Modernizing Data to Protect Mothers and Babies

The Surveillance for Emerging Threats to Mothers and Babies Network (SET-NET) was built as a system to collect information on moms and babies affected by Zika, but it has since adapted to include other exposures like COVID-19, cytomegalovirus, hepatitis C, mpox, and syphilis.

Even before the unexpected emergence of COVID-19, CDC’s SET-NET team had been working to expand the system’s capabilities by adding data on more diseases. When the pandemic hit, SET-NET was uniquely positioned to follow pregnant people with COVID-19 and understand its effects on babies.

SET-NET worked to report data to the public faster through preprint articles and web-based dashboards on birth and infant outcomes among pregnant people with COVID-19 in order to inform public health messaging and clinical guidance.

Tapping into Artificial Intelligence to track disease severity

SET-NET teamed up with Georgia Tech Research Institute (GTRI) to test out a new process for understanding the severity of COVID-19 in pregnant people. With thousands of cases being reported across the country, they needed a solution that could automate some of the work that was previously being done manually by scientists. By applying innovative natural language processing methods, CDC and GTRI were able to rapidly analyze structured data and unstructured text from thousands of patient health records and classify each case as asymptomatic, mild, moderate-to-severe, or critical.

Natural language processing is faster than classifying each record by hand, and initial tests showed that the method agreed with the clinician’s judgment 99.4% of the time. This method provides the ability to sift through many records quickly, and the results can help us better understand the increased risks of COVID-19 infection and the safety and effectiveness of COVID-19 vaccination during pregnancy.

Beyond COVID-19: Increasing connection for the future

Beyond the COVID-19 response, scientists are engaged in a leading-edge effort to identify opportunities to prevent and treat syphilis infections during pregnancy and understand the longer-term outcomes of infants born with and without congenital syphilis.

Through a pilot effort with two state health departments, CDC merged the National Notifiable Disease Surveillance System’s syphilis data on adults and infants to create a pregnant person-infant-linked dataset for SET-NET. This can help streamline reporting for health departments for surveillance. As we continue to face health threats, these efforts to improve data collection will help experts get information faster and better protect the nation.

The ongoing effort to expand SET-NET has massively advanced collaboration across infectious and non-infectious disease programs at CDC and is setting an example for how modernization can work across the agency. SET-NET is also one of two CDC programs (along with viral hepatitis) actively working to develop and pilot some of the new tools and processes being developed for the North Star Architecture in phase two of the Data Modernization Initiative.