An epidemiologist at a tertiary centre in Mumbai is reviewing a published study that investigated the association between occupational asbestos exposure and lung cancer. The study enrolled 200 workers with confirmed lung cancer and 200 workers without lung cancer, then retrospectively assessed their past asbestos exposure using employment records and interviews. The researchers calculated an odds ratio of 8.5 (95% CI: 5.2–13.8). A colleague questions whether this design was appropriate for establishing causation. What is the key limitation of this study design that must be acknowledged when interpreting the findings?

Explanation

## Key Limitation of Case-Control Studies: Recall Bias ### Why Recall Bias is the Critical Issue Here **Key Point:** In case-control studies, cases (those with disease) are more motivated and more likely to accurately recall and report past exposures than disease-free controls. This differential recall introduces systematic error that can artificially inflate the odds ratio. **High-Yield:** Recall bias is the MOST important limitation specific to case-control studies and is frequently tested in NEET PG because it directly affects the validity of findings. ### Mechanism of Recall Bias in This Study ```mermaid flowchart TD A[Lung cancer diagnosed]:::outcome --> B[Patient motivated to recall<br/>past exposures]:::action B --> C[Detailed memory of<br/>asbestos contact]:::action C --> D[Over-reporting of exposure]:::urgent E[No lung cancer]:::outcome --> F[Minimal motivation to recall<br/>occupational history]:::action F --> G[Vague or incomplete<br/>exposure memory]:::action G --> H[Under-reporting of exposure]:::urgent D --> I[Inflated Odds Ratio]:::urgent H --> I ``` **Clinical Pearl:** Workers with lung cancer may unconsciously (or consciously) over-report asbestos exposure because they are searching for an explanation for their disease ("rumination bias"). Healthy controls, lacking this motivation, may forget or minimize past exposures, creating a false impression of stronger association. ### Why the Other Options Are Incorrect or Less Critical | Limitation | Why It's Not the Best Answer | |---|---| | **Odds ratio vs relative risk** | True, but OR approximates RR when outcome is rare. Lung cancer prevalence in occupational cohorts is low, so OR ≈ RR. This is a technical limitation, not a validity threat. | | **Recall bias** | **✓ CORRECT — Directly threatens validity** | | **Sample size (n=200 each)** | With OR=8.5 and 95% CI excluding 1.0, the study is adequately powered. Small sample size is not the issue here. | | **Temporal relationship** | Case-control studies are inherently retrospective, but temporal sequence can still be inferred if exposure preceded disease onset. This is a design feature, not a flaw in this context. | ### How Recall Bias Affects the Odds Ratio **Mnemonic — RECALL BIAS IN CASE-CONTROL:** - **R**etrospecive assessment of exposure - **E**xposed cases remember better (motivated) - **C**ontrols forget or minimize exposure - **A**rtificially inflated association - **L**oss of validity - **L**ikelihood of bias: HIGH ### Mitigation Strategies (Not Asked, But High-Yield) 1. Use objective exposure records (employment files, dosimetry) instead of interviews alone 2. Standardize questionnaires and interview techniques 3. Blind interviewers to case/control status 4. Use biomarkers or environmental measurements when possible 5. Conduct sensitivity analysis assuming various levels of recall bias **Warning:** An OR of 8.5 is very strong, but recall bias could inflate a true OR of 4–5 to 8.5. The study is valid in design but vulnerable to this specific bias. [cite:Park 26e Ch 9]