By Melinda Young
Case managers and researchers need data that can be used to improve care coordination and prevent hospitalizations and ED visits. But the challenge is deciding which patient data are useful and which are not.
The authors of a recent study analyzed whether patient-reported health-related social needs (HRSNs) measures could help predict utilization and health outcomes. The findings were mixed, with some reason to believe that collecting HRSNs may aid efforts to improve care delivery. However, HRSNs are less useful in predicting utilization and outcomes.1
“Predictive value is of interest because we want to know who is more likely to have [worse] health outcomes and more utilization, and this could be a way to develop interventions to mitigate the impact,” says Samuel Savitz, PhD, lead study author and an assistant professor of health services research at the Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery at Mayo Clinic. “With our study, the results are mixed, but collecting social needs data seems to have some value for outcomes.”
There is a lot of literature on how social needs affect patients’ health and health outcomes, so healthcare systems will need these data to help them improve care delivery, Savitz says.
Savitz and colleagues identified patients with cancer, diabetes, or heart failure and studied their social needs, using standard social determinants of health (SDOH) questions. The patients received care at one of the Mayo Clinic’s main campuses in Rochester, MN, Arizona, and Florida. They also included some patients in several other states. Then, they used data to predict outcomes after baseline, including hospitalizations, 30-day readmissions, and death.
Savitz and colleagues used one model that did not include social needs data and another model that included those data. They compared the two to see how much predictive performance changed when the social needs data was added.
“In some cases, there was significant improvement; in others, there was not,” Savitz says. “When there was improvement, it was not a huge gain in performance, but a more moderate gain.”
Improving Outcomes with Data Collection
Healthcare professionals recognize the importance of SDOH, but they are less clear on how collecting these data can help them improve outcomes.
While the goal is for a dramatic and consistent effect, the results did not bear this out. “The fact that we’re not seeing it is because there are other factors that go into determining these outcomes,” Savitz explains. “In our study, we included an area deprivation index [ADI].”
The ADI is among other data with the potential to change utilization, including clinical factors and demographics. “The fact that we already had the area deprivation index in [the data] may have impacted the results,” Savitz notes. “We wanted to see the additional value of these social needs measures that were directly asked of patients when compared with variables already available [in charts].”
Another factor is that Savitz and colleagues only studied whether patients had been hospitalized and whether they were readmitted. They did not look at how many times patients had been hospitalized. It is possible that collecting social needs data could have a predictive quality for patients who are hospitalized several times in one year, but the study could not speak to that potential.
“That wouldn’t show up in our data,” Savitz says. “The way we measure in our study, we can’t speak confidently either way.”
Collecting SDOH can help with research, but that is a secondary use for the data. “The main use is about how you can use this information to connect patients to social services,” Savitz says. “We didn’t touch upon that. It’s still the main reason this information is collected.”
Savitz and colleagues also found that there was a lot of missing information. “A lot of social needs data were incomplete, and that may be another explanation for why we didn’t get predictive improvements,” Savitz says. “If there’s a way to improve the data, that could be a benefit.”
Social services organizations could use more complete data to develop programs to help people with various social needs. “If you don’t know the information, you can’t act on it,” Savitz says.
Healthcare organizations have been working on ways to collect social needs information more efficiently. One new method is to eliminate some of the questions to decrease the burden on patients filling out the information and on providers collecting the information. For example, one social needs question that was cut asks whether olive oil is a person’s main source of fat in their diet.
“There’s this thought that [the olive oil question] is not as valuable as asking about food insecurity,” Savitz explains. “If we keep the questions to the most essential ones, then hopefully that will reduce the burden and increase completion of the questions.”
The goal is to collect more complete data to improve care coordination, case management, and other healthcare services. “There is value to understand the interventions we can tie to this and to see how we can use this for prediction,” Savitz says. “I think there is value in having different uses for the information. If we’re asking patients to complete this, I think we owe it to patients to use the information to improve care. Otherwise, why do we collect it?”
REFERENCE
- Savitz ST, Inselman S, Nyman MA, Lee M. Evaluation of predictive value of routinely collected health-related social needs measures. Popul Health Manag 2023;Oct 30. doi: 10.1089/pop.2023.0129. [Online ahead of print].
Case managers and researchers need data that can be used to improve care coordination and prevent hospitalizations and ED visits. But the challenge is deciding which patient data are useful and which are not.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content