Cluster Randomized Trial of Hospitals to Assess Impact of Targeted Versus Universal Strategies to Reduce Methicillin-resistant Staphylococcus Aureus (MRSA) in Intensive Care Units (ICUs) (REDUCE-MRSA)

• Main Outcome: Patients With Nosocomial MRSA Clinical Cultures [ Time Frame: The 30-month time frame represents 12-month baseline and 18-month intervention periods. During these time periods, outcomes are defined as events occurring during attributed ICU time: from day 3 of the ICU stay until 2 days after ICU discharge. ]
  Hazard ratio for ICU-attributable MRSA+ clinical cultures comparing Baseline to Intervention period, by Arm, accounting for clustering by hospital.
  
  

• MRSA Bloodstream Infection [ Time Frame: The 30-month time frame represents 12-month baseline and 18-month intervention periods. During these time periods, outcomes are defined as events occurring during attributed ICU time: from day 3 of the ICU stay until 2 days after ICU discharge. ]
  Hazard ratio for ICU-attributable MRSA+ blood cultures comparing Baseline to Intervention period, by Arm, accounting for clustering by hospital.
  
  
• ICU-attributable All-pathogen Bloodstream Infection [ Time Frame: The 30-month time frame represents 12-month baseline and 18-month intervention periods. During these time periods, outcomes are defined as events occurring during attributed ICU time: from day 3 of the ICU stay until 2 days after ICU discharge. ]
  Hazard ratio for ICU-attributable positive blood culture from any pathogen, comparing Baseline to Intervention period, by Arm, accounting for clustering by hospital.
  
  
• Intervention Impact on Healthcare Costs [ Time Frame: 12-month period ]
  Costs (in dollars) per 1000 ICU-admissions associated with 3 ICU strategies to reduce ICU Bloodstream infection (BSI), (Arms 1-3).
  
  
• Blood Culture Contamination Rates [ Time Frame: 24-month time frame for this analysis represents a 6-month baseline and 18-month intervention period. ]
  Odds ratio for ICU-attributable blood culture contamination rates, comparing Baseline to Intervention period across Arms, accounting for clustering by hospital.
  
  
• Intervention Impact on Bacteriuria and Candiduria [ Time Frame: 30-month time frame represents 12-month baseline and 18-month intervention periods. ]
  Proportional hazard ratio for as-randomized, unadjusted, ICU-attributable bacteriuria, comparing Baseline to Intervention period across Arms, accounting for clustering by hospital. High-level bacteriuria is defined as ≥50,000 CFU/mL, high-level candiduria is defined as ≥50,000 CFU/mL.
  
  
• Intervention Impact on Mupirocin Susceptibility of MRSA Isolates [ Time Frame: 25-month time frame represents 7-month baseline and 18-month intervention periods ]
  Odds ratio for MRSA+ isolates from ICU patients expressing low-level mupirocin resistance (LLMR) and high-level mupirocin resistance (HLMR), comparing baseline to intervention period across arms, accounting for clustering by hospital.
  
  
• Intervention Impact on Chlorhexidine Susceptibility of MRSA Isolates [ Time Frame: 25-month time frame represents 7-month baseline and 18-month intervention periods ]
  Frequency of MRSA+ isolates from ICU patients with reduced susceptibility to chlorhexidine (CHG) (MIC >4 μg/ml), comparing baseline to intervention period across arms, accounting for clustering by hospital.
  
  

Baseline data involving 12 months of data for participating hospitals (July 2008 - June 2009) was collected prior to randomization to account for size and ICU baseline prevalence of MRSA in randomization scheme. Randomization occurred at the hospital level.

Eligibility survey was conducted to determine exclusion criteria.

As of May 2010, enrollment has been closed. 45 hospitals were randomized, but two were found to meet exclusion criteria and were excluded. As-randomized (or as-assigned) analysis included 43 hospitals, representing 74 ICUs. Individual (patient-level) subject enrollment during intervention is 74,256.