desirable in confirmatory clinical trials, but is not sufficient on its own to justify the use of a missing data method.

Feb 6, 2024 | bmj.com | Tim Morris |Suzie Cro

Dear EditorWe would like to thank Dr Blond for giving us the opportunity to further clarify this important point. As mentioned in our original reply, the “total effect” can be used to refer to different effects, depending on how it is defined.

Jan 31, 2024 | bmj.com | Tim Morris |Suzie Cro

Dear EditorWe thank Dr Blond for their interest in our article. Our statement that a treatment policy strategy cannot be used to handle truncating events is consistent with the estimands defined in the Young et al paper, which considers a number of estimands in the presence of competing events, but does not consider a treatment policy strategy (which, as described in our paper, cannot be defined in the presence of competing events). Young et al.