LB 9 Impact of Universal Admission Screening on Nosocomial Rates of Methicillin-Resistant Staphylococcus aureus (MRSA)

Saturday, March 20, 2010
Grand Hall (Hyatt Regency Atlanta)
Virginia Roth, MD , The Ottawa Hospital, Ottawa, ON, Canada
Natalie Oake, MSc , The Ottawa Hospital, Ottawa, ON, Canada
Kathryn Suh, MD, MSc , The Ottawa Hospital, Ottawa, ON, Canada
Monica Taljaard, PhD , The Ottawa Hospital, Ottawa, ON, Canada
Karam Ramotar, PhD , The Ottawa Hospital, Ottawa, ON, Canada
Natalie Bruce , The Ottawa Hospital, Ottawa, ON, Canada
Baldwin Toye, MD , The Ottawa Hospital, Ottawa, ON, Canada
Alan Forster, MD , The Ottawa Hospital, Ottawa, ON, Canada
Background: Healthcare and community-associated MRSA infections are increasing in many countries. MRSA detection at hospital admission allows for patient isolation and may prevent further spread from asymptomatic carriers.

Objective: To determine if the nosocomial incidence of MRSA is lower after implementation of universal admission screening.

Methods: Pre-intervention (Jan 2006 – Dec 2007), patients were screened for MRSA on admission based on risk factors (hospitalization in past 12 months, direct transfers, known MRSA). Post-intervention (Jan 2008 – Jun 2009), swabs of the anterior nares, rectum, insertion sites and open wounds were to be obtained from all patients admitted to our 1200-bed tertiary care hospital at the time of admission. Real-time PCR was used for MRSA detection; PCR positive specimens were confirmed by culture. PCR positive patients were placed on contact precautions. Systematic decolonization was not done. Nosocomial MRSA incidence pre- and post-intervention was compared using interrupted time series analysis.

Results: There was a significant increase in new MRSA carriers detected on admission post-intervention vs pre-intervention (2.2 vs 1.0 per 1,000 admissions; p-value<0.001), and a higher proportion were considered community-acquired (81 vs 72.6%; OR=1.6, 95%CI 1.0-2.6; p-value=0.03). The compliance rate for admission screening post-intervention was 83% vs 60% pre-intervention. Despite the increase in detected cases at admission, the post-intervention nosocomial MRSA incidence was not significantly different compared to pre-intervention (47.3 vs 44.5/100,000 pt days; incidence RR = 1.06, 95% CI 0.91-1.24, p-value=0.4) (Figure).

Conclusions: Universal admission screening detects significantly more MRSA carriers than risk factor-based screening and may result in improved compliance as it is straightforward and easy to audit. However, this intervention did not reduce nosocomial MRSA incidence in our hospital. Possible explanations include low healthcare worker compliance with hand hygiene and contact precautions, inadequate facilities, and imperfect admission screening compliance. As the detection rate in our study was low, universal screening may be more effective in settings with higher MRSA prevalence.