This practical exercise will be a revision on relative risk while we learn confounding.
We have already learnt that relative risk helps in identifying the risk of developing a disease in an exposed group versus risk of developing a disease in the non exposed group. While we learnt its formula as;
Relative risk = A/(A+B)/C/(C+D).
Let us learn the new term confounding in detail.
Confounding is a situation in which a measure of the effect of an exposure is distorted because of the association of exposure with other factor or factors that influence the outcome of interest.
It can be described as, Factor A is a risk factor for Disease B, X is a confounder if It is a risk factor for Disease B and is also associated with Factor A.
In the study of whether coffee consumption is a risk factor for pancreatic cancer, smoking is a confounder if: 1. It is a known risk factor for pancreatic cancer 2. It is associated with coffee drinking but is not a result of coffee drinking
Coffee consumption Pancreatic Cancer
(risk factor A) (disease B)
(X is a confounder here)
Another hypothetical example can be considered to understand confounding in a step by step manner.
For instance, to study if baldness causes CHD in men, an epidemiological study recruited 10000 bald and 10000 hairy men and followed for 10 years to see for CHD.
So the risk of CHD in bald men is 4.08 times more than in hairy men. This is a strong association but can we say if this is due to causal relationship or due to confounding effect.
Now let’s stratify and study the effect of age between old and young subjects.
Older subjects (aged greater than 65 years)
RR in the older men: (750/7,500)/(100/1,000) = 1
Younger subjects (aged between 40 and 64 years)
RR in the younger men: (25/2,500)/(90/9,000) = 1
These results suggest that there is confounding by age since because when stratifying and adjusting for age the risk is changed. Had there been no effect from confounding then the risk would have been 4 even after stratifying.
Thus age is a confounder in this study.
Let us illustrate yet another example of patients with bed sores and death for students to work out.
This study was carried out in 9400 patients among people aged 60 and above. Records of patients with and without bed sores were examined for outcome. The following data is grouped in the contingent table;
Risk of bed sores and death in high medical severity group
Calculate the risk and determine whether medical severity( high & low) is a confounder ?.
Let us consider another hypothetical example of confounder in a case control study discussing diabetes, CHD and age. The variable - age is a universal confounder and its effect shall be discussed subsequently.
Draw the diagram showing causal association between the variables. With the given data, determine, whether age(<40 & ≥ 40 is a confounder ?.