November 29, 2022 | Erin Bluvas, bluvase@sc.edu
Scientists from the USC Arnold School of Public Health and School of Medicine Columbia have developed an algorithm for identifying the gestational week when COVID-19 infection occurs. Published in PLOS ONE, the method leverages large amounts of data from electronic health records and enables researchers and clinicians to better understand the potential risks to both fetus and mother.
“Recent findings suggest COVID-19 to be associated with an increased risk for adverse pregnancy outcomes and neonatal complications; however, there has been limited knowledge pertaining to the timing of SARS-CoV-2 infection during the pregnancy,” says Chen Liang, who is an assistant professor with the Arnold School’s Department of Health Services Policy and Management (HSPM) and the USC Big Data Health Science Center. “Timing of viral infection is important because fetuses are more vulnerable to maternal complications and/or viral infection during certain gestational stages.”
Unfortunately, the quality of data (e.g., incomplete and inconsistent health records) and existing extraction methodologies fall short of identifying when these key moments occur. With their new algorithm, Liang and his team use available health records to accurately infer gestational age and delivery dates – enabling researchers to study the impact of COVID-19 on pregnancy.
For this project, the authors used electronic health records collected during antenatal, labor/delivery and postpartum care. Spanning up to 21 months, these records provide a rich source of data from multiple visits, health providers, clinical sites and information systems. The availability of these records in an electronic format is a relatively new development in health care research and has paved the way for big data approaches.
The emergence of big data methods has allowed scientists to make rapid progress in developing useful tools, such as this algorithm, for extracting useful information from previously underutilized, and often untapped, reservoirs of information contained within electronic health records. The authors’ creation is already in use by several cohort studies examining the impact of COVID-19 on pregnant women’s clinical inflammation and pregnancy complications in real time.
HSPM doctoral student Tianchu Lyu, who served as first author on the paper, presented the findings at the American Medical Informatics Association’s annual symposium earlier this month. By spreading the word about this algorithm, the Norman J. Arnold Doctoral Fellow and SmartState Center for Healthcare Quality Junior Scholar is extending the positive impacts of big data even further.
“Our algorithm is among the first that detects the exact gestational week of viral infection, early-stage pregnancy, preterm birth, early termination, post-term birth and other adverse clinical events,” Liang says. “Because viral infection at different stages of pregnancy is associated with different risks of fetus development and maternal status, it is also generalizable to viral infection aside from COVID-19 as well as adverse events that would affect pregnancy.”
Related:
Chen Liang named Fellow of the American Medical Informatics Association
Arnold School researchers enlisted in nationwide effort to tackle long COVID
Chen Liang brings informatics expertise to Department of Health Services Policy and Management