Early detection and prediction of disease outbreaks are crucial for public health service delivery, containment response, saving patient lives, and reducing costs. We propose a new data-driven statistical methodology for outbreak detection and prediction based on routinely collected hospital Emergency Department data. The time between consecutive ED presentations matching a diagnosis of interest forms the basis of a novel index measure to signal that an outbreak has occurred. We validate the method using historical presentations of influenza-like illness made to a large sample of public hospital EDs in 2020 and compare outbreaks identified by the method with the start of the first wave of COVID-19. The method shows promise within the field of disease outbreak detection.