Modeling antibiotic resistance to project future rates: quinolone resistance in Escherichia coli
- First Online:
- 238 Downloads
While the development of resistance to a new antibiotic is expected, the time course and degree of resistance that will develop are uncertain. Some best projections of the future extent of resistance can be highly impactful for activities, such as antimicrobial development, that require significant lead time. We focus on the surge among hospital isolates in fluoroquinolone-resistant Escherichia coli and use data on resistance and consumption to explore and quantify trends in increasing resistance and their relationship to antibiotic use from 2001 to 2007. A mixed-effects logistic regression model produced a good fit to the observed resistance rates during this period in the United States and Europe. The model contained significant effects of time, consumption, and country on developing fluoroquinolone resistance in E. coli. There was a larger projected increase in resistance for high fluoroquinolone-consuming countries projected to 2013: 45% (95% confidence interval [CI]: 38%, 53%) for high consumers vs. 33% (95% CI: 25%, 41%) for low consumers. The model was also used to obtain regional projections of resistance that can be used by local prescribers. In order to better understand and predict trends in antimicrobial resistance, it is vital to implement and expand current surveillance systems.
- 3.The European Antimicrobial Resistance Surveillance System (2007) EARSS Annual Report 2007. Available online at: http://www.rivm.nl/earss/Images/EARSS%202007_FINAL_tcm61-55933.pdf
- 7.Bates DM (2008) Linear mixed model implementation in lme4. Department of Statistics, University of Wisconsin, Madison, WIGoogle Scholar
- 8.SAS Institute Inc. (2003) SAS OnlineDoc 9.1.2. SAS Institute Inc., Cary, NC. Available online at: http://support.sas.com/onlinedoc/912/docMainpage.jsp