Main Menu Browse by Day or Program Author Keywords

300 Analysis of MRSA Transmission Dynamics in the Veterans Affairs System

Saturday, April 2, 2011: 3:30 PM
Coronado A (Hilton Anatole)
Makoto Jones, MD , University of Utah, Salt Lake City, UT
Jian Ying, PhD , University of Utah, Salt Lake City, UT
Tom Greene, PhD , University of Utah School of Medicine, Salt Lake City, UT
Molly Leecaster, PhD , University of Utah, Salt Lake City, UT
Benedikt Huttner, MD , University Hospital, Geneva, Geneva, Switzerland
Jeanmarie Mayer, MD , University of Utah School of Medicine, Salt Lake City, UT
Michael Rubin, MD, PhD , VA Salt Lake City IDEAS Center, Salt Lake City, UT
Matthew Samore, MD , VA Salt Lake City IDEAS Center, Salt Lake City, UT
Background: Few large multi-center studies have measured methicillin-resistant Staphylococcus aureus (MRSA) acquisition and healthcare-associated infection (HAI) rates. The 2007 Veterans Affairs (VA) MRSA Bundle that included active surveillance has provided data from a large number of hospitals to analyze MRSA transmission dynamics.

Objective: Determine the variability of MRSA acquisition and HAI incidence surrogates between hospitals. Measure the correlation between the two measures. Determine the overall temporal trends, 

Methods: We used patient census, active MRSA surveillance, and microbiology data from 112 acute-care VA sites from October 2007 through December 2009 for analysis. Only sites with data indicating at least 100 acute-care patient days per month were included. Acquisitions were defined as positive MRSA surveillance or clinical tests following negative surveillance tests, or in the absence of a surveillance test, the absence of a positive clinical cultures in the past year. Incident Nosocomial Clinical Cultures (INCC) were defined as the first positive clinical culture more than 48 hours after admission without prior positives in the last year. Both acquisition and INCC rates were denominated by patient-days at risk for the event and calculated at each hospital separately for ICU and non-ICU settings. The importation rate represents the fraction of MRSA positive admissions to the total number of admissions. Models analyzing INCC were calculated with clinical culture data only. Because rates from small hospitals may deviate from the overall mean purely by small numbers and chance, generalized linear mixed models were used to calculate “shrunken estimates.” To assess temporal trends, generalized estimating equations were used to model the effect of time on acquisitions and INCC as outcomes, while accounting for importation prevalence. Correlation between total acquisitions and INCC at each hospital were measured using Pearson’s correlation coefficient.

Results: A significant decline was seen in both importation-adjusted acquisition and INCC rates. The median model-derived acquisition rate estimate in the ICUs was .034 (IQR .029-.041), while in the non-ICU setting it was .027 (IQR .02-.037). The median INCC rate in ICUs was .034 (IQR .032-.039), while in the non-ICU setting, it was .015 (IQR .012-.018). The correlations between acquisitions and INCCs were .35 in non-ICU and .31 in ICU settings.

Conclusions: Overall MRSA acquisition and HAI trends appear to be decreasing, but there is appreciable variation between hospitals and settings. Variability after accounting for importation suggests that the MRSA transmission rate varies between hospitals. The correlation between acquisition rates and INCC rates is at least consistent with a central role of transmission in MRSA-HAI rates. Further analysis will attempt to determine individual- and ecological-level etiologies of the decline.