Improving Social Determinants Of Health Data

How to Improve Social Determinants of Health Data

Drs. Lisa Lines and Denise Clayton continue their conversation on social determinants of health data. In this video, they discuss our proprietary Local Social Inequity score (LSI) and how it helps organizations account for social risk.

This is the second in a four-part series of videos about social determinants of health. Other topics in this interview series include:

• Current Challenges with Social Determinants of Health Data
• Key Findings from Social Determinants of Health Analyses
• Use a Composite Score of Data to Find the Best Intervention Strategy

Denise Clayton: Hello and welcome to another episode of the RTI Health Advance series on social determinants of health. My name is Denise Clayton and I'm a health economist at RTI Health Advance. I’m joined today by my colleague Lisa Lines, a Senior Health Services Researcher. (0:15)

In our first video, we talked about current challenges with social determinants of health data and new approaches to incorporate better data to produce better health using the Local Social Inequity score or LSI. (0:28)

Today, we'll be talking about the applications of the LSI. To get us started Lisa, can you tell us about the goals or applications of the LSI in terms of helping organizations account for social risk and value-based payment models? (0:44)

Healthcare risk adjustment and social determinants of health data: accounting for social risk in value-based payment models

Lisa Lines: The Local Social Inequity score can be directly incorporated into a risk adjustment model that predicts healthcare expenditures. This ties back to making sure that providers aren't punished for factors outside their control and that their payments accurately capture the risk factors their population is experiencing. When you do that, you reduce the incentive for providers to cherry pick patients. (1:12)

Denise Clayton: Even though the LSI is at the neighborhood level, it can still be incorporated as a variable in an individual level risk adjustment model. You don't have to make major changes to risk adjustment. You can simply add a new variable to an existing model for example, one that has age, sex, and diagnoses or HCCs. (1:36)

I'm interested to hear more about other applications related to identifying which interventions to implement and which patients or members need those interventions. (1:50)

Using social determinants of health data to understand the characteristics of high-risk populations

Lisa Lines: We think the LSI could be very useful in targeting interventions to higher risk populations and understanding the characteristics of higher risk populations. Say you're working on a new intervention or support program for social determinants. How do you identify the people who need it and get it to them before they have a major health event? You can use the LSI to identify the top decile of the population at highest risk, the top percentile, or the 99.5^th percentile and then focus efforts on just those individuals at highest risk living in higher-risk communities. (2:29)

Just because we're risk adjusting for social determinants does not remove the obligation to address the social determinants. The artificial intelligence risk adjustment approach can be used to inform policies to address root causes. If you're a provider or payer or a policymaker and you have a budget to address social determinants of health, how do you know what's most important to address? (3:01)

Food insecurity, housing insecurity, transportation—they all sound very important. The neat part about our approaches is that you can run the causal models to tease out the root causes and use them to help guide decision making about where to invest and what to address—getting more bang for the buck really. (3:22)

Using social determinants of health data to evaluate the effectiveness of SDoH programs

Denise Clayton: We often talk to payers and hear that they want to implement all these new programs. But they also want to evaluate them and understand what's working, what's not, and why. The LSI has a role in evaluating a social determinants of health program or evaluating some other program and wanting to control for social risk factors. That's another application where the LSI comes into play. (3:55)

If you're trying to identify the causal effect of an intervention where a randomized trial didn't happen, having a good control variable or variables that help you with matching to create the best comparison group. Those variables help you improve the approach to quantify the effects of whatever you're trying to measure. You get the best estimate possible about what the true causal effect is. It can be used in evaluation of a non-social determinants of health program and be a control for social determinants of health. Or it can be used in an evaluation of actual social determinants of health programs as well. (4:35)

Lisa Lines: And the analysts don't have to decide amongst different social determinants of health variables and deal with collinearity and all the things that happened when you have multiple social determinants variables going in. Having one single composite measure is useful for evaluation. (4:56)

Denise Clayton: That's a great point. The LSI can be at different levels so if you're doing an evaluation at the individual level and you've got zip code level data to merge into the LSI you can do that. Or if you've only got county, you can merge it in that way. It's quite flexible because of the different level of granularity that it has. (5:15)

Thank you so much Lisa for talking with me today. These applications have a lot of potential to improve health and health equity. And thanks to the listeners for joining us for this discussion. Please stay tuned for more episodes in our social determinants of health series. If you'd like to learn more, you can contact us at RTIHealthAdvance.org. We're happy to talk with you further. (5:37)

Learn about health equity solutions from RTI Health Advance.