How Epic’s Cosmos Supported Clinical Research with De-Identified Data

Despite not having funding as a medical student, Jane Hinkle of Case Western Reserve University recently conducted award-winning clinical research in less than one week using EHR vendor Epic’s de-identified patient database, Cosmos.

Cosmos is a HIPAA limited data set combining the Epic EHR data of over 122 million patients.

“The concept that we have built around Cosmos is what we call self-service analytics, meaning that the big data in Cosmos is set up to be accessible to any user of Cosmos,” Dave Little, MD, a clinical informaticist at Epic, told EHRIntelligence in an interview.

“In the case of Jane, we saw a medical student who was able to go into this data with a targeted question and find the data that she needed to respond to that question based off of a population of 122 million patients,” he said. “That is historically unheard of in research circles.”

Hinkle’s study found that pediatric patients are three to five times more likely to get myocarditis from COVID-19 itself than from the COVID-19 vaccine.

Hinkle told EHRIntelligence that she was originally focusing on a study that used a big data tool to analyze obstructive sleep apnea. However, when the question about whether myocarditis was an adverse effect of COVID-19 vaccinations came about, Hinkle and her research mentor David Kaelber, MD, PhD, MPH, chief medical informatics officer at The MetroHealth System, shifted their focus.

“It was really exciting to work on something that was very pressing in the community right now,” Hinkle told EHRIntelligence in an interview.

If Hinkle had conducted this study 10 years ago when EHRs were new, she would have had to design very specific queries for use only at Case Western, Little noted.

“She maybe would have come up with a few hundred patients or so, or even less,” he said. “If you had gone back another 10 years, when most folks didn't have any electronic health records at all, and this was having to be done on paper, it would take months or years to compile enough cases of myocarditis after vaccination to make any sort of meaningful conclusions from that data.”

Hinkle said that she found the tool to be adaptable and easy to use.

“With this style tool in particular, there's so many different ways to ask the question that you're looking to ask,” she said.

Additionally, with so many health systems nationwide using Epic EHR systems, Hinkle said that she felt confident that she was getting a diverse array of health systems to represent how this problem is affecting the country.

Hinkle explained that the research team started with a simple inclusion criteria to target their study on the pediatric population, which brought the number of patient records from 122 million to 16 million. Then the researchers analyzed COVID-19 vaccination status in relation to myocarditis. 

Little said that Epic has been compiling and curating Cosmos data for a number of years, but it has just been within the past few months that Cosmos has been open for inquiries from physician users like Hinkle.

“I am really excited to see what our organizations will do with that data and what findings that they will generate,” he emphasized.

Little said that 38 percent of Epic customers currently contribute EHR data to Cosmos, however over 50 percent of customers are signed up to contribute data.

“From the time they sign up, it takes a while for them to start contributing their data and get access to data on the other end,” he explained.

Little said that as time goes on, he thinks additional organizations within Epic will become interested in participating in Cosmos to draw from the de-identified data that is available.

“The real power of Cosmos is in these big numbers of 122,000,000 patients that are currently available,” he said.

Hinkle said that she is excited about the implications big data tools like Cosmos have for clinical research.

“This is definitely going to be a future field of clinical research,” she said. “It's a very exciting thing to be a part of.”

Hinkle said that conducting this research cemented her decision to take an extra year at Case Western to earn her MPH.

“Working with all of these physician epidemiologists, I realized how much more training I would really need to be able to conduct this type of research professionally,” she said. “This experience really cemented that interest for me.”