How do you compute relative risk from a 2x2 table?

• A. RR = a/c
• B. RR = [a/(a+b)] / [c/(c+d)]
• C. RR = (a+d)/(b+c)
• D. RR = [b/(a+b)] / [d/(c+d)]

Answer: (B)

The idea behind relative risk is to compare how often the outcome occurs in the exposed group versus the unexposed group. In a 2x2 table, a represents the number with the outcome among the exposed, b the number without the outcome among the exposed, c the number with the outcome among the unexposed, and d the number without the outcome among the unexposed.

Compute the risk (probability) of the outcome in each group: risk in the exposed = a/(a+b), and risk in the unexposed = c/(c+d). Relative risk is the ratio of these two risks: (a/(a+b)) / (c/(c+d)).

Example: if there are 20 cases and 80 non-cases among 100 exposed (a=20, b=80) and 10 cases and 90 non-cases among 100 unexposed (c=10, d=90), RR = (20/100) / (10/100) = 2.0, meaning the exposed have twice the risk.

Other expressions don’t reflect risk within each group: a/c compares raw case counts without considering group sizes; (a+d)/(b+c) mixes totals in a way that doesn’t measure risk in each group; [b/(a+b)] / [d/(c+d)] compares non-disease proportions, not the risk of the disease.