At Socially Determined, we are deeply dedicated to the science behind our SDOH and HRSN data. That’s why in May 2026, Chintan Dalal, MS, PhD, our principal data scientist, Dan McMichael, our VP of Data Science, and Leah Broder, MPH, our chief customer officer shared a recent social risk research study at ISPOR 2026.
The study, “Social Determinants of Health Enable Early Identification of Premature Mortality Risk” (You can view the poster here) was initiated to evaluate the association between SDOH burden and age at death and the value of SDOH-based risk stratification for early prioritization. It was built with mortality data from Veritas Data Research.
In plain terms, the study set out to answer a specific question: “Independent of any known diagnosis, is a person's social risk associated with how early they die? More concretely, among people who have died, are those with higher SDOH burden more likely to have died before age 75, the standard threshold for premature mortality?"
If you’re interested in the data science behind the intersection of SDOH and real-world impacts on mortality, this post is definitely for you.
Here’s what the Socially Determined research team did.
The methods involved analyzing a national, de-identified mortality dataset of approximately one million individuals and linking person-level SDOH risk scores with age at and cause of death. Individuals were ranked by SDOH risk using multivariable stratification.
Premature mortality was defined as death before age 75 and quantified as years of life lost (YLL). Regression-based analyses were used to assess differences in YLL across SDOH risk strata and COD cohorts. Model performance was evaluated based on the concentration of YLL among higher SDOH risk rankings relative to uniform prioritization. Mean age-at-death differences between SDOH burden groups were estimated across COD categories, with uncertainty summarized using confidence intervals.
In effect, our team took our own SDOH risk scores, and once we had a record linkage, approximately 600,000 individuals were run through our risk model to find any strong signals that indicated high SDOH risk scores in different domains, finding that the likelihood of living past 75 drops.
Then, they took the model and estimation, stratified the cohort into percentiles ranked highest to lowest risk.
Here’s what the SDOH mortality study found.
The results were illustrative of Socially Determined’s larger goal. The SDOH-based risk ranking identified individuals with higher number of years of life lost relative to national averages. Individuals ranked at the highest SDOH risk accounted for approximately 2.7 times more premature mortality than expected under uniform prioritization. Higher SDOH burden was associated with earlier ages at death across multiple cause of death (COD) categories, with mean age-at-death differences between high- and low-SDOH burden groups ranging from 10 to 20 years for leading causes.
The researchers concluded that large-scale analyses indicate that SDOH burden is associated with higher years of life lost (YLL) and is reflected in the timing of premature mortality rather than recorded COD alone.
In turn, an SDOH-based risk ranking can be used to identify individuals at elevated risk of YLL and provides value for early prioritization and intervention. Integrating person-level SDOH risk data with mortality and cause-of-death information enables precise characterization of premature mortality patterns and informs population health management and prevention-focused investment.
There’s no Question: SDOH data has moved into applied sciences.
Making a real difference in social determinants of health, for both costs of care and patient outcomes, means constantly engaging with the science of data analysis. We are extremely proud of our team’s commitment to enriching not just our own understanding, but how healthcare sees this critical element of care. We’re also grateful for the contributions from Veritas Data Research.
If you have any questions, don’t hesitate to reach out today.