End points and trial design in geriatric oncology research: a joint European organisation for research and treatment of cancer--Alliance for Clinical Trials in Oncology--International Society Of Geriatric Oncology position article.
Clinical Trials as Topic
Selecting the most appropriate end points for clinical trials is important to assess the value of new treatment strategies. Well-established end points for clinical research exist in oncology but may not be as relevant to the older cancer population because of competing risks of death and potentially increased impact of therapy on global functioning and quality of life. This article discusses specific clinical end points and their advantages and disadvantages for older individuals. Randomized or single-arm phase II trials can provide insight into the range of efficacy and toxicity in older populations but ideally need to be confirmed in phase III trials, which are unfortunately often hindered by the severe heterogeneity of the older cancer population, difficulties with selection bias depending on inclusion criteria, physician perception, and barriers in willingness to participate. All clinical trials in oncology should be without an upper age limit to allow entry of eligible older adults. In settings where so-called standard therapy is not feasible, specific trials for older patients with cancer might be required, integrating meaningful measures of outcome. Not all questions can be answered in randomized clinical trials, and large observational cohort studies or registries within the community setting should be established (preferably in parallel to randomized trials) so that treatment patterns across different settings can be compared with impact on outcome. Obligatory integration of a comparable form of geriatric assessment is recommended in future studies, and regulatory organizations such as the European Medicines Agency and US Food and Drug Administration should require adequate collection of data on efficacy and toxicity of new drugs in fit and frail elderly subpopulations.