100. Development and Implementation of a 2-Tier Testing Algorithm for Clostridioides difficile: An Evaluation of Outcomes on Patients with Indeterminate Results at 90 Days

BACKGROUND: There is no definitive gold standard for accurate diagnosis of Clostridioides difficile (C difficile) infection. There is ample evidence that relying on a molecular test such as Polymerase Chain Reaction (PCR) for diagnosis, can lead to over diagnosis and unnecessary treatment. Combined, multi-step algorithms have been proposed to improve specificity of testing. The challenge remains in interpreting discordant or indeterminate results. Additionally, the risk of hospitalization due to lack of treatment for indeterminate results remains unclear. METHODS: To improve C difficile testing, a new 2-tier algorithm was implemented in 2019 starting with PCR testing. An indeterminate result was defined as a sample with a positive PCR and a positive Glutamate Dehydrogenase (GDH)/negative toxin result or a positive PCR and a negative GDH/positive toxin result. Indeterminate results were classified by episode severity and number. Patient records were reviewed by the Antimicrobial Stewardship (AS) physician and pharmacist to determine true infection versus colonization. Treatment was given as per recent IDSA Guidelines. All patients with indeterminate results were followed for 90 days for development of infection or hospitalization due to C difficile. Adults with stool samples submitted for testing between 6/1/2019 and 12/31/2019 were included. A total of 169 specimens were reviewed: 75 were positive, 72 were indeterminate (4 excluded from final analysis) and 22 were negative. RESULTS: Using a 2-tier testing algorithm, 68 (41%) of all results were indeterminate. Our AS classified 47 (69%) of those as infection and 21 (31%) as colonization. Patients with indeterminate results who were treated had a low incidence (8.5%) of reinfection requiring hospitalization in the following 90 days. There were no hospitalizations in the untreated group. Of patients with an indeterminate result who were treated, 42 (89%) were categorized as an initial episode of C difficile infection. CONCLUSION: Clinical correlation of indeterminate results is critical to algorithm interpretation. A combined approach with provider education, an electronic testing advisor, a 2-tier testing algorithm, daily monitoring and prescribing by the AS team resulted in favorable outcomes for patients with indeterminate results DISCLOSURES: All Authors: No reported disclosures