Reductions in Clostridium difficile Infection (CDI) Rates Using Real-Time Automated Clinical Criteria Verification to Enforce Appropriate Testing.

Abstract

C. difficile PCR testing identifies both colonized and infected patients, making it critical to only test patients that meet clinical criteria for C. difficile infection (CDI). We implemented an automated order-entry protocol that reduced inappropriate testing by 64% and hospital-onset (HO) CDI Standardized Infection Ratio (SIR) from 1.62 to 0.82.

Introduction

C. difficile infection (CDI) is diagnosed in over 450,000 patients annually.¹ CDI rates increased 3.5-fold from 2000–2008, coinciding with the widespread adoption of highly-sensitive PCR-based testing which cannot distinguish between colonization and active colitis.² Asymptomatic colonization can be present in 20–40% of hospitalized patients; inappropriate CDI testing can lead to false positive tests and unnecessary treatment.^2–3 While controversy over optimal CDI testing continues, strategies to enforce clinically appropriate testing are urgently needed.^4–6

We created a real-time computer physician order entry (CPOE) alert to enforce appropriate C. difficile testing and reduce CDI rates.

Methods

We conducted a pre/post intervention cohort study to evaluate C. difficile testing in adults hospitalized at a 417-bed academic hospital between 04/1/15–06/30/17. Baseline (4/1/15–3/31/16) and intervention (7/1/16–6/30/17) periods were compared, excluding a 3-month phase-in period (4/1/16–6/30/16). CDI testing was PCR-based and remained unchanged throughout the study period. The intervention involved automated real-time CPOE verification to enforce appropriate CDI testing criteria : (1) diarrhea (≥3 liquid/watery stools in 24 hours), (2) no alternate cause for diarrhea, (3) no laxative use within 24 hours, (4) no previous CDI test result within 7 days, and (5) age >1 year.^{5, 6} Clinicians were required to attest to criteria 1 and 2; criteria 3–5 were programmed to auto-populate on the ordering screen, including laxative name and time administered if given within 24 hours. Contraindication to testing resulted in a “hard stop” prompt instructing prescribers to either exit the order or submit the name of an approving Infectious Diseases (ID) or Gastrointestinal (GI) physician to override hospital protocol (Supplemental Figure 1).

To ensure adherence, infection preventionists reviewed overrides weekly. Approving ID/GI physician names were verified and physicians placing orders without appropriate approval received a warning email signed by ID/GI leadership and the Chief Medical Officer (CMO) that reiterated protocol criteria and reminded physicians that orders without approval are being monitored (Supplemental Text).

We evaluated: (1) National Healthcare Safety Network (NHSN) case counts/10,000 patient-days and standardized infection ratios (SIRs), (2) tests ordered in patients receiving laxatives within 24 hours, (3) repeat testing within 7 days, and (4) protocol overrides. Chi-squared tests compared changes in CDI testing and rates pre- versus post- intervention; quarterly SIRs were compared using t-tests.

Results

Baseline CDI testing rate decreased from 284/10,000 to 268/10,000 patient-days post-intervention (p=0.02). CDI testing in the hospital-onset (HO) period decreased 56% post-intervention, from 155 to 84 tests/10,000 patient days (p<0.001). At baseline, 49% of CDI tests were on patients receiving laxatives within 24 hours and 18% were ordered despite prior results available within 7 days. Testing while on laxatives decreased by 64%, from 77 to 24/10,000 patient-days pre- vs post-intervention (p<0.001), (Figure 1B). CDI tests reordered within 7 days also decreased by 64%, from 28 to 8/10,000 patient-days (p<0.001). HO CDI rates decreased 54%, from 17 to 7 cases/10,000 patient-days (p<0.001), resulting in a 51% reduction in the average quarterly HO SIR, from 1.62 to 0.82 (p<0.001) (Figure 1B). Improved testing protocol compliance was tied to monitoring and feedback with a templated CMO response to physicians bypassing the protocol without approval. In the first month of implementation, there were 22 unauthorized overrides; this decreased to zero by the end of the study period.

Figure 1.

Hospital-onset C. difficile orders decreased after launch of the automated real-time intervention, while community-onset orders were unchanged. The number of orders placed on patients receiving laxatives decreased sharply after CPOE launch.

Discussion

Proactive approaches to clinically-appropriate diagnostic testing can be important for high sensitivity tests such as the C. difficile PCR test which can identify colonization and lead to unnecessary treatment and concern.^1–3 Our real-time CPOE criteria-based testing protocol reduced inappropriate testing by 64% and HO C. difficile rates by 50% - without changing CDI testing method.

EHR strategies using passive alerts with information alone run the risk of being ignored over time and can have variable compliance.⁸ Our smart prompt provided clinicians with actionable data and also used a “hard stop” when testing criteria were not met. To address the rare but important possibility of CDI in complicated or high risk patients not meeting testing criteria (e.g., ICU patient on daily laxatives who develops abdominal distention and leukocytosis), physicians could override the protocol with ID/GI approval. This strategy encouraged thoughtful testing and provided an opportunity for specialist-level education of front-line physicians.

Electronic algorithms and protocols can often be circumvented; compliance monitoring and timely feedback are needed to achieve meaningful and sustainable changes.⁸ In our case, noncompliant physicians were sent email warnings signed by our CMO, sending a clear message that appropriate testing was an institutional priority while also providing an opportunity to educate physicians.

An important limitation of this intervention was the inability to capture the number of times a CDI test order was initiated but then cancelled due to the protocol, limiting our ability to describe the learning curve associated with this CPOE strategy. Nevertheless, the sustained decreases in overall testing strongly suggest decreases in order initiation.

Data on the harmful effects antibacterial agents on the gut microbiome are mounting and treatment of asymptomatic C. difficile colonization has been shown to increase future risk of colitis and recurrent disease.^9–10 In addition, oral vancomycin use increases the carriage rate of vancomycin-resistant enterococci, a drug resistant organism associated with healthcare-associated infections.¹⁰

As data showing the harms of over-testing and over-treatment for CDI emerge, CPOE strategies can be an effective training tool to improve use and stewardship of diagnostic tests.^{2, 3} Our electronic solution to enforce clinically-appropriate CDI testing is an example of one that integrates real-time CPOE alerts, specialist review, compliance monitoring and feedback, and leadership-level enforcement.