Mastering the design of clinical trials for older persons: The tension between external validity and feasibility

New interventions are needed to treat common medical conditions that are frequent causes of disability in older persons. The efficacy and effectiveness of these interventions is difficult to evaluate in older populations because traditional approaches to clinical trial design are unsatisfactory. Older persons often have comorbidity, disability and other problems that make them difficult to enroll, susceptible to adverse events, and less likely to respond, but the use of strict exclusion criteria tends to reduce the external validity of the findings. We identify three "decision" domains that should be addressed to identify the best compromise between feasibility and external validity. We propose a graphic tool that can help visualize the effect of critical decisions on external validity. Solid epidemiological foundations are needed to inform these decisions. Generalizability may be achieved gradually through the use of progressively more liberal inclusion criteria and more globally relevant distal outcome measures, after the safety and efficacy of the intervention is established in more narrowly defined samples.