LAUSR

Correspondence to Dr Ran Xu; ran. 2. xu@ uconn. edu © Author(s) (or their employer(s)) 2022. Reuse permitted under CC BYNC. No commercial reuse. See rights and permissions. Published by BMJ. Human actions are at the heart of epidemics. From government lockdown decisions to individual mask wearing, social distancing and vaccination choices, evolving actions of individuals and policy makers have conditioned observed patterns of COVID19 pandemic. In fact, scholars have increasingly called for integrating the interplay between human actions and dynamics of infectious diseases and offered theoretical models in support of those calls. 3 A common approach has been to use data (eg, on enacted policies or mobility) to quantify the impact of individual and policy choices on disease transmission, which enhances models’ fit to historical data. Less explored is closing the feedback loops connecting human behaviour and outbreaks. Such feedback loops follow the recognition that changes in human behaviour (individual and policy) are driven by (among others) the ongoing cases and deaths and the perceived risks. Consider one such behavioural feedback loop (which we call ‘riskdriven response’): During a pandemic, as infections and deaths rise, people and governments perceive a higher level of risk, which enhances the adoption of transmissionreducing Nonpharmaceutical interventions (NPIs). Then, as the cases and, consequently, the perceived risks decline, people relax the adherence to NPIs, leading to higher interactions and more cases, and the cycles continue. While intuitive and theoretically recognised such behavioural feedback mechanisms, with some exceptions , have not permeated the relevant research and the resulting policy insights. For example, examining over 60 models providing predictions to the Center for Disease Control’s (CDC) COVID19 forecast hub, we found only one model accounted for the riskdriven response endogenously. Here, we draw on recent studies to argue that this missing feedback is central to resolving some important empirical puzzles, providing novel policy insights and offering better forecasts.