For years, researchers have struggled to access public data that could better inform health policy and address health equity. Why is this data so hard to obtain and how can we make it more accessible?
This was the main question addressed last week in a webinar hosted by Mathematica, where Data Innovations Director Ngan MacDonald moderated a conversation with Senior Director of Health Equity Deliya Wesley and Amy O’Hara, director of the Georgetown University Federal Statistical Research Data Center.
The problem
MacDonald, Wesley, and O’Hara first discussed barriers to access for both administrative and commercial data. Administrative data (data collected from local, state, and Tribal governments) is challenging to access for legal and technical reasons. O’Hara cited the need to improve access “in a way that is consistent with the laws and regulations but isn't just restricted to like a bunch of people that know the secret handshake.”
MacDonald noted that the financial burden of commercial data is prohibitive for researchers. O’Hara added that there are additional challenges and issues with data quality, something she saw firsthand while at the Census Bureau. Wesley said that increasing access to data would help advance health equity, as it could lead to more robust analyses for indicators of health and well-being and provide a greater opportunity to achieve optimal health outcomes for more people.
Collaborating toward a solution
O’Hara stated that the first step in providing more access to the data is to increase awareness about the country’s Federal Research Data Centers. These data centers offer remote access to administrative data for projects approved via an application process.
While raising awareness about the centers is critical, researchers and policymakers may also need help navigating the required paperwork and fully understanding how to use the data. Georgetown, Mathematica, Northwestern, and the Robert Wood Johnson Foundation (RWJF) have teamed up on the Data for Access Project to help train researchers inside the labs across the country so they can get a quicker start with data sets.
“Ideally, you get your application approved and you get in the lab, and then you have access to somebody that knows how to code and will help you accelerate your progress to (finding answers) for the questions that you’ve articulated already,” said O’Hara.
Regional collaborations also offer researchers access to large amounts of data. For example, MacDonald pointed to the Chicago Area Patient-Centered Outcomes Research Network (CAPriCORN), which enables researchers to securely access electronic health record data in the greater Chicago area that wouldn't otherwise be possible-while also protecting patient privacy.
Centering community voices in health data is also needed to ensure a robust analytic approach. Wesley noted that Mathematica has an integrated in-house lived experience expert panel and Mathematica researchers have heard a lot in the last few years about the power of lived experience as a complement to research expertise.
Sustaining data access for the next generation
After offering a solution to challenges in data access, the panel turned to how to achieve a sustainable system that will prevent future researchers from facing the same barriers. The Data for Access Project aims to help drive that sustainability.
“Current research data centers are mostly centered on the coasts or at elite universities," said O’Hara. "And that’s ridiculous because that means that you’re making it harder for people that have the different lived experience…to even be able to propose to use the information.”
The hope is that offering access to researchers who haven’t traditionally worked with the data centers will result in the posing of different questions and in answers that resonate with groups that have not seen themselves reflected in data outputs.
A full recording of the webinar is above. Learn more about Mathematica’s Data Innovation Lab, the Federal Statistical Research Data Center at Georgetown University, and the Data for Access Project.
