Keywords: Special populations, geriatric patients, disproportionate stratified sampling.
It is generally recognized that in the development of therapeutic agents there is an opportunity to improve the evidence base for treating older patients. The 2003 ICH-E7 guideline along with subsequent regional regulatory documents and recommendations by academic working parties laid out evidence requirements to assess similarity in benefit-risk profile between older and younger patients and to ensure generalizability of clinical trial findings to the population of intended users of a drug. In spite of these efforts, current literature indicates that the representation of older, frail and special populations remains artificially and chronically low in clinical trials. At the same time, with increasing popularity of enrichment and adaptive designs, disproportionate stratified sampling has become progressively common in modern clinical trials. This allows unique and scientifically desirable opportunities for conducting studies in a population more frail than the “use-population” of a therapeutic, even if it is at the expense of full generalizability. In this talk I will discuss examples from three disease areas that are highly relevant in this context – cardiovascular disorders, degenerative neurologic disorders and cancer. I will also present key considerations and novel statistical methodology that allow drawing generalizable inference from clinical trials, especially those involving time to event endpoints, that involve disproportionate stratified sampling.