An epidemiologist is designing a study to investigate the association between oral contraceptive use and venous thromboembolism (VTE). She must choose between a case-control design and a cohort design. Which feature best distinguishes the choice of study design when the primary measure of association needed is relative risk rather than odds ratio?

Explanation

## Study Design and Measures of Association **Key Point:** The choice of study design determines which measure of association is **directly calculated** from the data. Relative risk (RR) is the natural output of cohort studies, while odds ratio (OR) is the natural output of case-control studies. ### Comparison of Study Designs and Their Measures | Feature | Cohort Study | Case-Control Study | |---------|--------------|--------------------| | **Direction of inquiry** | Exposed → Unexposed → Outcome (forward) | Outcome → Exposed/Unexposed (backward) | | **Directly calculates** | Relative Risk (RR) | Odds Ratio (OR) | | **Formula for RR** | Risk in exposed / Risk in unexposed | Cannot be directly calculated | | **When OR ≈ RR** | When disease is rare | When disease is rare | | **Time & cost** | Expensive, time-consuming | Efficient, economical | ### Why Cohort Studies Directly Measure RR **High-Yield:** In a cohort study: 1. Identify exposed and unexposed groups **at baseline** (before disease occurs) 2. **Follow both groups forward** over time 3. Calculate incidence in each group 4. RR = Incidence in exposed / Incidence in unexposed This is a **direct calculation** of the risk ratio. ### Why Case-Control Studies Calculate OR, Not RR **Clinical Pearl:** In a case-control study: 1. Start with people who **already have the disease** (cases) and those who don't (controls) 2. Look **backward** at exposure history 3. Cannot directly measure incidence (you don't know the denominator — total exposed in the population) 4. Must calculate OR instead: OR = (cases exposed / cases unexposed) / (controls exposed / controls unexposed) **Mnemonic:** **"Cohort = RR (forward time); Case-Control = OR (backward time)"** — The direction of time flow determines the measure. ### When Is Each Study Design Appropriate? - **Cohort design:** When you want RR directly, disease is not rare, and you have time/resources - **Case-control design:** When disease is rare, you need quick answers, or the disease has a long latency - **Approximation rule:** When disease is rare, OR ≈ RR, so case-control can approximate RR ### Application to the VTE Example If the researcher **must have RR** (not just OR), a **cohort study** is the correct choice because: - It directly measures the risk of VTE in OCP users vs. non-users - It follows both groups forward in time - VTE is not extremely rare, so RR is interpretable and clinically meaningful - RR directly answers: "What is the increased risk of VTE if a woman uses OCPs?"