November 1, 1992American Journal of Epidemiology
Sample Size for Studying Intermediate Endpoints within Intervention Trials or Observational Studies
Authors
Discussion
Loading...
Member takes
Implication
Methodological analysis reveals distinct sample size needs across research questions evaluating intermediate endpoints, highlighting design constraints for rare outcomes.
Key Points
- To determine the differing sample size requirements needed to evaluate intermediate biological endpoints across five distinct research questions in clinical trials and observational studies.
- Formulated sample size requirements for five key questions regarding intermediate endpoints: effect of intervention, association with risk factors, association with main outcomes, and mediation of intervention or risk factor effects.
- Applied statistical principles to an ancillary study of large bowel epithelial proliferation within the National Cancer Institute's Polyp Prevention Trial.
- Assessed the adaptability of these framework calculations to observational cohort designs evaluating rare health outcomes.
- Each of the five intermediate endpoint questions demands distinct statistical power and varying sample size calculations rather than a single unified estimate.
- Ancillary investigations nested within trials require significantly larger cohort sizes when the primary health outcome of interest occurs rarely.
Cite This Study
Freedman et al. (1992) studied this question.