Using Clinical Decision Support to Improve Evidence Based Testing and Diagnosis of Clostridium difficile Infection

BACKGROUND: Diagnosing Clostridium difficile infection (CDI) requires clinical understanding of the disease and knowledge of diagnostic testing limitations. It is important for providers to utilize CDI testing only in patients with suspected disease. Real-time polymerase chain reaction (PCR) assays are sensitive but cannot differentiate between symptomatic and asymptomatic patients. Individual hospitals have reported a 50% to 100% increase in the rate of CDI after substituting toxin tests with molecular tests such as PCR. We conducted a quality improvement project, implementing clinical decision support in ordering diagnostic testing of CDI, while measuring the number of diagnostic tests ordered and positive results. METHODS: We implemented evidence based clinical decision support into Cerner order entry system on March 1, 2016. The Cepheid Xpert C. difficile molecular test is used for diagnosis of CDI at our facility. The decision support included a message stating Òuse the test with caution in patients who are receiving tube feeds or recent laxative useÓ and prompted ordering providers to select one of three indications for using the test: 3 or more diarrheal stools per 24 hour period, leukocytosis with abdominal pain, or ileus. A control chart was used to monitor the number of tests ordered and positive tests per month (inpatient adults) for a total of 24 months; 14 months pre-intervention and 10 months post-intervention. RESULTS: A decrease in the number of tests ordered per month was seen post intervention. Average number of monthly tests ordered was 207 pre-intervention and 163 post-intervention. After controlling for patient-days per month, there was a 13.5% decrease in the number of tests ordered from a mean of 14.29 vs.. 12.37 tests per thousand patient-days per month. This resulted in special cause variation (Figure 1). There was no special cause variation detected with the number of positive PCRs per month, pre and post intervention. CONCLUSION: Implementing decision support into the electronic medical record may assist providers with evidence-based utilization of the C. difficile PCR by decreasing unnecessary testing. This decrease may also have an impact on overall hospital costs, antibiotic utilization, and public reporting related to CDI. DISCLOSURES: All authors: No reported disclosures.