(306) Examining Healthcare Disparities Among Survivors of Critical Illness With Sepsis

Introduction: Survivors of critical illness often face significant challenges, such as neurocognitive disorders, physical disabilities, and respiratory limitations. Healthcare disparities by race heighten such challenges. We hypothesize that 90-day and 365-day mortality will be lower among Black patients with sepsis who survive critical illness.

Methods: Using a large database of survivors of critical illness with sepsis, the association between race and 90-day and 365-day mortality is examined by race. A directed acyclic graph (DAG) is used to model causal presumptions for our analysis (daggity.net, Textor et al International Journal of Epidemiology 45(6):1887-1894, 2016.). Accelerated failure time (AFT) modeling is employed, controlling for our DAG-informed covariates. Non-parametric analysis is performed using Kaplan-Meier survival curves adjusted for age and sex.

Results: Our dataset comprised 62135 intensive care unit (ICU) survivors with sepsis (30659 (49.3%) male; 9402 (15.1%) Black race; mean age 64.7). Our DAG model reflects potential confounding for age and gender, with other covariates such as comorbidities, structural determinants of health, and socioeconomic status mediating the effect of race on 90-day and 1-year mortality (Figure 1). AFT models suggest that Black patients who survive the ICU are 21.3% less likely to survive to 90 days than White patients (coefficient for association between Black race and 90-day mortality = -0.239 (95% CI -0.405, -0.073);p = 0.005), and 19.5% less likely to survive to 1 year (coefficient = -0.217 (95% CI -0.287, -0.147); p < 0.001). Kaplan-Meier curves controlling for average age and sex (male) likewise reflected decreased survival for Black patients during the timepoint studied (Figure 2).

Conclusions: Our results suggest that Black patients who survive a hospitalization for sepsis that includes an ICU stay may be less likely to survive to 90 days and 1 year. Further research identifying effective interventions that may reduce healthcare disparities among ICU survivors is warranted.