sample size recalculation mid-trial based on unplanned look at data

#1
what is the current thinking on sample size recalculation mid-trial based on unplanned look at the data. suspect futility and want to either stop trial and start a new one or increase sample size. Please advise
 

hlsmith

Omega Contributor
#2
In a rigrous science based world, you planned the experiment based on background rationale supported by prior research presented in the literature and your own knowledge and work. Given this, you should wait out the study. True results may not have been discernible yet. The exception would be if it could be unethical to one of the treatment groups because of unplanned levels of detrimental effects or disproportionate effects.

In addition, given the sponsor or governing ethics committee or institutional review board, protocol deviations may need to be reported (or possibly approved prior to initiation) even when effect may not be detrimental or construed as detrimental by the investigator.

If deviations from initial protocol occur, complete transparency is mandatory when reporting findings.

Moreover, your decision depends on the context and content of the experiment along with potential policy.