Longitudinal Epidmiology of Clostridium difficile among Patients in the Veterans Affairs (VA) Healthcare System
Background: Clostridium difficile infection (CDI) is an important contributor to healthcare expenditures. Previous research has focused on primary CDI in the acute care setting. More comprehensive assessments of CDI burden are needed that account for multiple episodes, as well as episodes across inpatient and outpatient settings.
Methods: This was an open retrospective cohort study between 2005 and 2012 using data from the US Veterans Affairs (VA) system, including patients during years when they had at least one in-person healthcare encounter. Patients with a history of CDI between 2000 and 2005 were excluded. CDI episodes were classified according to the Center for Disease Control and Prevention's LabID metric in all inpatient settings. Patients were followed for the development of non-primary CDIs through the end of the study period. Incidence rates were calculated, and Poisson regression was used to evaluate trends in rates over time.
Results: During the study period, there were 78,493 CDI episodes from 58,879 unique patients, with an observed rate of 1.53 episodes per 1,000 Veteran-years at-risk. There was a significant decrease in the overall and HCFO-specific rates over time (p<.0001 for both). Conversely, the rate of CO cases increased from 0.99 to 1.08 per 1,000 Veteran-years from 2005 to 2012 (Figure 1). Approximately 78% of patients had 1 CDI episode, 13% of patients experienced 2 episodes, and 9% of patients had 3 or more (Max: 18). Among patients who developed non-primary CDI, the median time between episodes was 38 (IQR: 59) days, and more than 20% of non-primary cases occurred greater than 6 months after the primary episode. Nearly 60% of both primary and non-primary episodes were community-onset CDI.
Conclusion: CDI rates decreased overall due to decreasing HCFO-CDI rates and despite increasing CO CDI rates. In addition, more than 20% of patients experience multiple episodes. One-fifth of patients developing non-primary CDI did so greater than 6 months after their primary episode. Estimates of the economic burden of CDI should take into account both primary and non-primary cases that occur across the spectrum of healthcare delivery systems.
M. Jones, None
R. E. Nelson, Roche: Consultant and Grant Investigator, Consulting fee, Research grant and Research support
M. Rubin, None
M. Samore, None
See more of: Oral Abstract Session