OBJECTIVES: Sepsis is a severe response to an infection that often requires hospitalization. Admissions for sepsis are costly, estimated to be $27 billion annually. There have been recent shifts towards building sepsis-focused economic models for hospitals. The objective of this study is to identify predictors associated with high-cost hospital admissions and worse outcomes due to sepsis.

METHODS: This is a retrospective cohort study of patients admitted to Barnes-Jewish Hospital with a diagnosis of sepsis or septic shock from 2014 to 2017. Each admission was given a discharge disposition score (DDS) ranging from 1 (worst outcome: death during admission) to 6 (best outcome: discharged to home). To create a charge-outcome value, the total charges per day of each admission was divided by the DDS. The best value-outcome was defined as $0 to $2,500 per DDS and poor value-outcome was defined as greater than $10,000 per DDS. Binary logistic regression was used to identify variables associated with the poor value-outcome category. All charges were adjusted to 2021 dollars.

RESULTS: 3,321 patients (mean age 60.7, SD 15.7) were included in the analysis. The median total charges for all admissions were $114,429.68 (IQR $270,119.09) and the median hospital length of stay was 11.7 days (IQR 18.2). The presence of cancer (OR 1.58; 95% CI 1.35-1.84), chronic kidney disease (OR 1.42; 95% CI 1.20-1.68), heart failure (OR 1.24; 95% CI 1.05-1.45), and at least one antibiotic-resistant pathogen (OR 2.77; 95% CI 2.27-3.38) were associated with an increased risk of having poor value-outcomes.

CONCLUSIONS: In patients with sepsis, the presence of common chronic diseases, antibiotic resistant pathogens, and malignancy were associated with incurring higher hospital charges with worse outcomes compared to those without. Future steps include identifying specific hospital resources incurred in those with high-cost sepsis admissions.