Why It Matters
As soon as the first COVID-19 vaccines arrived in Dallas, Texas, leaders at Parkland Health & Hospital System were focused on equity. “We knew that there could be disparities in the distribution of it,” said Brett Moran, MD, Chief Medical Information Officer at Parkland. “We wanted to be very intentional [to avoid inequities].” To meet that goal, Parkland, Parkland Center for Clinical Innovation (PCCI), and their partners launched a comprehensive effort that incorporates both cutting-edge data analytics and more traditional approaches like church meetings.
Before the pandemic, Parkland and its community partners already had a strong foundation of collaboration. Parkland was working with Dallas County Health and Human Services to integrate public health functions into health care delivery. PCCI, an advanced data analytics organization affiliated with the health system, played a key role. For instance, PCCI developed a Digital Data Environment and a Community Vulnerability Compass (i.e., a Social Needs Index) for the whole county to evaluate social determinants of health and resiliency metrics at the block level within neighborhoods.
As a result, in early 2020, these partners were able to rapidly pivot to address the spread of the virus. They established a data-sharing agreement so PCCI could use current COVID-19 case reports from the county for Geographic Information System mapping and identifying where to focus their efforts. “We were blessed by the level of data and the partnerships we had,” said Steve Miff, President and CEO of PCCI.
Leveraging the Community Vulnerability Compass and their data science expertise, Miff and his team developed two indices specific to the pandemic: a COVID-19 Vulnerability Index and a COVID-19 Proximity Index. For individuals living within Dallas County, the Vulnerability Index incorporates stable risk factors, such as age and medical comorbidity, with dynamic risk factors, such as COVID-19 case density in the area and mobility. The personal risk score, or Proximity Index, is based on an individual’s potential exposure to COVID-19 cases via dynamic personal proximity to confirmed, active cases around them. The Proximity Index was used at Parkland to guide and inform patient care and patient-level care coordination and has also been made available to the public to promote personal awareness (see Figure 1).
Figure 1. Dallas County MyPCI: COVID-19 Personal Risk Score | Source: PCCI and Dallas County
Handling such large volumes of data, much of it sensitive, presented challenges. One was “how to present [data] in a way that maintains confidentiality,” Miff said. The team found ways to provide a view that is sufficiently granular to highlight hot spots, but not granular enough to disclose individuals’ information (see Figure 2).
Figure 2. Example PCCI Aggregate County-Level COVID-19 Data | Source: PCCI and NEJM Catalyst
The partnership initially used this rich data set to determine where to set up COVID-19 testing sites and whom to refer for testing. When the vaccine became available, they used it to distribute the limited supply in a strategic and equitable way.
After health care workers, adults aged 65 and older and those with certain underlying conditions were the first group of people eligible for COVID-19 vaccination. Parkland leaders knew that a large proportion of the community’s population met those criteria — about a million people — and that Parkland would be administering the vaccine to them. “The question was,” said Moran, “How do we do that equitably?”
Miff and his team created several analytics tools, which they used to rank all eligible individuals based on the COVID-19 Vulnerability Index. The ranking was intended to prioritize the most vulnerable, whether due to medical, behavioral, or social needs. Then the team began reaching out to those at the top of the list — through calls, emails, and text messages. Vaccines were administered by appointment at four mega-centers.
The partnership tracked COVID-19 vaccinations in Dallas County to monitor and assess their efforts. This tracking data also informed where they needed to focus outreach. By the second week of January 2021, they were vaccinating more than 4,000 people each day.
The partnership worked with Dallas County Health and Human Services (DCHHS), local thought leaders, community-based organizations, and advertisers to convey their messages to communities. Parkland experts made frequent appearances in the local media, including Black and Hispanic radio and television stations. At churches, they held educational events with Q&A sessions. They also enlisted prominent community members — such as pastors and sports stars — to film videos of themselves getting the shot and explaining why. They then circulated the videos on social media and through traditional media outlets.
At this stage of the COVID-19 pandemic, the challenges have evolved. “There’s plenty of [vaccine] supply but dwindling demand,” said Miff. “One of the things we’ve used analytics to do is identify blocks [within neighborhoods] with the highest number of unvaccinated individuals.” Parkland and the DCHHS teams are then using this data to determine where to locate pop-up vaccination sites. The county has also deployed volunteers to go door-to-door.
Now, the data are sending a new message: their hard work is paying off. Parkland has administered COVID-19 vaccines to more than 51 percent of community members at highest risk, those ranked 4 or 5 on the Social Needs Index scale by zip code — an impressive accomplishment.
Miff noted “the importance of having hyper-localized data — these data are so core to the whole journey through the pandemic.” Thanks to the data, the team was able to target efforts in a highly intentional way to reach individuals who most needed the vaccine. “One thing we learned from the pandemic was the importance of having this data available and the collaboration and partnerships to take action on them,” he said.
As crucial as technology has been in identifying needs, sometimes low-tech approaches are more appropriate for meeting those needs. “You can’t expect everyone to log in on their smartphone to schedule a vaccine appointment,” said Moran. That is why Parkland’s strategies have included everything from telephoning residents to knocking on doors to offering patients the vaccine on the spot when they came in for other appointments. Staff also help patients schedule COVID-19 vaccine appointments, whether by assisting the patient or their spouse in enrolling in the online patient portal, or just directly scheduling an appointment for them. “We went from the very savvy data analytics,” said Moran, “to the simple level of handing someone a piece of paper with instructions on when and where to go to get a shot.”
Editor’s note: The Parkland Health & Hospital System participated in the IHI COVID-19 Rapid Learning Initiative. IHI gratefully acknowledges Pfizer, Inc., and BD (Becton, Dickinson and Company) for their generous funding support of the COVID-19 Vaccine Rapid Learning Initiative and their leadership and expertise in the drive for vaccine distribution.
You may also be interested in:
NEJM Catalyst — “The Imperative for Integrating Public Health and Health Care Delivery Systems”
Bringing Health Centers Together to Tackle COVID-19
Identifying the Keys to Effective Vaccine Communication and Outreach