Introduction: Uninterrupted access to medical communications during disasters is necessary to optimize disaster preparedness, response, and patient outcomes. Robust, data-driven approaches to understand telecommunication network performance during and after disasters are necessary. The objective of this pilot study was to identify patterns in network disruption during natural disasters to enable characterization of performance and inform standards. Methods: Network performance data from multiple mobile and fixed cellular network carriers, provided by Ookla For Good, was analyzed to characterize performance during three major disasters: Hurricane Michael, Panama City, Florida (October 10, 2018), Hurricane Ian, Fort Myers, Florida (September 28, 2022) and Puerto Rico earthquakes (December 28, 2019 – January 11, 2020). Time series analysis and geospatial analysis were conducted on the available network performance data for these regions before, during, and after the disasters to identify patterns of change in network latency and download and upload speeds. Results: Despite substantial gaps in data (missing, uneven, and time-restricted datasets), distinct patterns of impact and recovery of network performance were identified. Panama City experienced the most severe disruptions, with significant declines in download and upload speeds and prolonged latency across both mobile and fixed-line networks. Fort Myers showed a similar trend, with marked drops in network performance following Hurricane Ian, although most providers demonstrated a strong recovery over time. In contrast, Puerto Rico was less affected by the earthquake, exhibiting minimal volatility in network performance. Conclusion: This analysis identified varying degrees of network resilience across different locations, carriers, and disasters. Additional data could permit the development of an intelligent model that is capable of predicting network performance during disasters. Understanding how hurricanes and earthquakes impact telecommunications infrastructure can inform stakeholders about expected network conditions, guide the development of resilient systems, and be used to set minimum standards that ensure continuity of medical communications during natural disasters.
Boyle et al. (Sun,) studied this question.