Teams of researchers from 12 health systems, 202 hospitals, use novel data collection methods to collaborate on research
UC Davis Health is collaborating with many other health systems in a data consortium aimed at answering specific questions about COVID-19. Researchers, clinicians and the general public are encouraged to submit their questions to COVID19questions.org, where teams of researchers evaluate their questions and collect data to make informed responses.
The goal of the program is for researchers to have access to large amounts of data for their studies, beyond what could be obtained from one hospital or health system. The answers to these questions are then shared online so that the general public can stay informed as an understanding of COVID-19 evolves.
Katherine Kim, a UC Davis associate professor at the Betty Irene Moore School of Nursing and the School of Medicine, Department of Public Health, is lead investigator for the team evaluating data from the UC Davis Health system—one of the 12 health systems, or sites, participating in the study. Each team is made up of professionals from a diverse array of backgrounds, according to Kim.
“We have data analysts and programmers at each site that actually run the software and pull the data to answer the question—run the statistics,” Kim said. “Then we have several other members who may have clinical backgrounds—our doctors and nurses—or have some public health expertise, so that when we answer these questions we have the experts also verifying what we’ve done to make sure that it’s all right.”
One unique aspect of this data consortium, called Reliable Response Data Discovery (R2D2), is that it allows hospitals to maintain control and security over their data, while still sharing it with outside researchers. Kim explained that the standard method for most studies that use widespread sources is to compile the data centrally.
“The real difference is most studies have a centralized database of all this data,” Kim said. “They either asked hospitals to give it to them or they’re using data from public health sources like the CDC and they’re running analysis on data that they have put together all in one database.”
When data is collected and kept in a central location, there is an increased risk of it being compromised or hacked. Many hospitals are wary of this potential breach of privacy which can hinder studies requiring large amounts of data. To tackle this challenge, researchers developed another method for receiving this information which relies on every hospital utilizing the same data model, according to Kim.
“Every hospital data stays at their hospital, and what we do is we send out this software packet that has what are the requirements, what kind of data do we want to include, what’s the actual analysis we need to do, how do we need to structure it,” Kim said. “That piece of software gets sent out to every hospital and they can run it against their data and return the results.”
By running the software at each individual hospital, researchers are able to compare COVID-19 cases with health data from those admitted for other illnesses, allowing for a wider breadth of analysis.
Some of the questions on the site include several patient variables; for example, one question asks “For patients with COVID-19 related hospitalizations, what is the average length of stay for those hospitalizations (in days) stratified by race, ethnic group, age group and gender?” The R2D2 network includes data from “not only academic health systems in wealthy metropolitan areas, but also small community hospitals, safety net hospitals and the VA,” which helps diversify the patient population, according to a press release.
Misty Humphries, a vascular surgeon and associate professor of surgery at the School of Medicine, Department of Surgery and the director of Research Experience in Surgery, was able to use data from R2D2 to study why COVID-19 patients were experiencing arterial thrombosis, or clotted arteries, and were unresponsive to standard treatments like blood thinners. She mentioned that R2D2 reduces exposure risk for those collecting COVID-19 data.
“Sometimes […] drawing blood from these patients or being around these patients is dangerous because you subject the person that is taking care of them to the virus,” Humpries said. “People that come in to draw the blood are also subjected to the risk of obtaining the virus if they’re not appropriately gowned and everything, so we try and minimize interactions.”
She added that “the electronic medical record is a tool that has not been fully capitalized on in health care,” but that standardized and accessible patient records for clinicians and researchers is a step in the right direction to provide comprehensive and better care for the general public.
Written by: Madeleine Payne — science@theaggie.org