The year 1948 is widely considered the beginning of the modern era of randomized clinical trials. In that year, results of a randomized trial of streptomycin for treatment of tuberculosis appeared in the Lancet. Despite resistance in some quarters, by the late 1960s the randomized trial had become widely accepted as the optimal way to evaluate new medical interventions. The ensuing decades saw an explosive growth of statistical methodology that improved both the efficiency of clinical trials and the reliability of results. Innovative trial designs continue to be introduced. Use of Bayesian methods, hampered early on by computational limitations, has become common in some clinical trials settings. The availability of genomic information has led to new trial designs geared to validate therapeutic targets as well as assess treatment effects. Emerging medical technologies such as neuroimaging and use of mobile devices to assess outcomes, and the increased accessibility of data from electronic records, will continue to pose new challenges to statisticians involved in the design, conduct and analysis of clinical trials. Finally: what will the p-value debate mean for clinical trials?