Bias

CHAPTER 15 Bias




The science of statistics has made great progress in this century, but progress has been accompanied by a corresponding increase in the misuse of statistics.


—Robert Hooke1


Bias is a normal human trait. It is the expectation of outcomes in certain scenarios, based on our prior experiences. It molds our behavior by encouraging us to choose situations we find comfortable and predictable. It unconsciously influences our decisions to act in a certain way.


For instance, if you were to attend a banquet it is very likely that you would choose to sit at a table with people you know. You feel comfortable with these people and, from past experience, you subconsciously predict that you will have an enjoyable time. In this way, your decision on where to sit is biased.


Now take the example of a surgeon who is about to perform a redo herniorrhaphy. This is a complex surgical procedure that requires a revision of an original hernia operation. Imagine there are two types of procedures for this operation, and the surgeon is skilled in both. She asks two of her colleagues for their opinion on the better type of procedure. One is a senior surgeon with extensive experience and the other is a junior partner with less experience.


Based on their respective experience levels, the senior surgeon recommends the first procedure, in opposition to the junior surgeon, who recommends the second. The inquiring surgeon is biased (but not to her patient’s detriment) in that she will most likely trust the opinion of the senior surgeon. However, a valid study is available that shows the senior surgeon’s preference has a higher rate of complications, all other things being equal. The inquiring surgeon must make a decision in the best interest of her patient but, if she still chooses to perform the senior surgeon’s preferred operation, she is letting bias influence her decision and is putting her patient at a potential disadvantage.


Bias is often subconscious and unintentional. However, since people who seek medical care must rely on the decisions of their health care professionals, bias in these situations can result in less favorable outcomes. Even though the layperson’s interpretation of bias carries a negative connotation akin to prejudice, bias is not a conscious decision to intentionally provide substandard care. Rather, it is acting within a comfort zone based on prior experience with similar patients. In order to minimize bias in the practice of medicine, we must acknowledge its existence and be willing to accept evidence that may alter the way we think.


Experiments are designed to re-create reality. We take a sample group of subjects and study them under controlled conditions, but we try not to alter the real situation that exists so that any conclusions we draw from our observations are valid and can be applied to the entire population. However, experiments are not real situations; they introduce elements that do not reflect the real situation with 100% accuracy. Bias in research is any influence that acts to make the observed results nonrepresentative of the true effect that is under investigation. Again, bias in research is not intentional but a biased study can result in an erroneous conclusion.



• Bias in statistics is an influence, possibly unrecognized, that introduces error into the process.


Certain experimental designs are more subject to bias than others, based on the way data are collected. The randomized, blinded, controlled trial was designed to minimize bias and is therefore given more credence than other trial designs. Besides trial design, bias can affect a research study in other ways. It can occur at any point—from the inception of the idea, throughout the trial process of sample selection and data collection, in the analysis, and even in the realm of publication. Bias in research is more detrimental than individual bias because the publicized results can ultimately influence a large number of practitioners.


The following are some areas where you might identify a source of bias. When we critique the literature, we should recognize the potential sources of bias and understand that they could ultimately have an effect on the conclusions upon which we base our recommendations.



Selection bias

Related posts:

  1. Types of Statistical Tests
  2. Ethics in Medical Research
  3. Measures of Disease
  4. Mathematical Principles

Stay updated, free articles. Join our Telegram channel
Tags:
Jun 18, 2016 | Posted by in BIOCHEMISTRY | Comments Off on Bias

Full access? Get Clinical Tree
Get Clinical Tree app for offline access