Canadian measles vaccine coverage data are collected cross-sectionally and are missing for many province, dose and year combinations, against roughly 1,100 reported cases between January and June 2026 compared with a previous annual average below 200. The authors curated a standardised coverage dataset from public sources and used Gaussian process models to impute first-dose coverage by province, territory and age, validating the approach against England's more complete data.
Why it is interesting: Shows how susceptibility inputs for outbreak models are constructed when coverage registries are incomplete, and how much of the immunity landscape is imputed rather than observed.