Clostridium difficile infection (CDI) remains a major cause of health care–associated diarrhea in industrialized countries and is a common target of antimicrobial stewardship programs (ASPs).
While the burden of CDI has been well described in tertiary metropolitan hospitals, there is a lack of published evidence from regional and rural hospitals. Our recent study published in the Journal of Hospital Infection explores the effect of an ASP on health care–associated CDI rates and the impact of CDI on length of stay and hospital costs.
Stuart Bond
The ASP functioned alongside infection control practices, including isolation of patients with antimicrobial-resistant organisms (including C. difficile), hand hygiene, personal protective equipment, and terminal cleaning. Timely feedback emails to medical officers contained information on patient-specific risk factors for CDI, current and prior antimicrobials, and suggestions for CDI treatment. The effect of health care–associated CDI on length of stay and hospital costs was investigated using a group of matched controls, identified retrospectively using hospital performance data. Prior antimicrobial and proton pump inhibitor use were also measured and compared with background use.
The results of our study demonstrated a stable health care–associated CDI rate of around four cases per 10,000 occupied bed days over a 5-year period, similar to the average Australian rate. The length of time over which CDI rates could be effectively examined prior to the intervention was limited by changes to C. difficile stool testing methods. Median length of stay was 11 days greater, and median hospital costs were AU$11,361 higher for patients with health care–associated CDI (n = 91) than for their matched controls (n = 172). It is likely that the increase in costs was associated with additional length of stay but also with increased investigation and treatment costs. Among the group of patients with severe disease (n = 8), only four received oral vancomycin according to Australian guidelines, possibly because of under-recognition of severity criteria. The response rate to emails was low at 19%, showing that other methods are additionally necessary to communicate CDI case feedback.
Third generation cephalosporins and beta-lactamase inhibitor combinations were over-represented in the health care–associated CDI group, where narrower spectrum antimicrobials such as beta-lactamase sensitive penicillins were under represented. Rates of prior antimicrobial use and proton pump inhibitor use were broadly in agreement with the literature.
Our study demonstrated that, in the Australian nonmetropolitan setting, there was a high burden of health care–associated CDI in terms of hospital costs and length of stay, even though our health district experienced CDI rates that were similar to the Australian average. Challenges associated with the study included maintenance of consistent data collection across multiple hospital sites without comprehensive electronic medical records, provision of timely email feedback, and dissemination of study results.
Analysis of prior antimicrobial use has allowed us to identify targets for ongoing antimicrobial stewardship activities, and we also intend to provide further education on recognition of severity criteria. These activities can be supported through daily antimicrobial stewardship ward rounds. Future research could involve application of appropriateness criteria to prior and current antimicrobial use in CDI patients in order to identify avoidable cases and use of more advanced statistical techniques such as multistate modeling to determine differences in outcomes between cases and controls.
Stuart Bond, BPharm, DipPharmPrac, is an antimicrobial stewardship pharmacist based at Wollongong Hospital in New South Wales, Australia.