A framework for comparing early warning systems across domains: A step toward a data-integrated public health EWS

Early Warning Systems (EWS) are crucial tools for public health, providing time for agencies to devise and enact control and mitigation measures in the face of emerging health threats. EWS offer similar benefits to agencies that manage other domains such as natural disasters or financial markets. After surveying various EWS, we develop a novel framework for characterizing EWS across domains. Key to this framework is the characterization of an Early Warning System´s domain and focal event; whether its aim is prediction, detection, or warning, whether its focus is model, systems, or infrastructure; the extent of human intervention required; and its input data. We believe this framework is quite novel, but more importantly, it serves as a reference to chart future projects. We use it to verify that an opportunity exists in developing public health EWS that integrate a spectrum of inputs from Web 2.0 and social media data, data from sensors, data from, say, electronic health records, as well as human opinions and behaviors.