Electronic health record-based detection of risk factors for Clostridium difficile infection relapse.

OBJECTIVE: A major challenge in treating Clostridium difficile infection (CDI) is relapse. Many new therapies are being developed to help prevent this outcome. We sought to establish risk factors for relapse and determine whether fields available in an electronic health record (EHR) could be used to identify high-risk patients for targeted relapse prevention strategies.
DESIGN: Retrospective cohort study.
SETTING: Large clinical data warehouse at a 4-hospital healthcare organization. PARTICIPANTS: Data were gathered from January 2006 through October 2010. Subjects were all inpatient episodes of a positive C. difficile test where patients were available for 56 days of follow-up.
METHODS: Relapse was defined as another positive test between 15 and 56 days after the initial test. Multivariable regression was performed to identify factors independently associated with CDI relapse.
RESULTS: Eight hundred twenty-nine episodes met eligibility criteria, and 198 resulted in relapse (23.9%). In the final multivariable analysis, risk of relapse was associated with age (odds ratio [OR], 1.02 per year [95% confidence interval (CI), 1.01-1.03]), fluoroquinolone exposure in the 90 days before diagnosis (OR, 1.58 [95% CI, 1.11-2.26]), intensive care unit stay in the 30 days before diagnosis (OR, 0.47 [95% CI, 0.30-0.75]), cephalosporin (OR, 1.80 [95% CI, 1.19-2.71]), proton pump inhibitor (PPI; OR, 1.55 [95% CI, 1.05-2.29]), and metronidazole exposure after diagnosis (OR, 2.74 [95% CI, 1.64-4.60]). A prediction model tuned to ensure a 50% probability of relapse would flag 14.6% of CDI episodes.
CONCLUSIONS: Data from a comprehensive EHR can be used to identify patients at high risk for CDI relapse. Major risk factors include antibiotic and PPI exposure.