The impact of inaccurate assumptions about antibody test accuracy on the parametrisation and results of infectious disease models of epidemics