Background: Objective was to quantify biases in screening for gastric cancer when comparing attenders to nonattenders using serum pepsinogen I (SPGI) level as primary test. Methods: In mid 1990s, all men aged 51–65 years ...