Earlier this year, and in collaboration with Netrin Health, we had the opportunity to support Milinda Polisetty, a graduate student from George Mason University’s Health Informatics program. She conducted research under Netrin’s preceptorship, applying predictive modeling to identify preventable high-cost utilization driven by both clinical and social factors.
Ms. Polisetty accomplished this through the application of machine learning and predictive modeling designed to equip care management teams with actionable insights beyond clinical and claims data to intervene early, mitigate high utilization, and improve outcomes.
Research Approach: Combining Health Informatics and SDOH Data
The full report is available ungated here. In summary though, her research utilized data from nearly 13,000 patients and merged clinical and claims data with social risk data powered by Socially Determined.
Outcome variables were defined around two high-risk groups: patients with three or more ER visits in a year, and those with annual costs exceeding $8,001.
The modeling process integrated over 20 variables, including illness band, pharmacy and medical risk scores, and chronic condition indicators such as hypertension, diabetes, and cardiac disease. These were analyzed alongside five of Socially Determined’s social risk domains: financial strain, housing instability, health literacy, food insecurity, and transportation barriers.
Findings: Social Risk as a Driver of Utilization
This analysis affirmed that it is, “Evident that healthcare overutilization is not solely a clinical issue but is equally driven by social risks.”
Frequent ER users were found to be heavily influenced by social risk, whereas high-cost patients were influenced by a combination of clinical burden, care quality gaps, and social risk.
Through this approach, 70 frequent ER users and 1,540 high-cost patients were identified for targeted support. These insights guided Netrin Health’s care management team in outreach, screening, and intervention strategies for these at-risk patients.
The research further demonstrated how social risks can cluster and reinforce each other, driving home an acute message: these are complex issues, and we have the data to help make tailored, focused interventions not just possible, but highly effective.
Finally, this research shows that the need for tailored interventions is paramount, but the impact of social risk shared between two patients can be radically different. This underscores the need to understand the entire social risk landscape. At first blush, an ideal intervention may face additional barriers.
From Insight to Action
Netrin Health quickly operationalized the findings. Using its infrastructure of care teams, workflows, analytics tools, and patient engagement systems, Netrin integrated a social risk lens into its existing programs. This ensured that outreach efforts were focused on the patients most likely to benefit from intervention, thereby maximizing the impact on both health outcomes and cost reduction. The framework continues to evolve through iterative model refinement, patient feedback, and care team collaboration.
Socially Determined Isn’t Just Our Name
Two of Socially Determined’s five core values are “Advance Wellness” and “Understand Impact.” We strive to help payers, health systems, partners, and life science companies do exactly that.
If researchers, intervention organizations, academic departments or institutions are eager to dive into SDOH work and benefit from our insights, don’t hesitate to contact us today.