Content area
Full text
Randomized controlled trials (RCTs) are the gold standard for evaluating new treatments and are required for the approval of new medicinal products.1 Unfortunately, randomization is often used to overcome poor methodology, when randomization itself is simply a tool to balance the groups being studied.2
The methodological superiority of the RCT stems from its capacity to preclude several forms of bias: by randomizing patients to treatments, groups are made statistically equivalent at baseline, at least in theory, such that observed differences at the end of the trial can be attributed with more confidence to the intervention instead of group differences. By blinding patients, doctors, evaluators, and biometricians, bias stemming from expectations and knowledge is precluded, as well as the possibility of fraud or subconscious preference on the part of raters, nurses, or statisticians. Therefore RCTs, if possible double blind, seem to be the most rigorous methodological standard available for evaluating complementary and alternative medicine (CAM).3-5 We do not know enough about CAM interventions and therefore have to employ methodologies which give us the highest possible certainty about the conclusions that we can draw from clinical studies. Internal validity is high and less likely to be challenged in RCTs, while other studies are often considered to be poor science, producing irrelevant results.6
While we do not question the accepted high internal validity produced by RCTs, we would nevertheless like to illustrate some of the weaknesses of the argument that RCTs are the primary design of choice for many types of medicine, including CAM. The main weakness derives from the fact that RCTs make presuppositions which often are not met in CAM interventions. We would also like to delineate where RCTs are the method of choice and where they are not. We argue for a staged and complementary methodological approach in evaluating CAM and finally we present some examples of outcomes research as a valuable tool in CAM evaluation.
THE PLACE OF RANDOMIZED CONTROLLED TRIALS IN THE EVALUATION OF MEDICAL INTERVENTIONS
Randomization and blinding, 2 procedures often combined in RCTs, are 2 methodological steps taken to enhance internal validity. Internal validity is usually defined as the certainty with which inferences from a study can be drawn because of its methodological soundness and rigor.7 In other words, internal validity...