Get access

Bias in (complementary) medical research

Authors


Abstract

Evidence-based decision-making is entirely dependent on identifying all relevant studies. However, a significant proportion of clinical trials are never published. Most forms of bias operate equally in mainstream and CAM research although some forms seem to be particularly relevant to CAM such as language bias. It seems clear that the dissemination of mainstream and CAM research findings is a biased process. Whether CAM is more prone to bias than mainstream medicine is debatable. It is important that the risk of bias is taken into account when involved in evidence-based decision-making in order to avoid unnecessary patient suffering and waste of limited resources.

An important step towards reducing the risk of publication bias is the prospective registration of all studies at the stage of study inception. The policy of compulsory trial registration was adopted by the International Committee of Medical Journals Editors (ICMJE) in 2004 which was influential but not sufficient to entirely prevent publication bias. Registered trials need to be publicly accessible in order to facilitate the identification and location of ongoing or unpublished trials. One such trial register has been established at the Institute of Medical Biometry and Statistics, University of Freiburg in close collaboration with the German Cochrane Centre. It is a primary register in the World Health Organization (WHO) registry network meeting the specific criteria of the ICMJE. Other trial registers exist for trials conducted in the USA or China. Ultimately, compulsory registering of studies with the additional help of government regulations and compulsory publishing may be required to prevent publication bias.

Ancillary