303 Burden of major hospital-onset device-associated infection types among adults and children in the United States, 2007

Saturday, April 2, 2011: 2:15 PM
Cortez Ballroom (Hilton Anatole)
Matthew E. Wise, PhD , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
R. Douglas Scott II, PhD , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
Katherine D. Ellingson, PhD , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
Jonathan R. Edwards, MStat , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
Teresa C. Horan, MPH , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
James Baggs, PhD , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
Scott K. Fridkin, MD , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
John Jernigan, MD, MS , Division of Healthcare Quality Promotion, CDC, Atlanta, GA
Background: Device-associated infections are a major target of the Health and Human Services’ Action Plan to Prevent Healthcare-Associated Infections (HAI), yet no systematic methods have been applied to contemporary data to estimate the burden of these infections in the US.  Due to declines in the incidence of these infections observed in national surveillance data, current burden estimates will be critical for ongoing policy efforts.

Objective: To estimate the national burden of hospital-onset central line-associated bloodstream infections (CLABSI), ventilator-associated pneumonias (VAP), and catheter-associated symptomatic urinary tract infections (SUTI) among adults and children in inpatient wards and critical care units and quantify the degree of uncertainty around these estimates.

Methods: We estimated the total number of critical care-days and inpatient ward-days in the US in 2007 using information from publicly available healthcare utilization data sources.  We applied device utilization ratios from the National Healthcare Safety Network (NHSN) to patient-day estimates to determine total device-days and then applied CLABSI, VAP, and SUTI rates from NHSN to device-day estimates to calculate the total number of infections nationally.  Using Monte Carlo methods, we performed the calculations 99,999 times, randomly sampling one value for each of the inputs from pre-specified distributions.  NHSN locations reporting >6 months of data during 2006-2008 that  had device utilization ratios or infection rates between the 25th and 75th percentile for the applicable location type were eligible for inclusion in input distributions.  The 2.5th and 97.5th percentile for the calculated total number of infections defined a 95% credible interval for burden estimates. 

Results: We estimated 176 million inpatient-days among adults and children in the US in 2007: 22 million (13%) in critical care units and 154 million (87%) in inpatient wards.  Overall we estimated 28 million central line-days, 9 million ventilator-days, and 44 million urinary catheter-days.  Of these, 12 million central line-days, 8 million ventilator-days, and 17 million urinary catheter-days occurred in critical care units.  In 2007 the 95% credible intervals for burden were 9,000-64,000 CLABSI, 7,000-37,000 VAP, and 39,000-139,000 SUTI; the interval for all three infection types combined was 79,000-200,000 infections.  Approximately 63% of CLABSI, >99% of VAP, and 36% of SUTI occurred in critical care units.

Conclusions: Hospital-onset device-associated infections remain an important cause of morbidity among adults and children in the United States.  Results of these simulations are largely consistent with published estimates of the overall burden of all HAI in 2002 given recent declines in the incidence of device-associated infections and accounting for the proportion of HAI that are device-associated.