Tracking viral warnings Resilience indicators to anticipate mosquito-borne disease outbreaks

Summary

Since interventions to control the spread of mosquito-borne disease outbreaks are limited, anticipating when and where outbreaks are about to occur could reduce the number of cases and allow public health services to prepare and respond more effectively. In this thesis, we focused on resilience indicators as a generic, model-free anticipation method. We found that resilience indicators can accurately predict outbreaks, but their practical use is often limited by the need for large amounts of detailed data. To address this, we tested alternative data collection strategies, like collecting shorter, high-resolution time series (« bursts ») or combining multiple data sources, and found these strategies gave reliable results. We also evaluated machine learning methods as early-warning tools, though resilience indicators generally performed better. Together, these findings support the development of early-warning systems that could improve outbreak preparedness and help reduce the public health consequences of mosquito-borne diseases.