Multicenter beta-blocker trials: certainties

In general, therapeutic evaluation is based on a comparison between the 'treated' and the 'untreated'. The validity of the inference is based on the methods used in the formation of such groups. Randomized treatment allocation avoids selection bias and leads to interpretable statistical tests of significance. In addition, it defines the point in time at which follow-up starts. The use of placebo in a double-blind manner achieves comparability of information and also makes differences in the use of concomitant medication during follow-up interpretable as effects of the drug under study. Multicenter cooperation may enhance the representativeness of the data for the kind of patient to which the results may apply and makes it possible to check results for consistency between the clinics. These features of the design of recent beta-blocker trials in post-myocardial infarction patients enhance confidence in the results.