This presentation is part of: Poster Session - Utility Theory; Health Economics; Patient & Physician Preferences; Simulation; Technology Assessment
ACCEPTABLE FAILURE RATES FOR ANTIBIOTIC THERAPY OF CENTRAL VENOUS CATHETER (CVC) ASSOCIATED BACTEREMIA
Eli N. Perencevich, MD, MS1, Keith S. Kaye, MD, MPH2, Larry J. Strausbaugh, MD3, Douglas D. Bradham, DrPH1, David N. Fisman, MD, MPH4, and Anthony D, Harris, MD, MPH1. (1) VA Maryland Health Care System and University of Maryland School of Medicine, Baltimore, Department of Epidemiology and Preventive Medicine, Baltimore, MD, (2) Duke University Medical Center, Department of Medicine, Durham, NC, (3) Portland VA Medical Center, Infectious Disease Section, Medical Service, Portland, OR, (4) Drexel University, School of Public Health, Epidemiology and Biostatistics, Philadelphia, PA
Purpose: Antimicrobial resistance is an increasing public health problem. Decreasing the duration of antibiotic therapy is an attractive strategy for delaying emergence of antimicrobial resistance. The paucity of data about optimal treatment durations for most clinical infections hinders the adoption of this approach and encourages overtreatment. Using a case scenario, this study used the methodology developed for contingent valuation analysis to identify acceptable failure rates of infectious diseases consultants (IDCs) for treatment of CVC-associated bacteremia. Quantifying an acceptable failure rate may help reduce excess treatment duration and improve numerical statements in physician risk communication. Methods: The authors developed a case scenario for a representative patient who developed uncomplicated CVC-associated coagulase-negative staphylococci bacteremia. The patient received standard-of-care therapy including catheter replacement and intravenous vancomycin therapy. In August 2003, the Infectious Diseases Society of America’s Emerging Infections Network (EIN) distributed the case description to its members. The EIN gave individual members one of 10 failure rates and asked if they would accept or reject the given value. The authors used logistic regression to evaluate the relationship between specific failure rates offered to EIN members and their willingness to accept them. Results: Overall, 374 (54%) of 687 EIN members responded to the questionnaire. Mean years in practice for responders was 15.9 years. Logistic regression analysis determined that the median acceptable failure rate was 6.8%. Thus, half the IDCs would have found a failure rate of 6.8% to be acceptable. 75% of IDCs would have found a failure rate of 1.6% acceptable and 25% of IDCs would have found a failure rate as high as 11.9% acceptable. Adjusting for IDCs years in practice did not alter these results. As expected, the acceptable failure rate and ranges in this study were all lower than those reported in a previous study using the same methodology in diabetic foot osteomyelitis (Perencevich, et al., Clin Infect Dis Feb 15, 2004) Conclusions: The quantified physician acceptable failure rates when combined with patient preferences to avoid treatment failure and duration-specific therapy failure rates might assist in selecting an optimal treatment duration. Thus, the methodology used may prove useful in delineating acceptable treatment failure thresholds, a first step in reducing durations of antimicrobial therapy. In addition, the numerical failure rates may help improve physician risk communication.
