A cohort study of 10,000 Indian smokers and 10,000 non-smokers followed for 10 years found 400 cases of lung cancer in smokers and 50 cases in non-smokers. What is the relative risk of lung cancer in smokers compared to non-smokers, and why is this measure preferable to odds ratio in this study design?

Explanation

## Calculating Relative Risk in Cohort Studies ### Step-by-Step RR Calculation **Incidence in smokers (exposed group):** $$\text{Risk}_\text{exposed} = \frac{400}{10,000} = 0.04 \text{ (or 4%)}$$ **Incidence in non-smokers (unexposed group):** $$\text{Risk}_\text{unexposed} = \frac{50}{10,000} = 0.005 \text{ (or 0.5%)}$$ **Relative Risk:** $$RR = \frac{\text{Risk}_\text{exposed}}{\text{Risk}_\text{unexposed}} = \frac{0.04}{0.005} = 8.0$$ **Key Point:** An RR of 8.0 means smokers are **8 times more likely** to develop lung cancer than non-smokers over the 10-year follow-up period. ### Why RR is the Gold Standard for Cohort Studies | Aspect | Relative Risk (RR) | Odds Ratio (OR) | |--------|-------------------|----------------| | **Study Design** | Cohort (prospective) | Case-control (retrospective) | | **What it measures** | Ratio of incidence/risk | Ratio of odds of exposure | | **Interpretation** | Direct probability comparison | Indirect association measure | | **When to use** | Disease incidence is measurable | Disease incidence cannot be measured | | **Validity** | Always valid in cohort studies | Approximates RR only if disease is rare | **High-Yield:** In a prospective cohort study, you can directly calculate cumulative incidence (risk) in both exposed and unexposed groups, making RR the natural and most interpretable measure of association. **Clinical Pearl:** An RR of 8.0 for smoking and lung cancer is consistent with real-world epidemiological data and represents a very strong causal association. This magnitude of risk is one reason smoking cessation is a cornerstone of public health in India. **Mnemonic:** **COHOR-RR** = **CO**hort studies use **RR**; **CASE-OR** = **CASE**-control studies use **OR**. ### Why OR Would Be Inappropriate Here Although you *could* calculate an OR from this cohort data, it would: - Obscure the true incidence rates - Be harder to interpret clinically - Overestimate the association if disease prevalence is high (as it is here: 4% in smokers) Since the disease is not rare (4% incidence in exposed group), OR ≠ RR, and RR is far more meaningful.