OBJECTIVES: Traditional randomized trials are the gold-standard for clinical research but may be long, expensive and, especially for CER, inefficient in providing evidence to practicing clinicians.  We, the REsearch in ADAptive methods for Pragmatic Trials (RE-ADAPT) investigators sought to test efficiency gains in CER trials by redesigning, then re-executing ALLHAT using adaptive trial methodologies. METHODS: ALLHAT was a traditionally designed, large comparative trial looking at long-term cardiovascular outcomes with different classes of anti-hypertensive drugs compared with diuretics.  We prospectively developed seven Bayesian adaptive designs incorporating combinations of early success/futility stopping, adaptive randomization, and arm dropping.  Using only information available when ALLHAT was designed, we examined operating characteristics of the designs to evaluate clinical and economic efficiencies (e.g. shorter trials, greater allocation to superior treatments).  One design was chosen a priorias the one we would have adopted for ALLHAT, including adaptive randomization with the potential for early stopping.  The design was executed using the ALLHAT data, preserving accrual order and re-sampling patients when adaptive randomization called for additional patients in a given arm.   RESULTS: ALLHAT enrolled 42,418 patients in four years, following patients for an additional four years and showed that none of the comparator treatments were superior to diuretics.  The recommended adaptive design reached the same conclusion as the original ALLHAT trial but would have stopped nearly one year earlier.  Also, the simulated trial randomized 28% more patients to calcium channel blockers, because it initially outperformed other arms during the accrual stage of the trial. CONCLUSIONS: CER trials are important to clinical research. However, conventional methods – even with early stopping rules as in ALLHAT– can be so time-consuming, that the clinical relevance of the study may be diminished before the results are published.  Adaptive methods, tailored to CER, can be more efficient, and produce more timely evidence.