What are the measures of effect estimated from incidence data?

• A. Relative risk, attributable risk (risk difference), population attributable risk, population attributable fraction
• B. Sensitivity and specificity
• C. Odds ratio and hazard ratio
• D. Prevalence ratio

Answer: (A)

Incidence data track new cases over a period, so the measures of effect we estimate are about how exposure changes the risk of developing disease. The measures listed here capture that directly: relative risk (risk in the exposed group divided by risk in the unexposed), which tells you how much more (or less) likely the outcome is with exposure; attributable risk (risk difference), which measures the extra risk in the exposed group that can be attributed to the exposure; and at the population level, population attributable risk and population attributable fraction, which estimate how much of the overall incidence in the population can be attributed to the exposure and what fraction could be prevented if the exposure were eliminated. These are all grounded in incidence data because they hinge on the occurrence of new cases and the comparison of risk between groups.

Other terms come from different contexts. Sensitivity and specificity describe how well a diagnostic test identifies disease, not the effect of an exposure on incidence. Odds ratio and hazard ratio can be derived in various study designs (and the hazard ratio relates to time-to-event data), but the direct incidence-based effect measures are the relative risk, attributable risk, and the population equivalents. Prevalence ratio uses prevalence data, which reflects existing cases at a point or period rather than new cases over time.