Multiscale model for forecasting Sabin 2 vaccine virus household and community transmission

Since the global withdrawal of Sabin 2 oral poliovirus vaccine (OPV) from routine immunization, the Global Polio Eradication Initiative (GPEI) has reported multiple circulating vaccine-derived poliovirus type 2 (cVDPV2) outbreaks. Here, we generated an agent-based, mechanistic model designed to assess OPV-related vaccine virus transmission risk in populations with heterogeneous immunity, demography, and social mixing patterns. To showcase the utility of our model, we present a simulation of mOPV2-related Sabin 2 transmission in rural Matlab, Bangladesh based on stool samples collected from infants and their household contacts during an mOPV2 clinical trial. Sabin 2 transmission following the mOPV2 clinical trial was replicated by specifying multiple, heterogeneous contact rates based on household and community membership. Once calibrated, the model generated Matlab-specific insights regarding poliovirus transmission following an accidental point importation or mass vaccination event. We also show that assuming homogeneous contact rates (mass action), as is common of poliovirus forecast models, does not accurately represent the clinical trial and risks overestimating forecasted poliovirus outbreak probability. Our study identifies household and community structure as an important source of transmission heterogeneity when assessing OPV-related transmission risk and provides a calibratable framework for expanding these analyses to other populations. Trial Registration: ClinicalTrials.gov This trial is registered with clinicaltrials.gov, NCT02477046. Author summary The oral poliovirus vaccines (OPV) have been one of the most effective tools for poliovirus eradication. These vaccines contain transmissible, live-attenuated viruses that promote mucosal immunity in the primary vaccine recipient and their nearby contacts. Although wild poliovirus is on the verge of eradication, circulating, vaccine-derived poliovirus (cVDPV) outbreaks are undermining eradication efforts and are an increasingly urgent public health issue. Models capable of accurately forecasting future outbreak risk and transmission will be crucial for informing strategies aimed at the eradication of both wild poliovirus and cVDPV. Here, we generated and calibrated a poliovirus transmission model that combines multigenerational household and community structure, intrahost infection, and immunity modeling to shedding data collected during a monovalent Sabin 2 OPV (mOPV2) clinical trial performed in Matlab, Bangladesh. This model provided a unique opportunity to determine whether the additional complexities imposed by household and community structure were necessary for accurate transmission modeling. Our study shows that household and community structure plays an important role in limiting poliovirus transmission and that its inclusion in transmission model is necessary for accurate outbreak risk prediction.