A dietary survey report indicates that the 5-year survival rate for breast cancer has increased following the introduction of a new screening method. However, autopsy data shows no change in overall mortality. What type of bias does this scenario illustrate?

Explanation

## Correct Answer: D. Lead time bias Lead time bias occurs when earlier detection of disease (through screening) appears to improve survival time without actually changing the natural history or mortality rate. In this scenario, the new screening method detects breast cancer earlier in its course. Patients now have a longer interval between diagnosis and death—not because they live longer, but because the disease was identified sooner. Autopsy data showing unchanged overall mortality is the critical clue: if screening truly improved outcomes, mortality should decrease. Instead, the apparent 5-year survival improvement reflects only the earlier point at which the disease became known. The patient's lifespan remains unchanged; only the "survival clock" starts earlier. This is a fundamental epidemiological principle in cancer screening evaluation. According to Park's Textbook of Preventive and Social Medicine, lead time bias is a major pitfall in interpreting screening efficacy and must be distinguished from genuine mortality reduction. In Indian cancer registries, this distinction is crucial when evaluating national screening programs (e.g., cervical cancer screening under NRHM), where apparent survival gains may mask unchanged population mortality. ## Why the other options are wrong **A. Berksonian bias** — Berksonian bias occurs in case-control studies when selection of cases and controls from hospital populations creates spurious associations between two diseases. It has no relevance to screening-induced apparent survival improvement. The scenario describes a population-level screening outcome, not a hospital-based case-control selection problem. **B. Detection bias** — Detection bias refers to differential accuracy in identifying disease outcomes between comparison groups (e.g., more intensive follow-up in one arm detects more events). While screening does detect more cases, the key finding here is unchanged autopsy mortality—indicating the bias is not about differential detection of outcomes, but about the timing of diagnosis relative to a fixed endpoint. **C. Survival bias** — Survival bias (or survivor bias) occurs when analysis excludes individuals who are no longer available for follow-up, leading to skewed estimates. This scenario involves complete follow-up to death (autopsy data available), so exclusion of non-survivors is not the mechanism. The problem is the earlier starting point of the survival clock, not loss of follow-up data. ## High-Yield Facts - **Lead time bias** = earlier diagnosis without mortality benefit; survival time increases but lifespan unchanged. - **Autopsy data unchanged mortality** = gold standard proof that screening did not reduce actual deaths, only shifted diagnosis timing. - **5-year survival vs. mortality rate** = survival is time from diagnosis to death; mortality is deaths per population; screening can inflate survival without changing mortality. - **Lead time bias in Indian cancer screening** = major concern in cervical cancer (Pap smear) and breast cancer programs; must use mortality reduction, not survival improvement, as efficacy measure. - **Screening efficacy evaluation** = always compare mortality rates (population-level) rather than survival rates (diagnosis-level) to avoid lead time bias. ## Mnemonics **LEAD TIME = Earlier Diagnosis, Same Death** LEAD = Longer Apparent Duration (from diagnosis to death, but actual lifespan unchanged). Think: 'I diagnosed you 2 years earlier, so you appear to survive 2 years longer—but you die on the same calendar date.' Use this when survival improves but mortality doesn't. **Screening Bias Discriminator** If **autopsy/mortality unchanged** → Lead time bias (diagnosis moved earlier). If **more events detected in screened group** → Detection bias. If **hospital cases only** → Berksonian bias. If **follow-up incomplete** → Survival bias. ## NBE Trap NBE pairs "improved survival rate" with screening success to lure students into choosing detection bias or assuming screening worked. The autopsy data (unchanged mortality) is the discriminator that forces recognition of lead time bias—a concept many students conflate with detection bias or survival bias. ## Clinical Pearl In Indian breast cancer screening programs, many centers report improved 5-year survival after introducing mammography, but national mortality data (GLOBOCAN, ICMR) show minimal change. This is classic lead time bias—patients know their diagnosis earlier but die at the same age. True screening efficacy must be proven by reduced population mortality, not improved survival rates. _Reference: Park's Textbook of Preventive and Social Medicine, Ch. 10 (Epidemiology of Chronic Diseases & Screening); Harrison's Principles of Internal Medicine, Ch. 81 (Approach to Cancer Screening)_