An epidemiologist in Mumbai is studying the relationship between maternal anemia during pregnancy and adverse birth outcomes (low birth weight, preterm delivery). She recruits 2000 pregnant women at their first antenatal visit, measures hemoglobin levels, and then follows them prospectively until delivery, recording outcomes. In contrast, a colleague in Bangalore begins with 800 mothers of low-birth-weight infants and 800 mothers of normal-birth-weight infants, then retrospectively assesses their hemoglobin records from pregnancy. Which statement best compares these two approaches?

Explanation

## Study Design Classification **Key Point:** The critical distinction is the **starting point**: - **Mumbai study (Cohort):** Starts with **exposure status** (hemoglobin level) → follows forward to **disease outcome** (birth weight) - **Bangalore study (Case-Control):** Starts with **disease status** (birth weight) → looks backward to **exposure** (hemoglobin) ## Detailed Comparison | Aspect | Mumbai (Cohort) | Bangalore (Case-Control) | |--------|-----------------|-------------------------| | **Starting point** | Exposed vs. unexposed pregnant women | Low-birth-weight vs. normal-birth-weight infants | | **Direction** | Forward in time (prospective) | Backward in time (retrospective) | | **Measure of association** | Relative Risk (RR) | Odds Ratio (OR) | | **Can calculate incidence?** | Yes (direct measurement) | No (starts with cases) | | **Time to completion** | Long (months to years of follow-up) | Short (retrospective data collection) | | **Cost** | High (prospective follow-up) | Low–moderate (chart review) | | **Selection bias risk** | Loss to follow-up | Selection of cases/controls | | **Recall bias risk** | Low (exposure measured prospectively) | High (maternal recall of pregnancy) | **High-Yield:** Both are valid designs, but they answer the question differently: - **Cohort RR** = Risk in exposed ÷ Risk in unexposed = $\frac{a/(a+b)}{c/(c+d)}$ - **Case-Control OR** = Odds of exposure in cases ÷ Odds of exposure in controls = $\frac{a \times d}{b \times c}$ ## Why Cohort for Maternal Anemia? **Clinical Pearl:** For common exposures (anemia affects ~30% of pregnant women in India) and common outcomes (low birth weight ~20%), **cohort studies are preferred** because: 1. You can measure **true incidence** of adverse outcomes in exposed vs. unexposed groups 2. You can calculate **relative risk directly** (more intuitive than OR) 3. You establish **temporal sequence** unambiguously (hemoglobin measured before delivery) 4. You minimize **recall bias** (exposure measured prospectively) **Mnemonic: RARE disease → Case-Control; COMMON disease → Cohort** - Anemia in pregnancy: ~30% → Common → Cohort preferred - If studying a rare genetic cause of low birth weight → Case-control would be better ## Why Case-Control Has Higher Recall Bias Here The Bangalore study relies on retrospective hemoglobin records or maternal recall of anemia symptoms from months earlier. Mothers of low-birth-weight infants may over-report or over-remember anemia (because they are seeking explanations for their baby's condition), introducing **recall bias**. [cite:Park 26e Ch 10; Epidemiology and Biostatistics]