Researchers envision better disease surveillance to improve public health

"Advances in identifying and gathering data, techniques for combining and processing that data, and technologies that enable deeper data analysis will facilitate a new paradigm in outbreak definition and detection. Our goal is to develop data analysis tools that can accurately determine both the probability that an outbreak will occur and its predicted severity," said co-author Elena Naumova, PhD, professor of public health and community medicine at Tufts University School of Medicine (TUSM).

"There are many obstacles to improving tools for disease surveillance. Varying definitions of what constitutes an outbreak, differences in the quality and source of data, and differences in how quickly diseases progress and spread -- all of these are realities that make it challenging to accurately and efficiently predict an outbreak," said Naumova.

In the current study, Naumova and co-author Nina Fefferman, PhD, describe cutting-edge methods that address these obstacles and have the potential to revolutionize the way disease data is processed and analyzed. Fefferman is an assistant professor in the department of ecology, evolution, and natural resources at Rutgers University as well as a research assistant professor in the department of public health and community medicine at TUSM.

"There is vast, untapped potential in the multitude of new data streams made available by technological progress. Data can be submitted instantly from the site of an outbreak investigation, it can be gleaned from websites and news items, it can be obtained from satellite images; our current methods, however, are not up to the challenge of tapping and appropriately analyzing these sources. New techniques are being developed that will dramatically shorten the time it will take us to detect an outbreak," said Fefferman.

"Dynamic mapping, multivariate visualization, flow mapping, outbreak signature forecasting, and large-scale simulations of infection spread are just a few of the tools being developed that might help us better detect the complexities of disease spread. These data analysis tools present huge data sets in such a way that researchers can better detect and interpret trends and anomalies," said Naumova.

"Additionally, we can gain further insight into an outbreak if auxiliary information related to the population is readily available. By including information about demographics, population migration and travel patterns, vaccination patterns, and large social gatherings, we can factor in the many contributors to an outbreak," said Fefferman.

According to the researchers, taking this auxiliary information into account will allow public health researchers to adjust their analyses by population, perhaps examining areas with limited vaccination coverage, or higher proportions of older adults, or heavy traffic from airports and tourism.

"Our hope is that improved disease surveillance tools will help us detect the unique signature of each outbreak, and as a result improve our ability to recognize the warning signs early and with greater accuracy," said Naumova.

Naumova is the director and Fefferman is the co-director of the Tufts University Initiative for the Forecasting and Modeling of Infectious Diseases (Tufts InForMID), which works to improve biomedical research by developing computational tools to assist life science researchers, public health professionals, and policy makers. The center is focused on developing methodology for analysis of large databases to enhance disease surveillance, exposure assessment, and studies of aging.

This study was funded by the US National Institute of Allergy and Infectious Diseases, part of the National Institutes of Health; as well as the Center for Discrete Mathematics and Theoretical Computer Science, and the Department of Homeland Security Command, Control and Interoperability Center for Advanced Data Analysis, both at Rutgers University.