Data from the ERASE-TB study, used to develop prediction models for Mtb infection among adolescent and adult tuberculosis household contacts and evaluate the risk of Mtb infection by HIV status. This is a cross-sectional dataset, with one row per person, indicating the participant interferon gamma release assay status, demographics and clinical history, and data on the index person with tuberculosis in the household (index case) and household-level variables. All candidate predictors for prediction model development are included.