Predicting patterns
Photos by Ashley Barnas Larrimore | Photo illustration by Jeffrey C. Chase December 21, 2023
Epidemiology researcher uses Google Trend search data to predict RSV pediatric hospitalizations
You do it all the time. Wake up with cold-like symptoms and start Googling. Google logs all search activity, allowing researchers to dig deep into data, glean insights into population behavior, and identify health-related trends.
Tarang Parekh, assistant professor of epidemiology at the University of Delaware College of Health Sciences, has done precisely that. Using historic modeling of Google Trends search activity from 2019 and a cross-sectional analysis of the 2019 Healthcare Cost and Utilization Project’s Kids’ Inpatient Database (HCUP-KID), Parekh and colleagues have identified a link between respiratory syncytial virus (RSV) searches and pediatric hospitalizations linked to RSV.
The study, recently published in the American Academy of Pediatrics’ Hospital Pediatrics journal, is the first of its kind, showing lead-time predictions that could have significant ramifications on overburdened healthcare systems.
“If people’s interest in RSV increases, and they’re searching for information on RSV, should we expect increased hospital admissions within a month or two? That’s the question we were asking,” Parekh said.
The research was inspired by Parekh’s wife, Sahiti Pemmasani, a co-author of the paper and a pediatric emergency fellow at Beaumont Hospital in Michigan. She observed an influx of RSV cases during her residency at Mt. Sinai Hospital in New York in 2020.
“She saw a significant increase in RSV cases, and ICUs were full of COVID patients,” he said. “So, in talking with pediatricians in West Virginia and California, we realized they were seeing similar increases in hospitalizations, and it made me wonder about potential patterns.
“Hospitals were not ready for the RSV spike,” Parekh said. “But we thought if we can predict an increase in RSV hospitalizations based on search data, hospitals could prepare and ensure they’re better equipped to handle an influx of pediatric hospitalizations.”
This could save lives. RSV kills between 100 and 300 children younger than 5 years old annually. Of more than 2 million cases reported annually, 58,000 resulted in hospitalizations. Pre-pandemic, RSV season often began in October and peaked in February, but that can vary by region.
“Post-pandemic, with COVID still circulating, there’s a lot of variation in the seasonality of RSV,” Parekh said. “We don’t know when it’s going to peak, and most of the time, the hospitals are not ready to tackle increased RSV hospitalizations.”
Parekh is hoping his research can change that. He found that for every one-unit increase in RSV searches over two consecutive months, there was a corresponding increase of 140 RSV-related hospitalizations. In the Midwest, an RSV surge was more apparent. Researchers identified a corresponding increase of 250 admissions for every one-unit increase in RSV searches.
“If we can predict behavior patterns and corresponding hospital admissions, then we can say with certainty we know the pattern and can expect the pattern,” Parekh said. “Then, physicians can confidently say they’re ready to tackle it.”
While the first RSV vaccine was approved in May 2023, there’s already a shortage due to high demand and limited supply. This further puts physicians in difficult positions where they must prioritize the most vulnerable.
“We know when flu strikes each year and are prepared with the vaccine, but with RSV, if we must change the recommendation, how will we ensure there’s enough vaccine for the populace?” Parekh asked. “It shows that while we have a solution, we’re still unprepared.
“If we can anticipate trends in higher RSV-related hospitalizations in the future, hospitals can be better prepared, and public health departments can advise people not to panic,” he said.
In the future, Parekh hopes to analyze the data from the 2022 HCUP-KID database to see whether a similar pattern exists and launch a pilot study to apply this modeling in a healthcare setting to determine whether an RSV surge can be predicted in a particular region.
Contact Us
Have a UDaily story idea?
Contact us at ocm@udel.edu
Members of the press
Contact us at 302-831-NEWS or visit the Media Relations website
ADVERTISEMENT
Resources
Office of Communications & Marketing 105 E. Main St. Newark, DE 19716 ocm@udel.edu Phone: 302-831-2792