A study comparing people with a condition to similar people without it, looking backward for exposures.
An observational design that starts with an outcome (cases with a condition or injury) and a comparison group without it (controls), then looks retrospectively at differences in prior exposures or risk factors. It's efficient for rare outcomes and useful for hypothesis generation, but is prone to recall and selection bias, and can only establish association, not causation. Confounding must be carefully controlled through matching or statistical adjustment, and results should be interpreted with appropriate caution relative to prospective designs.
This guide was auto-drafted and is pending editorial review.