On improving research methodology in clinical trials

Stat Methods Med Res. 2008 Jun;17(3):231-42. doi: 10.1177/0962280207080639. Epub 2007 Oct 9.

Abstract

Research plays a vital role within biomedicine. Scientifically appropriate research provides a basis for appropriate medical decisions; conversely, inappropriate research may lead to flawed ;best medical practices' which, when followed, contribute to avoidable morbidity and mortality. Although an all-encompassing definition of ;appropriate medical research' is beyond the scope of this article, the concept clearly entails (among other things) that research methods be continually revised and updated as better methods become available. Despite the advent of evidence-based medicine, many research methods have become ;standard' even though there are legitimate scientific reasons to question the conclusions reached by such methods. We first illustrate prominent examples of inappropriate (yet regimented) research methods that are in widespread use. Second, as a way to improve the situation, we suggest a model of research that relies on standardized statistical analyses that individual researchers must consider as a default, but are free to challenge when they can marshal sufficient scientific evidence to demonstrate that the challenge is warranted. Third, we characterize the current system as analogous to ;unnatural selection' in the biological world and argue that our proposed model of research will enable ;natural' to replace ;unnatural' selection in the choice of research methodologies. Given the pervasiveness of inappropriate research methods, we believe that there are strong scientific and ethical reasons to create such a system, that, if properly designed, will both facilitate creativity and ensure methodological rigor while protecting the public at large from the threats posed by poor medical treatment decisions resulting from flawed research methodology.

Publication types

  • Review

MeSH terms

  • Biometry / methods*
  • Clinical Trials as Topic / methods
  • Clinical Trials as Topic / standards*
  • Clinical Trials as Topic / statistics & numerical data*
  • Humans
  • Randomized Controlled Trials as Topic / methods
  • Randomized Controlled Trials as Topic / standards
  • Randomized Controlled Trials as Topic / statistics & numerical data
  • Research / standards*
  • Research / statistics & numerical data*
  • Research Design