Implementation of Two-Step Clostridioides difficile Testing Algorithm and Management of Possible Carriers

Background: Updated IDSA-SHEA guidelines recommend different diagnostic approaches to C. difficile depending on whether There are pre-agreed institutional criteria for patient stool submission. If stool submission criteria are in place, nucleic acid amplification testing (NAAT) alone may be used. If not, a multistep algorithm is suggested, incorporating various combinations of toxin enzyme immunoassay (EIA), glutamate dehydrogenase (GDH), and NAAT, with discordant results adjudicated by NAAT. At our institution, we developed a multistep algorithm leading with NAAT with reflex to EIA for toxin testing if NAAT is positive. This algorithm resulted in a significant proportion of patients with discordant results (NAAT positive and toxin EIA negative) that some experts have categorized as possible carriers or C. difficile colonized. In this study, we describe the impact of a multistep algorithm on hospital-onset, community-onset, and healthcare-facility–associated C. difficile infection (HO-CDI, CO-CDI, and HFA-CDI, respectively) rates and the management of possible carriers. Methods: The study setting was a 399-bed, tertiary-care VA Medical Center in Richmond, Virginia. A retrospective chart review was conducted. The multistep C. difficile testing algorithm was implemented June 4, 2019 (Fig. 1). C. difficile testing results and possible carriers were reviewed for the 5 months before and 4 months after implementation (January 2019 to September 2019). Results: In total, 587 NAATs were performed in the inpatient and outpatient setting (mean, 58.7 per month). Overall, 123 NAATs (21%) were positive: 59 in the preintervention period and 63 in the postintervention period. In the postintervention period, 23 positive NAATs (26%) had a positive toxin EIA. Based on LabID events, the mean rate of HO+CO+HCFA CDI cases per 10,000 bed days of care (BDOC) decreased significantly from 9.49 in the preintervention period to 1.15 in the postintervention period ( P = .019) (Fig. 2). Also, 9 of the possible carriers (22%) were treated for CDI based on high clinical suspicion, and 6 of the possible carriers (14%) had a previous history of CDI. Of these, 5 (83%) were treated for CDI. In addition, 1 patient (2%) converted from possible carrier to positive toxin EIA within 14 days. The infectious diseases team was consulted for 11 possible carriers (27%). Conclusions: Implementation of a 2-step C difficile algorithm leading with NAAT was associated with a lower rate of HO+CO+HCFA CDI pe