|
Abstract:
|
The R1 estimand E9 addendum addresses a gap between standard assumptions and clinical experience. Missing values, currently modeled as noise, are often strongly correlated with treatment outcomes. The estimand concept reconceptualizes patient dropout etc. as intercurrent events, informative qualitative data descriptive of or counterfactual to the treatment outcomes of interest. As current methods, surveyed, depend on strong assumptions, the guidance emphasizes appropriate visit schedules and monitoring practices, and sensitivity analyses to detect departures from assumptions. The recognition of gaps between traditional assumptions and clinical reality provides new opportunities for statistical leadership, from selecting endpoints and making scheduling and monitoring choices in trial designs to developing new, more robust methods embracing the paradigm of informative intercurrent events. It increases the value of conceptual understanding, interfunctional communications, clear goals, proactive planning and design, prevention, and statistics conceived as an engineering discipline, vs. focus on post-hoc analysis/adjustment. Key opportunities for leadership with examples are outlined.
|
