Deconstructing the Veterans Affairs MRSA prevention bundle.

During the 1970s and 1980s, methicillinresistant Staphylococcus aureus (MRSA) became endemic in hospitals worldwide [1]. By 2007, a large-scale epidemiological study determined that 94 000 cases of invasive MRSA disease and 19 000 MRSA-related in-hospital deaths occurred annually in the United States [2]. Using these estimates, an editorial suggested that the burden of MRSA deaths exceeded that of deaths due to human immunodeficiency virus or AIDS in the United States [3]. TheoptimalapproachtoMRSAprevention in hospital settings is debated [4, 5]. In general, 2 epidemiological approaches have been described: (1) “horizontal” programs that broadly attempt to reduce infections from all pathogens and (2) “vertical” programs that aim to eradicate a single pathogen [6]. Horizontal programs include hand hygiene improvement interventions and centralline association bloodstream infection (CLABSI) “bundles.” Vertical programs screen all or selected patients for MRSA colonization by obtaining a nares swab sample and place MRSA-colonized or MRSA-infected patients on contact precautions. Contact precautions include single-bed rooms and patient cohorting and require healthcare workers to wear gowns and gloves for all contacts. Some studies reported benefits of such screening programs [7], some reported benefits when screening was combined with decolonization [8], and some found screening and isolation to be ineffective [9, 10]. After a successful pilot study in the VA Pittsburgh Healthcare System, in August 2006 the Veterans Health Administration implemented a MRSA bundle in 17 US Department of Veterans Affairs (VA) medical centers [11]. The bundle included (1) admission, inhospital transfer, and discharge activesurveillance swab samples for MRSA; (2) contact precautions for patients known or found to be MRSA carriers; (3) efforts targeting improved hand hygiene; and (4) efforts encouraging culture change [12]. Importantly, and infrequently, mentioned in descriptions of the VA MRSA Initiative, support was also made available for a MRSA prevention coordinator at each site, who was responsible for coordinating implementation of the bundle. Thus, the VA bundle included 1 vertical intervention (swab samples and contact precautions) and 3 horizontal interventions (hand hygiene, culture change, and MRSA prevention coordinator). By October 2007 the bundle was implemented in all acute-care (except mental health) units in 150 of 153 VA medical centers. By most measures the VA MRSA Initiative was a success. From initiation to June 2010, healthcare-associated MRSA infections declined by 62% in intensive care units (ICUs) and by 45% in nonICUs. MRSA acquisition declined 17% in ICUs and 21% in non-ICU settings. Interestingly, a subset of hospitals reported healthcare-associated vancomycin-resistant enterococcal and Clostridium difficile infection rates. After the initiative, vancomycin-resistant enterococcal infections were eliminated in ICUs and declined by 73% in non-ICUs. C. difficile infections declined 61% in non-ICU settings, with no change in ICUs [12]. How do we explain the MRSA declines reported by the VA? A nationwide decline in MRSA seen during the study period [13] might have contributed to these in-hospital declines; however, MRSA positivity on admission was relatively flat during the study period, so any outside impact should have been negligible. Importantly, infections declined 2–3 times more than incident colonization, suggesting that factors other than surveillance swab samples and isolation played a role in reduced MRSA infections. Thus, it seems that one benefit of the bundle was to reduce Received 22 February 2012; accepted 24 February 2012; electronically published 4 April 2012. Correspondence: Eli N. Perencevich, MD, MS, CADRE, Iowa City VA Medical Center, 601 Hwy 6 W, Iowa City, IA 52246 ([email protected]). Clinical Infectious Diseases 2012;54(11):1621–3 © The Author 2012. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals. [email protected]. DOI: 10.1093/cid/cis277