526. An EMR-Based Diagnostic Stewardship Intervention for GI mPCR Aimed at Reducing Inappropriate C. difficile Tests

Abstract

BackgroundDiagnostic stewardship is an emerging tool that can be used to prevent overuse of diagnostics. Because GI mPCR (GI multiplex PCR panel) tests can be ordered on formed stool, the test has lower pre-test probability for Clostridium difficile (C. difficile) infection than traditional singleplex PCR. Furthermore, after 48hours of admission, most other targets on the GI mPCR are no longer clinically relevant. Any C. difficile testing on inappropriate specimens may increase the rate of Lab ID events (positive C. difficile tests after 3 days of admission) without improving detection of true infections.MethodsIn January 2018, our 700-bed academic medical center implemented an informatics-based intervention that restricted ordering of the GI mPCR to the first 48 hours of hospitalization. After 48 hours, providers were required to contact microbiology to request an exception (see Figure 1). Singleplex PCR testing for C. difficile was available throughout admission. Orders for the GI mPCR test require the provider to note whether the patient had >3 loose stools in the previous day. Statistical analysis performed with STATA software.ResultsA total of 282 late (after 48 hours of admission) GI mPCR tests were ordered in the 104 days before restriction and 210 late tests were ordered in the 104 days after. Late GI mPCR tests (before and after restriction) resulted in diagnoses other than C. difficile less than 5% of the time (20 of 492 tests). 11.7% (24 of 210) of late GI mPCR tests were ordered for patients who did not have >3 loose stools in the previous day. Prior to restriction, 15% (41 of 282) of Lab ID events from GI mPCR were for patients who had already tested positive for C. difficile earlier in the same admission. Following the intervention, there was a decreased proportion of GI mPCR tests that were positive for C. difficile (from 14.5% to 11.3%, P = 0.26), as well as a significantly decreased rate of Lab ID events detected by GI mPCR, from 7.2/10,000 patient days to 4.0/10,000 patient days (P = 0.01).ConclusionAccurate diagnosis of C. difficile infection is important for treatment and prevention efforts, yet these data show that many rapid GI mPCR tests are inappropriately ordered on patients who may not have loose stools and who are unlikely to have an alternate diagnosis. EMR-based restriction on the GI mPCR ordering time reduced Lab ID events of C. difficile infection without missing important alternate diagnoses.Figure 1.Disclosures All authors: No reported disclosures.