More on Analytics

Getting a handle on social determinants of health requires investing in predictive analytics, EHR integration

Social determinants are important to track and can lead to improved finances, but there need to be better and more consistent methods of tracking.

Jeff Lagasse, Associate Editor

Increasingly, hospitals and health systems looking to tighten up their finances -- and improve clinical care -- are attempting to address social determinants of health, the nonclinical factors that influence people's ability to maintain their health.

These social determinants can encompass anything from financial security and transportation to home stability and literacy. Understanding social determinants has become more important as providers try to manage their costs and thrive in a cost-sharing or value-based framework.

The problem lies in tracking these social determinants. They're typically not integrated into a health system's electronic health records, which puts the onus on things like predictive analytics to make sense of it all.


Ryan Bengtson, senior vice president of clinical innovation at Connance, a Waystar company, sees the evolution of EHR systems as the reason why they historically haven't taken social determinants into account.

Most of the major players, he said, have created a module in their EHRs to manage social determinants, but a lot of organizations aren't collecting the information -- at least not in a  standardized way that can be used on a data system.

Getting a better handle on them can yield a significant return on investment, and one of the areas in which this tends to be the case involves readmissions.

"Readmissions is a high-focus area where hospitals have incurred a lot of additional costs," said Bengtson. "CMS started the Hospital Readmission Reduction Program, and included penalties for hospitals that had excess readmissions.

"We looked at building predictive models looking at the social determinant model that we had, and they were significantly more accurate when factoring in the social determinants. They're more influential than the fact that a patient had co-morbidity factors."


In fact, according to Connance's own data, about half of hospital readmissions are rooted in social determinants. Statistics like that make it clear that health systems would do well to allocate resources to tracking and addressing them.

"Every organization has a finite number of physical resources and human resources to put toward these things, and they're looking for ways to apply this" said Bengtson. "Once you stratify the population, that gives you a roadmap around what those services are, what they need."

The intersection between the highest-risk social determinants patients and the total cost-of-care patients are where systems should start focusing their efforts, he said.

"If you've got a cohort of patients that are high-risk, or costs are rising, you need to be able to focus and understand those needs and identify those resources in your community," said Bengston.

Partnering with those community resources is an integral piece of the puzzle, and to understand which resources to focus on, a system needs to understand what the prevailing social determinants are in the immediate geographic area.

"It is very market-specific, more geographically focused" said Bengtson. "We certainly find there are organizations in more financially depressed areas that have issues with food deserts or lack of transportation, even in the catchment of a typical health system. You're going to find areas that are better off, and areas that are struggling. There's no rule of thumb for what they need, but it does help identify either care gaps or an understanding of the prevalence and severity of some of these challenges."


So systems for tracking social determinants are in place, and an organization has crafted a strategy based on the geographically specific needs of its population. What are some of the results that can be attained?

Geisinger tried it. It used social determinants to focus on a higher-risk subset of its diabetes population, and by using that information to to target more of those patients with an intensive treatment program, they were able to see a $6,000-per-patient benefit, said Bengtson. That's a significant return on investment.

Connance also did a study looking at patient no-shows for ambulatory appointments.

"We had a customer who was having almost one in four of their appointments no-show," said Bengtson. "By using social determinants to show who had financial factors that could impact their decision, we were able to do targeted outreach. Patients are offered ridesharing, and we tell them it's covered by insurance.

"There was a 40 percent reduction in no-show appointments by working in that way," he said. "Patients were better engaged."


Many have begun screening for social determinants on an encounter-by-encounter basis, and in some case they're capturing that information in a codified manner. But in the future, organizations will have to abandon that encounter-by-encounter approach and identify patients at scale.

"It's not a lot different from Google and Amazon and others who have information about our consumer preferences," said Bengtson. "Not to get too Big Brother-y, but there's a role for technology to share with providers what these social determinants are."

The future, he said, consists of more codified criteria for what those factors are, and having the information integrated more consistently to the EHR, sitting alongside the clinical factors that are typically tracked.

"We've got to be looking more proactively at how we invest the resources … so we're not dealing with excess length of stay, admissions, and all the other things that whittle down the bottom line," said Bengtson.

Twitter: @JELagasse

Email the writer:

Air Max 1 Ultra Flyknit