Achieving and maintaining low rates of hospital-onset Clostridioides difficile

To the Editor—The Cambridge Health Alliance (CHA) is a safety-net organization consisting of 277 beds spread between 2 community teaching hospitals and 3 emergency departments. We adopted polymerase chain reaction testing (PCR, Cepheid, Sunnyvale, CA) for Clostridioides difficile infection (CDI) in 2011 and, similar to other hospitals, we noted an increase in our hospital-onset (HO) CDI rate after adopting the more sensitive molecular assay. This increase occurred despite excellent hand hygiene practices, private rooms with contact precautions, daily bleach disinfection of high-touch surfaces, ultraviolet disinfection after terminal cleaning, and an antimicrobial stewardship program. A performance improvement project initiated in 2015 led our organization to successfully reduce the HO-CD standardized infection ratio (SIR) to <1 at 2 hospitals.¹ We have been able to sustain a low HO-CD rate over the past 3.5 years despite changes in the National Healthcare Safety Network’s (NHSN’s) risk stratification methodology² and without imposing additional testing restrictions on providers. In fact, our SIR has been maintained at <1 and has continued to decline.

Methods

The initiation of an incentive program in 2015 led our institution to successfully implement a plan in February 2016 to drive the HO-CDI SIR to <1 via an automated nurse testing protocol (NTP). The goal of the NTP was to optimize identification of patients with community onset (CO) CDI to avoid inaccurate attribution of HO-CDI and inflation of the SIR due to a delay in stool specimen acquisition. In the NTP, which was embedded in Epic healthcare software (Verona, WI), documentation of diarrhea on the flowsheet during hospital days 1–3 led to a nurse best-practice alert that, when accepted, automatically led to stool collection and PCR testing of the specimen as well as initiation of contact precautions. We were careful to educate nurses and providers that a positive PCR test did not necessarily require antibiotic therapy because PCR detects C. difficile bacteria with the gene for toxin production but does not detect the toxin itself. Patients with recent risk factors for CDI or clinical characteristics of disease (eg, fever, severe diarrhea, or leukocytosis) were started on treatment, whereas those who had other reasons for diarrhea (eg, medications, diet, or laxatives) were advised to continue with watchful waiting.

In July 2016, a combined glutamate dehydrogenase (GDH) antigen and toxin assay for toxins A and B (Abbott Diagnostics, Chicago, IL) was implemented for patients experiencing diarrhea beyond hospital day 3 due to its increased specificity. Diagnostic stewardship measures included limitations of C. difficile testing to stool samples taking the shape of the container and 1 test per patient per week.

In January 2018, the NHSN changed its risk stratification according to the last C. difficile test performed.³ This change had the potential to negatively impact our healthcare-associated CDI metrics. Hence, we report an update on our performance since our prior publication.

Results

Immediately after implementation of the NTP in February 2016, both CHA sites observed fewer HO-CDI cases than would be predicted according to bed size, hospital affiliation, number of CO-CDI cases, and laboratory test used (Fig. 1). Upon adoption of the GDH/toxin assay in July 2016 and the change in the NHSN risk stratification in January 2018, this improvement was sustained.

Fig. 1. Clostridioides difficile standardized infection ratio (SIR) over time.

Discussion

Our implementation of an automated process for identifying CO-CDI via the NTP, a process with minimal burden to staff and providers, has driven our CDI SIR to <1, and this change has been maintained for nearly 4 years.

Notably, indeterminate GDH and toxin results are reflexed to PCR testing. Although this methodology has the potential to increase our SIR due to the fact that the PCR test is of higher sensitivity, we have maintained this testing strategy (ie, not requiring approval from the infectious disease team to obtain this test) because our excellent performance has been maintained. We argue that delaying PCR testing could lead to delayed diagnosis and treatment of true CDI and therefore pose a danger to patient care.

We also posit that restrictions or hard stops that prevent providers from ordering C. difficile testing in the setting of recent laxative use have the potential to delay identification and treatment of CDI. A recent publication supported this argument,⁴ reporting that, despite 2017 IDSA-SHEA guidelines⁵ that recommend against testing for CDI if a patient has received a laxative within the preceding 48 hours, patients who recently received laxatives had no difference in CDI or symptom severity. Had we followed the IDSA-SHEA guidelines, a diagnosis of CDI would have been missed in nearly one-third of the laxative-treated cohort.

Controversy persists regarding the optimal testing strategy for diagnosing CDI. Our approach has been effective in maintaining a low SIR without imposing the need for approval from the infectious disease team or the laboratory, without a hard stop for laxative use in the preceding 48 hours, and without imposing PCR testing restrictions for patients with indeterminate GDH and toxin results. Empowering nurses in both diagnostic and antimicrobial stewardship efforts for CDI continues to be an effective strategy for reducing HO-CDI.