1214. Hospital transfer network structure as a risk factor for C. difficile infections
Session: Oral Abstract Session: New Insights into C. difficile Transmission and Reporting
Saturday, October 5, 2013: 8:45 AM
Room: The Moscone Center: 200-212
Background: Healthcare associated Clostridium difficile infection (CDI) is an important cause of morbidity. CDI rates are increasing and CDI is now one of the most common healthcare associated infections. The environment is an important risk factor for CDI, but not all cases can be attributed to the hospital where they are diagnosed. Patients with CDI could have acquired the disease from exposures in the hospital where the diagnosis was made, the community or another healthcare facility. We hypothesize that how a hospital is connected via transfers to other hospitals is associated with the hospital’s CDI rate.

Methods: Using the Healthcare Costs and Utilization Project State Inpatient Database for Arizona for 2003 to 2007, we identified transfers as two separate hospitalizations with a 0 day interval. Of the 3,496,258 stays, 76,485 transfers were identified. We calculated the number of source hospitals (in-degree) for each of the 88 hospitals in Arizona. For the analysis, we aggregated CDI incidence and covariates to the hospital level and restricted the sample to the 46 general medical and surgical hospitals with data for each of the five years under study. Using generalized estimating equations, we regressed the log CDI on the fraction of patients aged >= 65, the median length of stay, an indicator of size >= 200 beds, teaching status and log in-degree. The in-degree of a hospital is the number of unique hospitals from which transfers originate.

Results: We find increasing the log in-degree by 1 is associated with a 0.19 unit (p = 0.0007) increase in the log CDI. The mean in-degree is 9.3 (inter quartile range: 2 – 12). Setting the covariates at the mean or modal values, we estimate an increase of in-degree from 9 to 10 is associated with a 0.011 cases per 100 stays increase in incidence. The mean incidence was estimated as 0.93 cases per 100 stays (IQR: 0.35 – 0.94).

Conclusion: We find that an increase in the number of hospitals from which transfers are accepted is associated with increased incidence of CDI adjusting for bed size and patient mix. These findings suggest that a network based approach to studying and designing CDI interventions are needed.

Jacob Simmering1, David Campbell2, Linnea Polgreen, PhD1 and Philip M. Polgreen, MD3, (1)Pharmacy Practice & Science, University of Iowa College of Pharmacy, Iowa City, IA, (2)Computer Science, University of Iowa College of Liberal Arts and Sciences, Iowa City, IA, (3)Division of Infectious Diseases, Department of Internal Medicine, University of Iowa Carver College of Medicine, Iowa City, IA


J. Simmering, None

D. Campbell, None

L. Polgreen, None

P. M. Polgreen, None

Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.