Using Electronic Health Information to Risk-Stratify Rates of Clostridium difficile Infection in US Hospitals
| Autor: | Ying P. Tabak, Richard S. Johannes, Xiaowu Sun, L. Clifford McDonald, Marya D. Zilberberg, Dawn M. Sievert, Karen G. Derby |
|---|---|
| Rok vydání: | 2011 |
| Předmět: |
Adult
Microbiology (medical) medicine.medical_specialty Epidemiology Hospitalized patients Toxin assay symbols.namesake Risk Factors Internal medicine medicine Electronic Health Records Humans Poisson Distribution Poisson regression Intensive care medicine Disease Notification Aged Cross Infection Clostridioides difficile business.industry Incidence (epidemiology) Middle Aged Clostridium difficile Hospitals United States Community-Acquired Infections Infectious Diseases Population Surveillance Multivariate Analysis Risk stratification Clostridium Infections symbols Electronic data Health information business |
| Zdroj: | Infection Control & Hospital Epidemiology. 32:649-655 |
| ISSN: | 1559-6834 0899-823X 0195-9417 |
| Popis: | Background.Expanding hospitalized patients' risk stratification for Clostridium difficile infection (CDI) is important for improving patient safety. We applied definitions for hospital-onset (HO) and community-onset (CO) CDI to electronic data from 85 hospitals between January 2007 and June 2008 to identify factors associated with higher HO CDI rates.Methods.Nonrecurrent CDI cases were identified among adult (≥18-year-old) inpatients by a positive C. difficile toxin assay result more than 8 weeks after any previous positive result. Case categories included HO, CO-hospital associated (CO-HA), CO-indeterminate hospital association (CO-IN), and CO–non–hospital associated (CO-NHA). C. difficile testing intensity (CDTI) was defined as the total number of C. difficile tests performed, normalized to the number of patients with at least 1 C. difficile toxin test recorded. We calculated both the incidence density and the prevalence of CDI where appropriate. We fitted a multivariable Poisson model to identify factors associated with higher HO CDI rates.Results.Among 1,351,156 unique patients with 2,022,213 admissions, 9,803 cases of CDI were identified; of these, 50.6% were HO, 17.4% were CO-HA, 9.0% were CO-IN, and 23.0% were CO-NHA. The incidence density of HO was 6.3 per 10,000 patient-days. The prevalence of CO CDI on admission was, per 10,000 admissions, 8.4 for CO-HA, 4.4 for CO-IN, and 11.1 for CO-NHA. Factors associated (P< .0001) with higher HO CDI rates included older age, higher CO-NHA prevalence on admission, and increased CDTI.Conclusion.Electronic health information can be leveraged to risk-stratify HO CDI rates by patient age and CO-NHA prevalence on admission. Hospitals should optimize diagnostic testing to improve patient care and measured CDI rates. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|