A case-control study of 200 women with breast cancer and 200 age-matched controls without breast cancer was conducted to investigate the association between oral contraceptive use and breast cancer risk. The investigator calculated an Odds Ratio of 1.8 (95% CI: 1.2–2.4). Which investigation would be most appropriate to assess for potential confounding by age in this analysis?

Explanation

## Confounding Assessment in Case-Control Studies **Key Point:** Stratified analysis is the primary method to detect and assess confounding in epidemiological studies. By dividing the data into strata (subgroups) based on the potential confounder, you can compare stratum-specific measures with the crude measure. ### Stratified Analysis for Confounding Detection **High-Yield:** Confounding is suspected when: 1. The crude measure (crude OR = 1.8) differs substantially from stratum-specific ORs 2. The stratum-specific ORs are similar to each other but different from crude OR 3. The crude OR lies between the stratum-specific ORs ### Interpretation of Stratified Analysis Results | Finding | Interpretation | |---|---| | Stratum-specific ORs similar to each other but different from crude OR | **Confounding present** — age confounds the association | | Stratum-specific ORs similar to crude OR | **No confounding** — age does not confound | | Stratum-specific ORs differ markedly from each other | **Effect modification** — age modifies the effect of OCP on breast cancer | | Crude OR between stratum-specific ORs | **Confounding likely** | **Clinical Pearl:** In this study, age was used as a matching variable in the case-control design (200 age-matched controls), which should have controlled for age at the design stage. However, stratified analysis can still assess whether residual confounding by age remains. ### Why Stratified Analysis is the Investigation of Choice 1. **Non-parametric approach:** Does not assume linearity or specific functional form 2. **Transparent:** Results are easy to interpret and present 3. **Detects both confounding and effect modification:** Can identify heterogeneity across strata 4. **Appropriate for case-control data:** Works with OR directly without requiring conversion **Mnemonic: STRATUM** — **S**tratified analysis **T**ests **R**esidual **A**ssociation **T**o **U**ncover **M**odification and confounding.