How should a cohort study select an appropriate comparison group?

• A. Randomly selected from the general population without considering baseline characteristics
• B. Similar baseline characteristics except exposure; aim for similar sociodemographics and risk profiles
• C. Always unexposed with different risk profiles
• D. Only matched on age

Answer: (B)

The key idea is controlling confounding by making the comparison group as similar as possible to the exposed group at baseline. In a cohort study, you want differences in outcomes to reflect the exposure itself, not preexisting differences in risk factors. So the comparison group should have similar sociodemographic characteristics, health histories, and other risk profiles to the exposed group, except for the exposure being studied. This similarity helps ensure that any difference in outcomes is more likely due to the exposure rather than other factors.

Choosing a comparison group at random from the general population without considering baseline characteristics can leave important differences—confounders—uncontrolled, leading to biased results. An always-unexposed group with different risk profiles would also bias the comparison. Matching only on age may still leave many other confounders unbalanced. By aiming for similar baseline characteristics and risk profiles, you better isolate the effect of the exposure and enhance the study’s internal validity.