AI Revolutions
Florence Nightingale’s rose diagram, which transformed raw hospital mortality data into a visual representation, revolutionized disease surveillance. Since then, AI and machine-learning algorithms have been developed to complement manual techniques, allowing for the identification of hidden patterns and innovative solutions to old problems.
Early-warning systems for disease surveillance have benefited immensely from AI algorithms, which can parse, filter, classify, and aggregate text for signals of infectious-disease events at unprecedented speeds. HealthMap, an internet-based infectious-disease surveillance system, is one example of such a system. It uses natural-language processing to search through text posted across the web, providing worldwide coverage and hyperlocal situational awareness.
AI has also led to advances in diagnostic classification, enabling public health authorities to differentiate among various pathogens and respond accordingly. For example, a recent AI-driven image classification tool has been created to ascertain the antibiotic susceptibility of bacteria with a highly scalable approach, enhancing our ability to track antibiotic resistance globally.
When an outbreak has been identified, the next step is to stop the outbreak by first tracing and then cutting off routes of transmission. AI can help identify sources of infection by combining whole-genome surveillance sequencing and machine learning to automatically mine patients’ electronic medical records (EMRs) for data related to an outbreak. The Enhanced Detection System for Healthcare-Associated Transmission (EDS-HAT) is one such system that has identified multiple, otherwise-undetected outbreaks by detecting hidden transmission patterns in EMR data.
While AI has shown promise in infectious-disease surveillance, there are also critical limitations to consider, such as the lack of standardization in data collection and the potential for biased algorithms. Addressing these limitations is essential for improving implementation in the future.
COVID-19 Pandemic
AI and machine learning techniques have been increasingly used in infectious disease surveillance, especially during the Covid-19 pandemic. These methods have helped identify outbreak hotspots, monitor adherence to nonpharmaceutical interventions, and optimize resource allocation for testing. One example is Eva, an AI algorithm used in Greece to screen travelers for Covid-19 at the border of the country. The algorithm uses reinforcement learning to target travelers for polymerase-chain-reaction Covid-19 testing, resulting in identifying substantially more cases than those identified with the use of alternative strategies. Wearable devices such as smartwatches and smart rings can detect early signals of an impending outbreak. Furthermore, AI and machine learning can help transform individual behavior into population health information, generating meaningful insights from otherwise difficult-to-interpret, multidimensional data.
Roadblocks and Limitations
However, there are also roadblocks and future directions to consider in the use of AI and machine learning in infectious disease surveillance. These include the availability and quality of data, representation of selected populations in databases, issues of individual privacy, and the limits of AI. Although AI can improve surveillance infrastructure, it cannot replace the cross-jurisdictional and cross-functional coordination required for the collective intelligence necessary to fight novel and emerging diseases. Collaborative surveillance networks are needed for ongoing endemic surveillance if we are to be prepared for the next pandemic. The future of infectious-disease surveillance will feature emerging forms of technology, including but not limited to biosensors, quantum computing, and augmented intelligence. Nevertheless, the success of the next generation of AI-driven surveillance tools will depend heavily on our ability to unravel the shortcomings of our algorithms, recognize which of our achievements are generalizable, and incorporate the many lessons learned into our future behavior.
Brownstein, J.S., Rader, B., Astley, C.M., & Tian, H. (2023). Advances in Artificial Intelligence for Infectious-Disease Surveillance. New England Journal of Medicine, 388(17), 1597-1607. doi: 10.1056/NEJMra2119215. PMID: 37099342.