Methods: Using the Nationwide Inpatient Sample database, we identified adult patients (¡Ý 18 years) with CDI by ICD-9-CM codes. The trends of CDI incidence, mortality and hospital charges were evaluated by Poisson regression. The risk for HOCDI and factors to predict in-hospital death of CDI patients were analyzed by logistic regression.
Results: 3,337,910 cases of CDI were identified out of a total of 318,703,355 hospitalizations (1.05%). Incidences of non-HOCDI and HOCDI were 0.42% and 0.63% respectively. In the 10-year study period, CDI incidence increased with an annual rate of 3.3% (P<0.001). The annual incidences of HOCDI and non-HOCDI increased with a rate of 1.4% and 2.0% respectively (P<0.001). After adjusting for demographics, length of hospital stay and Charlson index, transfer from long-term health facilities (OR=2.02, 95% CI 1.83-2.23) and admission to a teaching hospital (OR=1.10, 95% CI 1.05-1.15) were independent risk factors for HOCDI. The in-hospital mortality of CDI associated hospitalization decreased from 9.7% in 2005 to 6.8% in 2014 (P<0.001). Transfer from long-term health facilities significantly predicted the risk for in-hospital death in CDI patients (OR= 1.34, 95% CI 1.32-1.36). The sum charge of all CDI hospitalizations increased with an annual rate of 2.0% (P<0.001). The median charge per CDI hospitalization increased during 2005-2009 (P<0.001), and then decreased during 2010-2014 (P<0.001).
Conclusion: During 2005-2014, the mortality in CDI hospitalized patients decreased, but CDI incidence in acute care hospitals increased, resulting in increased total CDI associated hospital charges. Patients transferred from long-term healthcare facilities increased the risk for HOCDI and CDI associated in-hospital mortality. They should be considered as high risk patients for CDI surveillance when developing mitigation strategies.
J. Weinberg, None
See more of: Poster Abstract Session