Clinical trials. How should they be designed?

The fundamental principles of clinical trial design include (1) a priori formulation of a specific research question, (2) precise description of the population to be studied, (3) limitation of the effects of chance by predetermining the study hypothesis and using methods that account for multiple statistical testing, (4) calculation of the sample size necessary to answer the research question, and (5) limitation of the potential bias whenever possible through blinding and random allocation of subjects to treatment and control groups (Table 2). Although the randomized controlled trial is regarded as the gold standard for determination of efficacy of therapy, cost and sample size are frequently rate-limiting. Other techniques have been developed to supplement clinical trials and include meta-analysis, data base (observational) research, and decision analysis. In the current climate of concern over staggering health care costs and variable practice patterns, research emphases are shifting from cost containment to maximizing value of health-related services and assessing patient outcomes.