Reducing Readmission for Sepsis by Improving Risk Prediction Algorithms

Abstract

Unplanned readmissions after sepsis, rates of which range from 17.5% to 32%, pose substantial challenges for health care systems. Associated costs for sepsis surpass those for other critical conditions. Existing readmission risk models rely primarily on clinical indicators, which limits their predictive accuracy for patients with sepsis. This review explores how integrating social determinants of health into readmission models can enhance model precision and applicability for predicting 30-day readmission among sepsis survivors. Although socioeconomic status, neighborhood deprivation, and access to health care are known to influence postdischarge outcomes, these social determinants of health are underused in current risk algorithms. Evidence shows that incorporating social determinants of health into predictive models significantly improves model performance. Furthermore, failure to account for health disparities driven by social determinants of health in high-risk populations can exacerbate existing inequities in health care outcomes. The integration of social determinants of health into sepsis readmission risk models offers a promising avenue for improving prediction accuracy, reducing readmissions, and optimizing care for vulnerable populations. Future research should focus on refining these models and exploring postdischarge monitoring strategies to further mitigate the burden of sepsis readmissions.