A 52-year-old man from rural Maharashtra presents to a PHC with a 3-month history of productive cough and low-grade fever. Chest X-ray shows an upper lobe infiltrate. Sputum smear microscopy is positive for acid-fast bacilli. He is started on anti-TB therapy. During case investigation, the health worker notices that the patient's wife (48 years old) and his eldest son (28 years old) also have similar respiratory symptoms and are sputum-positive. The health worker concludes that TB transmission within the family is 100% and reports this as evidence of high household transmission risk. However, the district epidemiologist suspects this observation may be affected by a systematic error in the investigation process. Which type of bias is most likely affecting this conclusion?

Explanation

## Understanding the Scenario The health worker identified TB in the index case and then actively searched for TB among household contacts, finding cases in the wife and son. This targeted investigation of household members creates a systematically different case-finding process compared to what would occur in the general population. ## Type of Bias: Surveillance Bias (Ascertainment Bias) **Key Point:** Surveillance bias (also called ascertainment bias) occurs when cases are identified through differential detection or investigation intensity. In this case: 1. The index case was identified through routine clinic attendance (passive case-finding) 2. Household contacts were then actively investigated with heightened suspicion 3. This active case-finding in contacts is more intensive than passive surveillance in the general population 4. The apparent "100% transmission" reflects differential case detection, not true transmission risk ## Why This Matters Epidemiologically **High-Yield:** Surveillance bias inflates the apparent association between exposure (household contact) and disease (TB). The true household transmission risk is likely much lower — many exposed household members may have subclinical infection or may not develop active disease. **Clinical Pearl:** This is why TB contact tracing data must be interpreted cautiously. Active case-finding among contacts artificially increases the apparent secondary attack rate compared to passive surveillance in the general population. ## Distinguishing from Other Biases | Bias Type | Mechanism | Why Not This Case | |-----------|-----------|-------------------| | **Berkson's bias** | Selection bias in hospital/clinic populations | No hospital-based selection here; this is community-based | | **Healthy worker effect** | Occupational cohorts appear healthier | No occupational cohort; this is household contacts | | **Attrition bias** | Loss to follow-up in cohort studies | No cohort with differential follow-up; cross-sectional investigation | **Mnemonic:** **SAFE** — Surveillance bias affects Ascertainment, creating False Elevation of association. The detection method (active vs. passive) differs between groups. ## Correct Interpretation The epidemiologist should adjust the analysis by: - Recognizing that active case-finding in household contacts is more sensitive than passive surveillance - Comparing household transmission rates only to similarly actively-detected contacts in other families - Not comparing to general population passive surveillance rates - Adjusting for detection intensity when calculating secondary attack rates