Sarcouncil Journal of Medical Sciences

Abstract: Infectious disease surveillance in the United States has long been facing challenges of delayed feedback, inefficient data infrastructure, and limited predictive capacities, which have been evident in recent outbreaks. However, the burgeoning development of big data ecosystems coupled with machine learning algorithms has made it possible to transform the situation by improving the timeliness, accuracy, and robustness of infectious disease control in the United States. This review draws upon the current literature for a synthesis of the integration of big data and machine learning in infectious disease surveillance and control in the United States. Conventional sources of public health-related data are presented, in addition to new sources of digital, genomic, and non-conventional data sources such as electronic health records, syndromic surveillance, mobility datasets, social media data, wearable biosensing, and genomic pathogen sequencing. Finally, different machine learning paradigms such as supervised learning, unsupervised learning, and deep learning are presented in terms of their applicability for detection, forecasting, and risk assessment. Practical applications of machine learning for early warning of outbreaks, disease control, resource allocation, and precision medicine for public health are presented with an emphasis on the United States. Finally, future directions for research in machine learning-related applications in disease control are presented. This review cumulatively signifies the promise of machine learning-enabled disease control for improving the accuracy, speed, and robustness of infectious disease control in the United States.