Predicting the risk of C. difficile infection to improve fluoroquinolone use

Kaiser Foundation Research Institute, Oakland CA

Investigators

Abstract

PROJECT SUMMARY/ABSTRACT: Clostridium difficile infections (CDIs) are associated with significant and debilitating morbidity and can result in severe outcomes such as surgical intervention or death. The primary modifiable risk factor for CDI is antimicrobial use. The proposed research will focus on the risk of infection following the use of fluoroquinolones, an antimicrobial class associated with CDI outbreaks. We propose to develop a risk score for the prediction of CDI among patients who receive a fluoroquinolone prescription in the outpatient setting. Our long-term goal is to enhance the safe use of fluoroquinolone agents in the outpatient setting by providing risk information to clinicians at the time of prescribing through alerts embedded in the electronic medical record. These alerts will provide patient-level prediction of CDI risk to aid clinician decision-making when a fluoroquinolone prescription is being considered. The incorporation of this risk score into an alert will offer a more accurate risk management strategy than asking the clinician to synthesize a patient's individual predictors of CDI on an ad hoc basis. The objectives of this application are to quantify the risk for CDI following fluoroquinolone use and to develop a risk score to identify patients at high-risk patient for CDI. To accomplish these objectives, we propose the following specific aims: (1) Describe the epidemiology of Clostridium difficile infections in patients who filled a fluoroquinolone prescription in the outpatient setting;and, (2) Develop a risk score to predict Clostridium difficile infection among a cohort of patients who are prescribed fluoroquinolones in the outpatient setting using patient demographic and clinical characteristics available through automated means from the electronic medical record. To accomplish these aims, we will identify a retrospective inception cohort of patients prescribed a fluoroquinolone in the outpatient setting and, among this cohort, describe the epidemiology of CDI in the six months following fluoroquinolone use. A CDI risk score will then be developed from patient characteristics to predict the risk of CDI among this cohort and to categorize patients by their predicted risk for CDI. This project has potential to significantly reduce the significant mortality and morbidity associated with CDI infection by improving prescribing practices of fluoroquinolones. Our preliminary data show that 29% of CDIs occurred subsequent to fluoroquinolone use. Eventually, if this risk score can be successfully used by clinicians to prescribe fluoroquinolones to patients at lower risk for CDI only, we plan to adapt this research to develop risk scores based on EMR data for other adverse outcomes.