Empirically optimising complex interventions: factorial experimental design
Location: Room 9.58a, Worsley Building, University of Leeds
Refreshments will be served from 12:30.
12:30-13:00: Annual General Meeting
13:00-13:15: Event Refreshments
13:15-14:00: Rebecca Walwyn, Leeds Institute of Clinical Trial Research
(View recording of presentation)
Complex healthcare interventions, such as psychotherapy or surgery, are defined as interventions that “contain several potentially interacting components”. Typically complex interventions tend to be represented by a single treatment variable in analyses, just as drugs would be in a conventional drug trial. Complex interventions are thus treated as black boxes, with randomised evidence built up on whether one package works relative to others. Collins et al have proposed that factorial trial designs are used to estimate the individual and combined effects of the components of a complex intervention. A potential barrier to the uptake of Collins’ proposal, however, is the recommendation that factorial designs are only used in clinical trials when it is safe to assume that there will be no interactions or when the trial is powered to detect realistic interactions. This recommendation grew out of the proposal by Peto and others to use factorial trials as a way of evaluating independent healthcare interventions efficiently, with 2x2 factorials providing ‘two independent answers for the price of one’. I will argue that this recommendation only applies if there is interest in estimating simple effects rather than main effects and interactions. I will use the AFFINITIE and ENACT programmes to motivate and illustrate