Timed Essay - Bias & Confounding

Question 1

Recite general bias introduction, with quotes.

Structure:

• Quote
• Sub-groups

Answer 1

Barratt et al. essentially defines bias as “any systematic error in an epidemiological study that results in an incorrect estimate of the true effect of an exposure on the outcome of interest.”
Bias can be broadly broken down into three main sub-groups; selection bias, information bias and confounding. Each can pose differing problems for epidemiologists in both the design and analysis stages of studies.

Question 2

Recite selection bias introduction, with quotes.

Structure:

• Quote
• Simple further explanation

Answer 2

Selection bias can be viewed as “The error introduced when the study population does not represent the target population.” (Delgado-Rodriguez et al). In a nutshell this means that there is a fundamental difference between either; those who are participating in the study and those who are not or between the different groups selected for the study.

Question 3

Recite information bias introduction, with quotes.

Structure:

• Quote
• Often referred to as?
• Example
• Explanation of terms referred to.
• Another form of information bias

Answer 3

Barratt et al. informs us that information bias is caused by “systematic differences in the way data on exposure or outcome are obtained from the various study groups.”
This distorted information is often referred to as misclassified due to the error leading to a subject being categorized incorrectly within the study. For example, a frequent drinker could be wrongly classed as an infrequent drinker.
The damage this causes to validity can be somewhat determined by the type of misclassification that occurs. Paraphrasing Delgado-Rodriguez et al., differential misclassification is when the misclassification differs between groups of a study and non-differential misclassification is when it does not.
Another, more situation specific form of information bias is ecological fallacy. As the name suggests, this bias is only produced during ecological studies. In Layman’s terms it is defined as the bias produced when ecological scale results are used to make inferences at the individual level. This would assume that every person represents the average characteristics concluded from the data. For example if the mean age was 35, the study inferences would be based upon every person being 35 years old.

Question 4

Recite confounding introduction, with quotes.

Structure:

• Set-up, quote
• Example with reference
• Further example explanation

Answer 4

Confounding also provides us with an ulterior explanation to a study’s conclusion. “It occurs when an observed association is in fact distorted because the exposure is also correlated with another risk factor.” (Barratt et al.). For example, the prevalence of increasing birth order and the likelihood of Down syndrome are positively correlated. (Rothman et al.). In this scenario it was discovered that maternal age was a confounding factor, in that it was independently associated with the outcome and the exposure, but did not lie on the causal pathway between the two.