Burden Of Bloodstream Infection Caused By Extended-Spectrum Beta-Lactamase-Producing Enterobacteriaceae Determined Using Multistate Modeling At A Swiss University Hospital And A Nationwide Predictive Model

OBJECTIVE. To obtain an unbiased estimate of the excess hospital length of stay (LOS) and cost attributable to extended-spectrum beta-lactamase (ESBL) positivity in bloodstream infections (BSIs) due to Enterobacteriaceae.DESIGN. Retrospective cohort study.SETTING. A 2,200-bed academic medical center in Geneva, Switzerland.PATIENTS. Patients admitted during 2009.METHODS. We used multistate modeling and Cox proportional hazards models to determine the excess LOS and adjusted end-of-LOS hazard ratio (HR) for ESBL-positive and ESBL-negative BSI. We estimated economic burden as the product of excess LOS and average bed-day cost. Patient-level accounting data provided a complementary analysis of economic burden. A predictive model was fitted to national surveillance data.RESULTS. Thirty ESBL-positive and 96 ESBL-negative BSI cases were included. The excess LOS attributable to ESBL-positive and ESBL-negative BSI was 9.4 (95% confidence interval [CI], 0.4-18.4) and 2.6 (95% CI, 0.7-5.9) days, respectively. ESBL positivity was therefore associated with 6.8 excess days and CHF 9,473 per BSI. The adjusted end-of-LOS HRs for ESBL-positive and ESBL-negative BSI were 0.62 (95% CI, 0.43-0.89) and 0.90 (95% CI, 0.74-1.10), respectively. After reimbursement, the average financial loss per acute care episode in ESBL-positive BSI, ESBL-negative BSI, and control cohorts was CHF 48,674, 48,131, and 13,532, respectively. Our predictive model estimated that the nationwide cost of third-generation cephalosporin resistance would increase from CHF 2,084,000 in 2010 to CHF 3,526,000 in 2015.CONCLUSIONS. This is the first hospital-wide analysis of excess LOS attributable to ESBL positivity determined using multistate modeling to avoid time-dependent bias. These results may inform health-economic evaluations of interventions targeting ESBL control. Infect Control Hosp Epidemiol 2013; 34(2): 133-143