[Skip to Navigation]
May 2, 2017

Addressing Bias and Conflict of Interest Among Biomedical Researchers

Author Affiliations
  • 1Charles Perkins Centre, Faculty of Pharmacy, University of Sydney, Sydney, Australia
JAMA. 2017;317(17):1723-1724. doi:10.1001/jama.2017.3854

Bias in research is ubiquitous and the goal of every researcher should be to reduce bias and make its potential sources transparent. Discussions of research bias are often limited to the statistical definition of bias, ie, a systematic error or deviation from the truth in results or inferences. This means biased studies will have systematically different (either larger or smaller) effect estimates than studies that are not biased. This type of bias can be reduced through methodological rigor. For example, selection bias or baseline confounding can be reduced by randomization; performance and detection biases can be reduced by blinding. Common tools for measuring bias, such as the Cochrane risk of bias tools, focus on identifying problems with the study design and execution.