An Algorithm Addressing the Problem of Overdiagnosis of Clostridioides difficile Infections in Hematopoietic Stem Cell Transplantation Recipients: Effects on C. difficile Infection Rates and Patient Outcomes

Abstract

Diarrhea of other causes and Clostridioides difficile colonization are common in patients hospitalized for hematopoietic stem cell transplantation (HSCT). It has been well recognized that these issues tend to decrease the specificity of stool testing for C. difficile infection (CDI). The best way to address this problem is uncertain. In September 2018, we initiated a project with the goal of addressing the apparent problem of overdiagnosis of CDIs in our HSCT population. Using the quality improvement tool Model for Improvement, we introduced a C. difficile stool testing and CDI diagnosis algorithm with the aim of decreasing unnecessary inpatient CDI diagnoses and treatments. In this study, we examined the effects of the algorithm. We reviewed all HSCT admissions for the 2 years before introduction of the algorithm and the subsequent 3 years, recording all stool submissions for C. difficile determination and CDI. At the close of the study, we recruited our advanced practice providers (APPs) to review all CDIs following algorithm initiation and provide feedback on the ease of use of the algorithm and potential improvements to the overall process. Stool submissions for C. difficile determination decreased from 38.0/1000 to 20.6/1000 inpatient days (P < .001) and CDI diagnoses decreased from 5.5/1000 to 2.4/1000 days (P=.007). Patients admitted for a first allogeneic HSCT, first autologous HSCT, or HSCT readmission showed similar proportionate reductions. No detrimental effects on hospital length of stay, overall survival, progression-free survival, rate of readmission post-HSCT, incidence of acute graft-versus-host disease, or incidence of recurrent CDI were noted following algorithm introduction. A strategy of education, monitoring/feedback, and ease of algorithm access proved effective in inducing provider compliance. APPs rated the algorithm high on ease of use. We conclude that the use of an algorithm defining criteria for C. difficile testing, diagnosis, and treatment was associated with significantly decreased CDI diagnoses on a HSCT inpatient unit without apparent adverse effects.