Viorika/istock via Getty Images

Addressing the Social Determinants of Health with AI, Partnerships

UVA Health System is leveraging artificial intelligence and community partnerships to address individuals’ social determinants of health.

In the healthcare industry today, it is widely understood that optimal health outcomes require addressing patients’ clinical and non-clinical needs – their social determinants of health.

So much of individuals’ health is determined by factors beyond the doctor’s office. Where someone lives, works, and plays has a direct impact on her well-being, and it’s critical for health systems to gather and understand their patients’ social determinants data.

However, for many healthcare organizations, it can be challenging to know where to start with addressing patients’ social determinants. A lack of industry standards makes it difficult to collect and share this essential information, and healthcare entities may not be equipped to address unmet social needs.

These obstacles have been highlighted even more as the industry has increasingly understood the important role social determinants play in overall health, Amy Salerno, MD, MHS, Director of Community Health and Well-Being at the University of Virginia (UVA) Health System, told HealthITAnalytics.

Amy Salerno, MD

“Industry-wide and society-wide, there’s a lot more recognition of the social determinants of health. In Charlottesville specifically, we’ve been looking at inequities and disparities in our local community and among our patients, and noting the tie to health outcomes,” she said.  

“Because the health system is not necessarily the expert in addressing housing instability, food insecurity, or other socioeconomic factors, we need to partner with organizations in our community that are the experts and create a robust network to support individuals’ social needs.”

In 2018, UVA Health System started to partner with community organizations to help patients beyond clinical settings. The organization selected a technology-based referral network, called Pieces, to better connect patients with community resources that meet their specific needs.

Before UVA could achieve better community health outcomes, the entity needed to develop a comprehensive network to connect patients with community groups. 

“We had to consider what a robust network and partnership should look like, where we respect and honor the expertise that our community partners bring to the table. We wanted to be able to address these complex issues using shared decision-making,” Salerno said.

“That led us to look for solutions that align with the strategic objectives of UVA from an operational standpoint, but that also serve the broader community and not just our patients. We wanted it to be a win-win, both for our community and for UVA.”

UVA implemented a real-time artificial intelligence platform that integrated directly with hospital EHRs. The platform leverages natural language processing tools to extract social determinants and clinical risk factors from unstructured notes for more timely interventions.

For UVA, the technology could help hospitals improve patient care in several key metrics, Salerno said.

“The areas where we have the chance to make the biggest impact is around hospital length of stay and readmissions,” she said.

“For us, it’s been huge just to have a starting point of really robust data to tell us exactly where we are compared to other National Hospitals. The tool can also tell us where we have our biggest gaps in spaces to be able to make improvements.”

The platform has also helped UVA detect gaps or disparities in COVID-19 testing and outcomes.

“A lot of my work is around health equity and identifying potential disparities in testing access – individuals who may have met testing criteria based on their symptoms, but maybe didn’t receive testing,” Salerno said.

“We’ve been able to look for care access disparities and outcomes disparities in the ICU, admissions, and mortalities. And that allows us to think about changing our processes and look at the data to see if these changes improved our disparities.”

The health system also adopted a fully-linked case management platform that can be used by nearly any community-based organization, hospital, or clinic. The platform enables closed-loop referrals and care plans that help health systems address social determinants.

“Multiple different types of agencies in our community have adopted the platform,” Salerno said.

“We’ve been able to use this platform to stand up our own community calling in response to coronavirus. It acts almost like a case management and call log platform for all of the individuals looking for access to a physician and call line, especially if they have yet to be part of our UVA system. So, they don’t have yet a medical record number, but we can connect them to social services if they need them.”

An essential part of implementing the platform and ensuring its success was working together with community partners to find quality solutions, Salerno explained.

“We had community organizations help us assess tools and look at the options, and tell us what was important to them. We also have a community coalition of organizations, including housing providers, food providers, mental health providers, transportation groups, and education leaders, that are helping us implement the tool in a community-wide way,” she said.

“You have to strategically align the goals of your organization with those of your community and have conversations with them about what would be helpful. What are their fears, and how can you support them?”

Also critical for an effort like this: good data.

“You have to have really robust, quality data before people are willing to invest the resources for change,” she said.

“We’re an academic medical center, so for us, data is king. Having a comprehensive data platform that uses natural language processing to identify our patients’ needs, and enables us to connect them to resources, was a really great place for us to start.”

Going forward, these cross-sector partnerships, coupled with innovative platforms and tools, will help healthcare organizations better address patients’ social determinants of health, leading to improved outcomes.

“Health is so dependent on the non-clinical factors of your life, like food access, safe and affordable housing, employment, and other determinants. Healthcare providers are not the experts in any of those factors, although all of those factors contribute to overall well-being,” Salerno concluded.

“Making sure that you’re partnered in creating a comprehensive system to support people along their journey and in taking care of their health and well-being is critical. Knowing that your community partners are experts in what they do, and being able to lean into them and ask for their help – those mutually beneficial partnerships are essential.” 

Next Steps

Dig Deeper on Artificial intelligence in healthcare