Many children from a particular community coming to a hospital were detected to have acute lymphoblastic leukemia (ALL). It was assumed that it is due to the presence of cytotoxic waste in the water of that community. If a case-control study has to be done to find whether the chemical and ALL are associated, what will be taken as the control?

Explanation

## Correct Answer: A. Children from the area exposed, but unaffected with the disease In a case-control study, the control group must be selected from the **same source population** as the cases, differing only in disease status. Since cases (children with ALL) are from the exposed community, controls must also be from that same community but without the disease. This ensures that both groups have had equal opportunity for exposure, making any difference in exposure frequency between cases and controls attributable to the exposure-disease association rather than selection bias. Selecting controls from the exposed but unaffected population allows valid comparison of exposure prevalence between the two groups. This is the fundamental principle of case-control study design: cases and controls should be comparable in all respects except disease status and exposure history. Controls from the same community eliminate confounding by geographic, socioeconomic, or environmental factors unrelated to the specific chemical exposure being investigated. This design maximizes the ability to detect a true association if one exists, as per Park's Textbook of Preventive and Social Medicine guidelines on case-control methodology. ## Why the other options are wrong **B. Children from the area not exposed and affected with the disease** — This violates the fundamental principle of case-control design: controls must come from the same source population as cases. Selecting disease-affected children from a non-exposed area introduces selection bias and confounding. These children's ALL may have different etiologies unrelated to the chemical exposure being studied, making comparison invalid and preventing assessment of the exposure-disease association in the exposed community. **C. All children with ALL irrespective of exposure status** — This is not a control group at all—it is a second case group. Controls must be disease-free individuals. Comparing cases with other cases provides no information about exposure differences and cannot establish association. This fundamentally misunderstands the case-control study structure, which requires disease-discordant groups with comparable exposure opportunity. **D. Children coming to your OPD, who do not have the disease** — This represents a hospital-based control group from an undefined, non-comparable source population. It introduces Berkson's bias and selection bias because OPD attendees may have different exposure patterns, socioeconomic status, and healthcare-seeking behavior compared to the community-based cases. Controls must originate from the same community to ensure comparability and valid exposure assessment. ## High-Yield Facts - **Case-control study controls must be selected from the same source population as cases**, differing only in disease status. - **Matching controls to cases on exposure opportunity** (same community, same time period) prevents confounding and selection bias. - **Hospital-based controls introduce Berkson's bias** because hospitalized individuals have different exposure and disease patterns than the general population. - **Exposed but unaffected controls** allow direct comparison of exposure prevalence between cases and controls, enabling calculation of odds ratio. - **Selection bias in case-control studies** occurs when controls are chosen from a different source population than cases, invalidating the study. ## Mnemonics **SOURCE SAME** **S**ame source population, **A**ll comparable factors except disease, **M**atch on exposure opportunity, **E**xposed controls (unaffected), **S**election bias avoided, **A**ssociation detectable, **M**eaningful comparison, **E**xposure prevalence calculable. **CC-CONTROL Rule** **C**ommunity-based (same source), **C**omparable exposure opportunity, **C**ontrol = disease-free, **O**pportunity to be exposed (same), **N**o selection bias, **T**rue association detectable, **R**elevant to study population, **O**dds ratio valid, **L**ike cases in all but disease. ## NBE Trap NBE pairs "exposed but unaffected" with "control" to test whether students understand that controls must come from the same source population as cases. The distractor options (non-exposed cases, hospital OPD controls) exploit confusion between case-control and cohort study designs, or between source population and convenience sampling. ## Clinical Pearl In Indian public health investigations (e.g., RNTCP cluster studies or water-borne disease outbreaks), controls are always selected from the same village/community as cases to ensure both groups faced the same exposure risk. Hospital-based controls from different areas would miss the true association and mislead public health interventions. _Reference: Park's Textbook of Preventive and Social Medicine, Ch. 10 (Epidemiological Study Designs); Robbins & Cotran Pathologic Basis of Disease, Ch. 7 (Neoplasia and epidemiology)_