For each test and study, authors can define one or more covariates for example setting, age, severity of disease, or specific features of the test such as threshold. These covariates can be used for identifying subgroups that may have different test accuracy results.