In 2016, the World Health Organization identified 21 countries which could eliminate malaria by 2020. Monitoring progress towards this goal requires tracking ongoing transmission. In addition, understanding the spatio-temporal pattern underlying declines in malaria transmission can inform efforts to maintain elimination and prevent re-introduction. However, traditional methods to measure malaria transmission such as parasite prevalence or the entomological inoculation rate are not suitable in near-elimination settings due to cases being sparse and highly focal in space and time.

I will present approaches which apply Bayesian network analysis to individual level case data from two elimination settings, El Salvador and China. Using this approach, we estimate the likelihood of connectivity between observed cases and derive individual reproduction numbers as well as their variation through time and space. Individual reproduction numbers, R_c, describe the state of transmission at a point in time and differ from mean reproduction numbers, which are averages of the number of people infected by a typical case. We also jointly estimate the number of likely unobserved sources of infection. We use these estimates within spatio-temporal geostatistical models and carry out time series analysis to map how transmission varied over space and time and to provide estimates of the timeline to elimination and the risk of resurgence.