ABSTRACT
OBJECTIVES: Diagnosis of Clostridium difficile infection (CDI) can be challenging due to high colonization rates. Unlike PCR-only testing, two-step algorithm testing (that includes toxin and PCR) may help differentiate colonization from active infection, but it is unknown if this type of testing impacts treatment decisions. We examined the association between changing CDI diagnostic methods, the way the testing results were displayed, and the rates of CDI-specific treatment. METHODS: We performed a retrospective analysis of positive C. difficile cases over 2 years, a year preceding and following our institution's transition from PCR to two-step testing. During the PCR period, results were displayed in the electronic medical record as 'positive'. In the two-step period, positive results were either displayed as 'likely colonized' or 'toxin positive'. Rates of CDI-specific therapy and adverse patient outcomes (30-day mortality and intensive care unit admission) were compared among the three groups. RESULTS: A total of 610 patients had positive results over the study period. Of the 354 patients in the PCR group, 329 (93%) were treated with CDI-specific therapy. Of the 142 patients in the likely colonized group, 59 (42%) were treated. All 114 patients in the toxin-positive group were treated. Multivariate analysis of patients who were PCR positive or likely colonized showed that tests sent in the two-step era were less likely to be associated with treatment for CDI (odds ratio 0.05, 95% CI 0.03-0.09). DISCUSSION: We found a correlation between changing the type of test and the way the results were displayed and reduction in CDI-specific antibiotic use without restricting clinician diagnostic ordering.