Sub-national estimation of surveillance sensitivity to inform declaration of disease elimination

A fundamental question in the global commitment to polio eradication is how long a period of absence would be consistent with regional elimination, and the safe withdrawal of the oral polio vaccine is contingent on the answer. We present a statistical framework to model the time-varying sensitivity of two key components of polio surveillance - environmental sampling and clinical cases of acute flaccid paralysis - for detecting infection on a monthly basis at the local government authority level. We use this to estimate the probability of freedom from infection (FFI) at a critical prevalence level that is consistent with interruption of transmission, given the absence of virus in collected samples. We validated this framework against two periods of poliovirus absence in Nigeria (2014-2016 and 2016-2020). Our model highlights substantial heterogeneity in surveillance sensitivity over time and space and, given this, concluded an 85% probability (95% uncertainty interval: 77.1-90.2%) of the country being free from WPV1 infection after 23 months without detection from July 2014. Detection of WPV1 in July 2016 demonstrated that circulation had indeed persisted during this time. In contrast, we conclude a probability of 98% (97.9-98.7%) by the time elimination of the serotype was officially declared in 2020. The inferred probability of FFI during both time periods was found to be consistent with the retrospectively known status of regional elimination. This supports the validity of applying this framework prospectively to inform the certification of wild poliovirus elimination from remaining endemic regions, and to determine the resolution of cVDPV2 outbreaks.