When the official figure and the WHO/UNICEF estimate disagree: a pre-specified regional analysis of reported versus WUENIC vaccine coverage, 2000-2025

Official national immunization coverage informs funding and target tracking, while the WHO/UNICEF Estimates of National Immunization Coverage (WUENIC) reconcile them with surveys and other evidence. This paper measures the gap between official and WUENIC coverage for DTP3 and MCV1 in 195 countries and areas, 2000-2025, by WHO region (2025 WUENIC revision). The analysis follows an unpublished plan recorded before the analysis; its timing cannot be verified. One registered stop condition was not followed: the plan said to stop on any ``header change'', every request sent a fixed custom User-Agent, and the survey host refuses the library default header, so the survey-dependent checks are reported outside the registered evidence and used for no conclusion. In the median country-year official and WUENIC figures are identical in every region and almost every year (40.3% of DTP3 country-years differ by zero); the gap lives in a tail: 10.5% to 13.0% of country-years report DTP3 coverage at least 10 percentage points (pp) above WUENIC, and one fifth to one third in the African Region (exploratory). The registered test found a mean within-country increase in the DTP3 gap of 0.56 pp from 2017-2019 to 2020-2022 (95% bootstrap interval 0.07 to 1.08; sign-flip p = 0.03, which assumes symmetric changes, doubtful given their skew; an exact sign test gives p = 0.12), below the registered 1 pp threshold of practical size. Its ``widening'' and ``no widening of practical size'' clauses both hold and no priority was registered, so the registered classification is indeterminate. The same contrast three years earlier was -0.04 pp (-0.75 to 0.66); a births-weighted version was not distinguishable from zero. Only the African Region's interval was above zero (1.68 pp, 0.12 to 3.41). A denominator decomposition against UN population estimates was too imprecise to apportion the administrative gap.