AUSTIN (KXAN) — New research from scientists at the University of Texas at Austin suggests greater access to for-profit health data could improve public health, especially when it comes to predicting flu season.
In UT Austin research published in the journal PLOS Computational Biology, researchers found that out of the more than 600 flu-related data sets they studied, some of the best predictions came from electronic health records collected by athenahealth.
The company provides cloud-based services for healthcare companies. Researchers found that combining this private data with information collected by the Centers for Disease Control and Prevention would improve flu forecasts.
In fact, this combination of data was found to be 15 percent more accurate than using the CDC data alone. The researchers added that using the top five data sources they found — two from surveillance data from the CDC and three from athenahealth — were far more effective at predicting flu risks than other data sets.
The research was aided by the supercomputers at UT’s Texas Advanced Computing Center.
The findings were carried out by Professor Lauren Ancel Meyers’ lab with UT’s Department of Integrative Biology.
“The sort of enterprise of trying to predict outbreaks and, specifically, outbreaks of flu is a new thing, and it’s totally in its infancy,” Meyers explained.
“Flu is one of the most dangerous infectious diseases that we experience every year here in the United States,” she said, noting her lab chose to study flu because there is so much data available on it.
She said plenty of people have tried using new data to predict flu outbreaks before, from online searches of symptoms and social media posts to pharmacy records.
“What we’ve thought less about and what our study specifically focuses on, is what is the best data for the job?” she said.
Meyers explained outside of academic research, it can be difficult to get a hold of private data sets like these, because companies like to protect the information they own and avoid the privacy issues that can come with sharing health data.
“Our study suggests it might be worth trying to cross some of those hurdles because the data can be quite powerful,” Meyers said.
The researchers explained computer models used for predicting upcoming flu seasons are “not very accurate,” in part because it can be tough to decide what type of data to plug in.
But they believe making more data from health care companies accessible to researchers and public health agencies could help better predict when the flu season will peak, how long it will last, and how many people will get sick.
The researchers also believe this method of improving the data used could also be used to forecast viruses like dengue fever and chikungunya.
The problem is, companies don’t go around readily offering up health care data, Meyers explained. Supplying that extra data and removing identifying information may require more time and money for health care companies, she said.
“These are commercial organizations and we have to figure out how we also cover the costs of making this kind of information available in a protected way, which is no small feat,” Meyers said.
But her lab believes it will be worthwhile to find new ways to access this kind of information. To predict the flu virus better, she believes it will require partnerships between public agencies, private companies, researchersand technology companies to harness the best information.
“Hopefully this will inspire athenahealth to come up with a sustainable way of making their data available and that other companies like them will follow suit,” she said.
Athenahealth sent KXAN a statement for this story that read in part:
“At athenahealth, we support the research initiatives of academic researchers and public health agencies across the country by providing carefully de-identified and/or aggregated data sources around topics like influenza, opioid prescribing trends, obesity, and the impact of the Affordable Care Act. We agree that electronic health records offer a treasure trove of important health information and look forward to future collaborations.” – Josh Gray, Vice President, athenaResearch, athenahealth
Flu season this year
Dr. Ajay Zachariah, a family medicine physician at Seton Family of Doctors at Round Rock, explained the two most important ways to avoid the flu are to get the vaccine early (before November) and to practice good hygiene, covering your mouth and nose when you sneeze, washing your hands with soap and water, and using hand sanitizer.
He noted the people most likely to come down with the flu are children under the age of five and seniors older than 65.
Zachariah added that the most common type of influenza virus in other countries now is Flu A.
“It is difficult to predict at this early stage how effective this year’s vaccine will be compared to last year,” he said.
Dr. MaryAnn Tran, an infectious disease physician at Baylor Scott & White Medical Center in Round Rock, explained the components of this year’s flu vaccine have changed from last season, noting the influenza A and B components are different.
“It is not possible to predict what this flu season will be like and each flu season varies in severity,” she said.
Tran added that the CDC recommends everyone six months of age or older to get the flu vaccine.
“It is the best way to protect yourself from the flu,” she said, also noting the live vaccine (LAIV4 (FluMist Quadrivalent) can be given this season, which is a change from the previous two flu seasons.