EBM Episode 5: Bias
This is the fifth episode in our series on EBM. In it we examine biases in EBM and what they mean for healthcare practices.
First, let's take a moment to recognize that nobody is free from bias and that nobody is immune to external influences. Our biases play a big role in who we are as individuals. Perhaps the best studied form of bias is one that is related to financial conflict of interests. It is well known because it is a little bit easier to measure than others, and it is easier to track down to a direct source such as a financial contribution from a pharmaceutical. Aspirational financial incentives such as "if my results are favourable to this company, they may pay me one day" are hard to measure.
An extreme form of bias is outright cheating to change (or make up) the results of experiments to show the desired outcome. Retraction Watch reports on much misconduct which can often be traced down to bias. In some cases, the bias is that positive results will lead to a paper with more impact that will advance the author's career.
Dealing with bias, typically involves disclosure of conflicts of interests which may disqualify someone from committees and recommending bodies or be used to disclaim the results of a trial. Peer review is also meant to deal with bias by allowing external referees to point out biases. We propose that if you see biased research, you should be able to correct it and reproduce it yourself (ask us how).
For a more complete survey of bias in research see Identifying and Avoiding Bias in Research.