Social Determinants Of Health: Social Risk Factors

Social Risk Factors in Social Determinants of Health Data

Drs. Lisa Lines and Denise Clayton discuss the different levels of data and how to apply them together to find the best intervention strategies in the final video of our series on social determinants of health.

This is the fourth installment 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
• Using LSI to Improve Social Determinants of Health Data
• Key Findings from Social Determinants of Health Analyses

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. (0:14)

Previously, we talked about challenges with social determinants of health data, RTI's Local Social Inequities score (LSI), and some key applications and findings from our own analyses. Today we're talking about the importance of neighborhood level data, along with individual level data. (0:32)

How individual level data is used for social determinants of health

First, it's important to acknowledge the role of individual level data for social determinants of health. Lisa, can you tell us why and how individual level data is valuable? (0:44)

Lisa Lines: Individual level data on social risk factors enables payers and providers to identify people and their needs because you can't address a need if you don't know it exists. Neighborhood level data is not meant to replace the individual level data on risk factors that exist at the individual level like food insecurity. (1:12)

Denise Clayton: That's a great point and it’s also important to note that individual level data not only helps identify who and what, but those individual circumstances can change over time. You need up-to-date, individual level data. That's the data point you need if you're trying to determine who to target for a given intervention. (1:39)

How neighborhood level data is used for social determinants of health

Denise Clayton: You mentioned food insecurity. That's a great example of the importance of neighborhood level data because the right intervention for someone experiencing food insecurity might be different if they live in a healthy food desert versus an area with lots of food access where the root issue is income or even transportation. Can you talk about the value of neighborhood level data, both on its own and in conjunction with individual level data? (2:05)

Lisa Lines: There are some measures where the neighborhood level is the right level of data. Think about air pollution. There's indoor air pollution and outdoor air pollution. If you're measuring outdoor air pollution, you're going to be measuring at the neighborhood level. Same with something like access to public transportation. Whether there are airports nearby, manufacturing factories, solid waste facilities – all of these are important in terms of causing pollution like air pollution and water pollution. Everybody in that neighborhood is subjected to that pollution. (2:29)

Then there are situations where the neighborhood level data complements individual data. So, say that a particular child has severe asthma, and they end up in the emergency department. If this patient lives in a place where there's low quality or older housing stock maybe there is an environmental issue that's exacerbating that asthma like old carpets or poor indoor air quality. (3:26)

Knowing those neighborhood level data points on housing stock air quality can help you address the underlying cause of that asthma. Or maybe they're living near a highway. There’s not much you can do about that, but you won't know about that unless you have that neighborhood level data. Without that neighborhood level information, you continue to address the asthma attacks after they happen instead of addressing the root cause. (3:53)

Merge social determinants of health data sets to find the best healthcare intervention strategies

Denise Clayton: There are examples like that across a lot of different disease states even related to screening or the adoption of vaccines. In some cases, there might be neighborhood level data points that can help you understand what the best intervention is. It might be an individual level intervention for a specific disease state but given the context of the individual in a certain neighborhood there might be other interventions that will help not only that individual, but others as well. I'm glad you raise that asthma point, that a good example. (4:32)

A lot of people ask how do you work with individual level and neighborhood data together. How do you merge these data sets? How does that work? (4:44)

Lisa Lines: What people typically do is put lots of different indicators into a model. What we have done instead is this single score—sort of a composite measure—an omnibus measure that captures over 150 different indicators going into that score. That way you don't have to chase down 20 different variables and worry about collinearity. Instead, you have one variable at the zip code level, at the census tract level, or the county level. The census tract level and even smaller is ideal because you're dealing with smaller numbers of people, so the estimates are not as broad. (5:42)

If you just have zip code level information that will also be fine. It tells you something about the area where a person lives. County levels are the least ideal. Think about Los Angeles County with 10 million people. It's hard to make an estimate that's valid across 10 million people. If you can get the census tract or zip code level, that's much better. (6:13)

And if you have individual level data—like an individual's address—you can get census block level if you have their nine-digit zip code. But a five-digit zip code will work. And if you merge that individual level data with these zip codes into the scores or individual predictors you can combine the power of the information that you have about the individual with the place where they live. That gives you a much better picture of someone's particular situation than just looking at neighborhood or just looking at individual. (6:52)

Denise Clayton: A nice thing about the LSI is it exists at the census tract, zip code tabulation area, or county level like you were saying. But in addition to having that LSI variable which is that omnibus measure of risk, the data set that goes into creating that score is the 150 plus variables so the LSI's there when you need that high level overall risk score that correlates strongly with life expectancy at birth. As part of the data sets at that granular geographic level we also have all those 150 plus input variables so all the environmental or the food related factors are all in there as well. That can be combined with individual level data, in addition to the LSI which is a powerful tool. (7:41)

Lisa, thank you so much for talking with me today and to everyone for joining us for this discussion. I always enjoy getting to talk about health equity and sharing our insights on social determinants of health. If you'd like to learn more about our Local Social Inequity score or how it could be used in your organization, you can contact us at RTIHealthAdvance.org. We're happy to talk with you. (8:10)

Learn about health equity solutions from RTI Health Advance.