|
Abstract:
|
The scale and rapid spread of the COVID-19 pandemic has created an unprecedented public health emergency. In the absence of effective treatments and vaccines, disease surveillance and public health interventions are vital for mitigating the toll of COVID-19. Due to limited testing capacity, the number of cases was under-reported. In this paper, we propose a hierarchical epidemic model to estimate COVID-19 prevalence and transmission rates by explicitly accounting for under-ascertainment. In addition, we propose a causal inference method using marginal structural equations to estimate the effects of non-pharmaceutical interventions (NPIs), such as mask wearing and stay-at-home order, on transmission rates. Our statistical framework uses aggregated US county-level datasets on COVID-19 case rates, social distancing behaviour, and COVID-19 NPIs from multiple sources. We model the mean of the recorded case rates as a function of effective reproduction numbers and ascertainment rates, both of which can vary across time and region. The reproduction numbers are then modelled as a function of demographic covariates, time and time-varying behavioural and intervention covariates, while the asc
|
