More on Population Health

University of Virginia Health System emphasizes community-first approach to social determinants

With the help of natural language processing and predictive capabilities, UVA has forged community partnerships during a time of need.

Jeff Lagasse, Associate Editor

The COVID-19 coronavirus is creating a wealth of health-related hardships for healthcare consumers, and, while infection is obviously at the top of that list, there are many secondary challenges resulting from the pandemic, especially when it comes to social determinants of health. The virus' spread is exacerbating these SDOH challenges, but the University of Virginia Health System is finding ways to address them by emphasizing a community-first approach.

The strategic decision to partner with community organizations came in 2018. Leadership at the system saw that there was a direct relationship between SDOH and length of stay for patients, and estimated that about 80% of a person's health status is determined by non-clinical factors, such as access to healthy foods, stable housing, reliable transportation and well-paying jobs.

These factors have always existed, but the ongoing health crisis has only served to exacerbate them.

"It's highlighted all of the other SDOH even more," said UVA Health's Dr. Amy Salerno, who treats patients in general medicine, geriatrics and palliative care. "It impacts people's risk for having coronavirus – if someone becomes homeless for example, you then have to go into a shelter where you have group living, and your risk of exposure is much higher."

Add to that the access challenges common to at-risk populations, and the pandemic has made the situation combustible. UVA's approach was to select a technology-based referral network to better connect patients to community resources. The mission to partner with those in local communities to improve patient health required building a strong network to connect the system with community groups who were already working to achieve similar goals.


UVA Health selected Pieces Technologies as a partner in this endeavor. One of the biggest technological assists from this team-up came in the form of natural language processing. Many health systems screen their patients for social needs based on questionnaires. While UVA is in the process of implementing that, the entire system is already using NLP to record conversations and doctor's notes, which are then embedded into the EHR. The result allows care teams to identify social determinants and connect patients to relevant local resources.

"We can proactively identify those who have the need, look for disparities and connect them to the community resources they need," said Salerno. "It's been amazing."

Clinicians have access to all of the data the health system is documenting, and with EHR integration can then pinpoint the most pressing social determinants for a given individual and then present that information to the relevant teams. Staff can also measure improvements in outcomes over time, an ongoing process that only adds to the wealth of information at the care teams' disposal.

Armed with this knowledge, UVA's Office of Community Partnership has been able to forge relationships with community organizations covering such things as transportation, secondary education, housing, food and mental health. These varied entities have begun to work in tandem, and the system itself has been able to identify where the community needs to invest more resources and where the priorities should be in terms of funding and grant applications.

"It's helpful for the community benefits office to think of where the community benefit dollars are going to be the most impactful," said Salerno.

Dr. Ruben Amarasingham, founder and CEO of Pieces, said the relationships between SDOH and key system outcomes are becoming more and more evident.

"Length of stay and avoidable hospitalizations for chronic conditions have been linked to social determinants and how health systems can tackle those problems," said Amarasingham. "It has gotten the attention of critical operators, and that's important because it's impactful when it comes to health system outcomes."

Previously, health systems typically had a one-way orientation to the community. What UVA Health's approach has done is create a community-first mindset, giving local organizations a major voice in how workflows and operations are conducted. In that way, the system is approaching the problem of SDOH from a dual perspective.

"I think that's especially important now during the pandemic," said Amarasingham. "If we're going to defeat this, the importance of communities working together is going to be critical."

"It's especially critical for the financial impact from the pandemic," said Salerno. "More and more people are suffering from some of the more acute social needs, but the community organizations themselves are working through this as well."

While the tech-enabled community-first mindset is still in relative infancy at UVA Health, Salerno expects that in time the data will show improved outcomes, which could be a lifeline for socially challenged patients during a confusing and stressful time in human history.

"The prediction side has been really critical for helping us identify the patients who have social needs," said Salerno. "If we didn't have the NLP prediction side of things, we would never know if we have any issues.

"Meeting the community where they are is really critical," she said.

Twitter: @JELagasse

Email the writer: