During an evidence-based medicine seminar, medical students are evaluating various study designs and interpreting their statistical outcomes. Match each specific clinical question raised by the students to the most appropriate epidemiological measure of disease frequency or association that provides the answer.
- What is the excess incidence of lung cancer specifically caused by smoking in a cohort of patients?Attributable Risk (AR)
- What are the odds of prior asbestos exposure in patients presenting with mesothelioma compared to healthy patients?Odds Ratio (OR)
- How many patients must receive a new chemotherapy regimen instead of standard care to result in one additional case of severe neutropenia?Number Needed to Harm (NNH)
- What is the proportional decrease in the risk of myocardial infarction when comparing patients on a new lipid-lowering drug to those on placebo?Relative Risk Reduction (RRR)
- How many patients must be treated with a novel biologic agent to prevent one exacerbation of rheumatoid arthritis?Number Needed to Treat (NNT)
Answer
The clinical inquiries match their corresponding epidemiological measures as follows: excess incidence corresponds to Attributable Risk, odds of retrospective exposure corresponds to Odds Ratio, patients treated to cause an adverse event corresponds to Number Needed to Harm, proportional decrease in risk corresponds to Relative Risk Reduction, and patients treated to prevent an event corresponds to Number Needed to Treat.
Each specific clinical question inherently describes the mathematical definition of a unique epidemiological measure. Recognizing phrases like 'excess incidence' (difference in rates), 'proportional decrease' (relative change), and 'patients treated to prevent/cause' (inverses of absolute differences) allows for direct mapping to standard biostatistical metrics.
Step-by-Step Solution
Key Concept
Selecting the correct measure of disease frequency or association requires identifying the study design (e.g., case-control vs. cohort) and the clinical intent (e.g., relative benefit vs. absolute patient threshold).