The use and pitfalls of large randomized controlled trials / 中华流行病学杂志

ABSTRACT

Modern epidemiology is the art and science of investigating quantitatively regularities or general laws regarding applied healthcare issues.The validity of epidemiological studies is primarily determined by the study design and the precision by the sample size.Large randomized controlled trial (RCT) is thus the most rigorous and most precise epidemiological study design.Due to ethical concerns,RCTs can however be used only to evaluate medical interventions.Rigorousness of study design and sample size required for a study are inversely related to the anticipated size of effect to be evaluatedthe smaller the effect,the more rigorous the study design and larger the sample size are required.Thus,large RCTs are necessary and called upon when and only when the effectiveness to be proved is relatively small;large effectiveness can be verified with small or medium-sized RCTs or even observational studies.In the stages of scientific research,large RCTs are confirmatory rather than original investigations on new hypotheses,whereas the value of a study is ultimately determined by the importance and novelty of the research question rather than methodology and the P value.Overemphasis on large RCTs has been causing1) overemphasis on interventions of small or moderate effect;2) overemphasis on confirmatory studies and on size of study and funding and weakening original creative work;3) increasing the risk of research resources,medical activities,and patients' well-being being hijacked by pharmaceutical companies.