How this calculator works
Risk is the event count divided by the number of observed people in a group. Relative risk divides the exposed risk by the unexposed risk. Values above one indicate a higher observed risk in the exposed group; values below one indicate a lower observed risk. The odds ratio is a different association measure and can differ substantially for common outcomes. These quantities require a suitable study design, comparable follow-up and consideration of confounding. They do not establish causation or diagnose an individual.
Formula
Exposed risk = a/(a + b). Unexposed risk = c/(c + d). Risk ratio = exposed risk/unexposed risk. Risk difference = exposed risk - unexposed risk. Odds ratio = ad/(bc).
Worked example
If 20 of 100 exposed people and 10 of 100 unexposed people experience an event, their risks are 20% and 10%. Relative risk is 2, and the absolute risk difference is 10 percentage points.
Advertisement
Frequently asked questions
Does relative risk two mean a 100 percentage-point increase?
No. It means twice the reference risk. Increasing from 10% to 20% is a 100% relative increase and a 10 percentage-point absolute increase.
What if the unexposed group has no events?
The finite risk ratio is undefined. A positive exposed risk over zero is displayed as infinite, while two zero risks give an undefined ratio.
Can I use a case-control table to estimate risk ratio directly?
Generally no. Case-control sampling does not supply the population risks. Its odds ratio may be appropriate under the study design.