Re: Statistical methods to compare functional outcomes in randomized controlled trials with high mortality
Colantuoini and colleagues describe statistical methods for the analysis of randomised controlled trials when the intervention may reduce mortality, but where functional outcomes are also measured. The concern here is that the functional outcomes might be truncated due to death in the control group if the intervention improves survival. Thus, worse or unchanged functional outcomes in the intervention group might be the result of survival of patients who, without the intervention, would have died. The three methods they describe are to simply compare functional results and accept the above possibility; to conduct a type of sensitivity analysis; or to construct a composite endpoint that includes both mortality and functional outcomes. From a decision-making point of view I would like to propose a further alternative. My proposal is to consolidate the mean utility of the functional outcome with deaths to calculate a QALY (or more strictly a DALY) difference. In essence this is a refinement of the composite endpoint method proposed by the authors.
1. Colantuoni E, Scharfstein DO, Wang C, Hashem MD, Leroux A, Needham DM, Girard TD. Statistical methods to compare functional outcomes in randomized controlled trials with high mortality. BMJ. 2018; j5748.
Competing interests: No competing interests