Racial, gender and language based disparities in sepsis: a public health perspective.

Existence of linguistic, gender and ethno-racial differences in patients with sepsis remains relatively unknown, especially in the public health domain. Retrospective analysis of data reported by a hospital in South Brooklyn to a New York State sepsis registry was undertaken over a 24-month period. Inclusion criteria were age over 18, available linguistic, gender and ethno-racial data, and registration in the New York State sepsis registry. Patients with missing data fields were excluded from the study. Primary outcome was the correlation of gender, race, ethnic & language-based differences with overall sepsis-based mortality. Secondary outcomes were the correlation of the same demographic variables with rates of mechanical ventilation, vasopressor use, intensive care unit (ICU) admission rates & overall incidence of sepsis and septic shock. Results: 677 patients were included in the final analysis in this single center retrospective observational cohort study and multiple statistically significant primary and secondary outcomes were found. Non-English-speaking patients had a higher incidence rate of sepsis-based mortality when compared to their English-speaking cohorts. The incidence rate difference is -0.36 (95% CI -0.49 to -0.22), with a P-value < 0.0001. A higher rate of vasopressor use was noted among non-English speaking patients when compared to their English speaking counterparts. The difference in incidence rates was -0.35, (95% CI -0.46 to -0.25) with a P-value of < 0.0001. Non-English-speaking patients had a higher incidence of receiving mechanical ventilation when compared to English-speaking cohorts. The incidence rate difference is -0.35 (95% CI -0.48 to -0.22), with a P-value < 0.0001. Non-English-speaking patients had a higher ICU admission rate with an incidence rate difference of -0.30, at a P value < 0.0001. Non-English-speaking patients had a higher sepsis incidence rate (0.70, 95% CI 0.63 to 0.78) compared to English-speaking patients (0.30, 95% CI 0.25 to 0.36), with a P-value of < 0.0001. Non-English-speaking patients experienced a higher incidence of septic shock (0.70, 95% CI 0.63 to 0.78) compared to English-speaking patients (0.30, 95% CI 0.25 to 0.36), with a P-value of < 0.0001. Caucasians showed statistically significant and higher rates across all primary and secondary outcomes albeit with greater statistical fragility. No significant differences were noted with regards to the impact of gender on all outcomes. Conclusion: Significant and multiple linguistic and ethno-racial differences were noted in this single center study with regards to sepsis-based morbidity and mortality outcomes. These differences need to be validated in larger, multi-center trials and could inform future efforts focused on identifying higher risk subsets in patients presenting with sepsis and septic shock in a public health setting.