Improving Care Through AI One Zip Code at a Time

By Matt Phillion

A new health equity AI tool is able to aggregate clinical and social data across 4,500 factors to provide a more granular view of health disparities in varying communities.

The tool, Radian, was created by Lightbeam Health Solutions to provide a no-cost health equity solution for healthcare organizations and leadership. Using just a ZIP code and pulling from an array of publicly available data, it provides demographics within one, 10, and 100 miles of the provided location; offers AI-informed areas of focus and impact for that population; and delivers clinically supported recommendations to address top vulnerabilities.

Lightbeam had already made moves to strengthen its bench in terms of AI development. “When Lightbeam completed its transaction with Jvion [AI], Jvion was well known, highly regarded in the market because of its clinical AI models: avoidable admissions, avoidable ED utilization, and more,” says Blake Marggraff, CEO with CareSignal, Lightbeam’s deviceless remote patient monitoring subsidiary. “They had also developed some pretty impressive IP around health equity. And when I say impressive, I mean that in a quantitively defensible sense: Their health equity AI’s performance is off the charts relative to something like the CDC social vulnerability index.”

This was happening right around the time CMS started to shift toward health equity being a component of new models.

“A lot of the industry is looking toward health equity and equity AI, and Radian is designed to show what that means, showing everyone from executives to licensed social workers at rural health clinics what we can do by taking data from publicly available resources, then applying robust machine learning and deep learning protocols to figure out where to focus for the best impact,” says Marggraff.

How it works

The idea behind the tool is to offer a common foundation that enables any organization, not just those who are ahead of the technology curve, to use AI to improve outcomes. To do this, Radian pulls data from publicly available resources such as the CDC and FDA, but also less obvious resources that relate to social determinants of health like the office of Housing and Urban Development. Then, rather than focusing on the problems in a given geographic region, it looks at the solutions that might help improve outcomes in that area.

“It leverages models designed not just to identify the biggest problems today and in the future, but the biggest opportunities,” says Marggraff. For example, in a given region, it might zero in on preventive screenings for elderly men followed by education for patients on access to care and how they can get the appointments they need.

“There are thousands of different patient-specific recommendations, but Radian boils it down to 25 for a given community and makes those recommendations based on a selected radius,” says Marggraff.

This geographical and distance-related component can help organizations tailor their plans. A single clinic with 20,000 patients may want to look at a smaller radius to help with their most underserved members, while a larger practice or system may want to use a 100-mile radius to address more patients’ needs.

Using the tool simply involves entering a ZIP code and a distance parameter, from which Radian identifies top trending issues in the area and generates educational and marketing materials that the organization or provider can distribute to patients. These AI-generated materials take the region’s population into account to maximize their impact.

An inflection point

Radian is intended as a steppingstone toward more targeted care, Marggraff explains.

“We want to have the greatest possible impact on real people’s lives and on the organizations who serve those people,” he says. “Health equity is at this moment at an inflection point, almost like where behavioral health was in 2015–2016: one of those situations where doing the right thing is the right thing for the patient and for the market. This requires making sure everybody has a common foundation, and it needs to be equitable not just for the patient but for organizations as well.”

Making sure that organizations can prioritize where they will make the greatest impact, and implement the right solution at the right time, has consistently informed Lightbeam’s strategy. Now, healthcare’s increased openness toward technologies like telehealth has helped usher in the use of AI-based concepts like Radian.

“We’re now seeing organizations can start [taking advantage of AI] without needing to go through a three- or six-month process just to use technology that benefits their staff, patients, and members,” says Marggraff. “There’s so much more beyond that. I’d like to scale beyond the current offering; that goes without saying. But what we’re already seeing, we’re generating all sorts of dashboards from top five health systems but also small associations and community health centers. And those are the ones who will benefit the most from this technology.”

The impact of staffing shortages

As providers move out of the industry at alarming numbers, can targeted, high-impact solutions for patients help address this growing issue?

“Even before the pandemic, the staffing crisis was expected to get worse at least through 2026,” says Marggraff. “The solution to that problem is technology that elevates teammates to the top of their licenses, that meets patients where they are, and allows organizations to move from reactive to proactive.”

Technology like Radian can be a piece of solving that puzzle, he says. “It’s not enough to just show an organization where their biggest problems are,” Marggraff says. “It’s not even enough to make really good recommendations. You need to partner together to provide scalable solutions to deploy those recommendations, including for populations who have traditionally been tough to support.”

This is one of the positive effects of the shift to value-based care, says Marggraff. “We’re always optimizing with our partners for what is both financially sustainable and clinically impactful,” he says. “Going back five or six years, this was sometimes hard to do, but more and more it’s not—and that’s about as energizing as it gets, when the mission and the financial upside are aligned.”

A sense of urgency

If there’s any particular challenge, it’s that groups will wait until the last minute to invest in health equity, says Marggraff. “Just like with any change, it doesn’t seem urgent until things are already on fire,” he says. “I’d rather see things not get to that point.”

AI also remains a nebulous term to those who don’t work with it. “When our conversations start, AI sounds great, but most folks still don’t know what that means,” says Marggraff. “Jvion goes beyond analytics and statistics—a true cutting-edge opportunity for identification.”

There’s often an “aha moment” for users, Marggraff says, when they are both intrigued by the technology but can also see that it’s addressing something that will rapidly grow into a pain point.

“I think we’re at a maturation point where it’s more than just a good solution for parts of the problem, but rather all the puzzle pieces to be put together in the right way to solve for, from analytics to prediction to scalability to long-term downstream patient support,” says Marggraff. “All of those things coming together in one place—it’s about the defragmentation of healthcare. Reducing entropy from U.S. healthcare is about as ambitious a vision as you can get.”

Matt Phillion is a freelance writer covering healthcare, cybersecurity, and more. He can be reached at matthew.phillion@gmail.com.