Predictive value of individual social risk versus neighborhood-level social vulnerability for trauma outcomes.

BACKGROUND

Neighborhood-level indices serve as proxies for social risk in clinical research, although self-reported social determinants of health may better identify vulnerable patients, especially in the context of traumatic injury. We hypothesized that self-reported social determinants of health have stronger predictive value for short-term trauma outcomes than neighborhood aggregates.

METHODS

Adult inpatients with trauma tic injury (2020-2024) who completed hospital-administered social determinants of health screeners were assigned Social Vulnerability Index, Area Deprivation Index, and Environmental Justice Index scores based on census tract and categorized as high risk/low risk using median splits. Patients were also categorized as at risk/no risk across the 8 self-reported social determinants of health domains, measured by the screener. Primary outcomes were 30-day readmissions and hospital length of stay. Regression models measured the association of social determinants of health measures with outcomes, adjusting for injury and patient factors.

RESULTS

A total of 3,115 patients completed social determinants of health screeners with 917 screened pre-trauma (median 152 days) and 2,198 screened post-trauma (median 17 days). Social Vulnerability Index and Environmental Justice Index predicted readmission risk (odds ratio 1.41 [95% confidence interval 1.07-1.86], P = .014, and 1.37 [1.00-1.87], P = .047). Social isolation was associated with significantly greater odds of readmission (odds ratio 2.17 [95% confidence interval 1.12-4.76], P = .032). No other social determinants of health domains predicted readmissions. Social isolation and stress were associated with longer length of stay (β = 1.44 days [95% confidence interval 0.36-2.52], P = .009, and β = 1.31 days [95% confidence interval 0.37-2.24], P = .006). No neighborhood measures predicted length of stay.

CONCLUSIONS

Social isolation demonstrated a stronger association with 30-day readmissions than neighborhood measures for trauma inpatients. Both social isolation and stress predicted hospital length of stay. Interventions to mitigate isolation and to increase outpatient support after discharge may be effective in reducing readmission in high-risk trauma patients.