Advertisement

Impact of a multipronged approach to reduce the incidence of Clostridioides difficile infections in hospitalized patients

Open AccessPublished:September 05, 2022DOI:https://doi.org/10.1016/j.ajic.2022.08.027

      Highlights

      • The incidence of C. difficile dramatically decreased after a set of interventions.
      • A simple admission question by nurses identified cases present on admission.
      • A 2-step testing algorithm decreased the diagnosis of new C. difficile infections.
      • The reductions in new Clostridioides difficile infections were durable.

      Background

      Effective approaches to reduce Clostridioides difficile infections (CDI) in hospitalized patients are needed. We report data from 3 years preceding and 3 years following interventions that proved successful, with detailed analysis of all cases the first year after implementation.

      Methods

      Interventions included a nursing protocol to identify cases present on admission by asking if the patient had 1 or more liquid stools in the last 24 hours, and a 2-step testing algorithm with samples positive by polymerase chain reaction (PCR) for the C. difficile toxin gene reflexing to an enzyme immunoassay (EIA) for the toxin antigen.

      Results

      Healthcare-associated infections due to CDI fell from ∼160 in each of the preceding 3 years to <65 in each of the subsequent 3 years (P < .001), while the ratio of observed-to-expected hospital-onset cases diminished to ∼0.50 (P < .02). In the first year, 395 samples were PCR(+), but only 118 (29.9%) of these were EIA(+). 55 (46.6%) of the PCR(+)/EIA(+) samples were from hospital day 1 or 2 and classified as present on admission. The mean time from stool collection to report of PCR results was ∼7.5 hours, and the EIA took on average only 68 additional minutes to be reported.

      Conclusions

      The number of incident CDI cases can be dramatically decreased by implementing an admission screening question and a 2-step testing algorithm.

      Key Words

      The Emerging Infections Program (EIP) of the Centers for Disease Control and Prevention (CDC) recently estimated from data collected at designated EIP locations that there were ∼236,000 cases of healthcare-associated Clostridioides difficile infection (CDI) in the United States in 2017, resulting in >16,000 in-hospital deaths.
      • Guh AY
      • Mu Y
      • Winston LG
      • et al.
      Trends in U.S. burden of Clostridioides difficile infection and outcomes.
      This analysis defined health care-associated CDI as someone >1 year of age with a stool sample positive by a C. difficile toxin or molecular assay who was hospitalized >3 days, in a healthcare facility in the previous 12 weeks, or resided in a long-term facility. Reducing such infections has been a major goal of antimicrobial stewardship programs (ASPs),
      • Dellit TH
      • Owens RC
      • McGowan Jr., JE
      • et al.
      Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship.
      ,
      • Barlam TF
      • Cosgrove SE
      • Abbo LM
      • et al.
      Implementing an antibiotic stewardship program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America.
      and the same report found the adjusted burden of these infections had fallen by ∼6%/year and 36% overall from 2011 to 2017.
      • Guh AY
      • Mu Y
      • Winston LG
      • et al.
      Trends in U.S. burden of Clostridioides difficile infection and outcomes.
      It also noted that the CDC's National Healthcare Safety Network (NHSN) found a decline of 20% from 2015 to 2017 in the risk-adjusted C. difficile standardized infection ratio (SIR).
      • Guh AY
      • Mu Y
      • Winston LG
      • et al.
      Trends in U.S. burden of Clostridioides difficile infection and outcomes.
      ,
      Centers for Disease Control and Prevention
      2017 National and State Healthcare-Associated Infections Progress Report.
      The SIR compares observed to expected numbers of hospital-onset (HO) CDI, which focusses on cases in hospitalized patients and uses a similar definition of a positive test on a sample from hospital day (HD) 4 or later, but from data reported to the NHSN multi-drug resistant organism (MDRO)/CDI module via the laboratory-identified event reporting option.
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].
      Our hospital established an ASP in 2011 and documented a 21% decrease in antibiotic days of therapy/1000 patient-days from 2010 to 2018, with a more profound 58% decrease in fluoroquinolone use during this time.
      • Katzman M
      • Kim J
      • Lesher MD
      • et al.
      Customizing an electronic medical record to automate the workflow and tracking of an antimicrobial stewardship program.
      Nonetheless, we were experiencing increasing numbers of CDI cases that met the NHSN surveillance definition for a health care-associated infection (HAI) due to CDI
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      CDC/NHSN Surveillance Definitions for Specific Types of Infections [Chapter 17].
      — defined as a positive test on a sample from HD 3 or later
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Identifying Healthcare-associated Infections (HAI) for NHSN Surveillance [Chapter 2].
      plus evaluation by an infection preventionist
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].
      (this definition includes tests 1 day earlier than the purely laboratory-based definitions described above). However, careful analysis showed much of the rising case counts could be attributed to increased patient-days and changes in testing methodology or reporting requirements, while our hospital's SIR had been ≤1 since 2013.
      Notwithstanding the above explanations, our hospital was still documenting an average of 14 cases/month in 2017, despite attention to cleaning protocols, extending contact precautions until discharge for most CDI patients, and instituting after-action reviews. We therefore introduced a concerted set of interventions to reduce the number of incident CDI cases. This report analyzes our data before and after implementation of these successful interventions.

      Methods

      Setting

      Milton S. Hershey Medical Center is an ∼550-bed academic health center, including an ∼125-bed children's hospital. The electronic medical record (EMR) is Cerner Millennium PowerChart (Kansas City, MO), which supports customized MPages to collect data from relevant patients.

      Laboratory methodology

      All CDI testing is done in our hospital-based laboratory on unformed stool samples only. Before April 2011, samples were tested by a tissue culture-based toxin assay.
      • Kelly CP
      • LaMont JT.
      Clostridium difficile infection.
      Subsequently, samples were tested by a molecular assay, first with a loop-mediated isothermal amplification (LAMP) assay (Illumigene C. difficile, Meridian Bioscience, Cincinnati, OH), and since June 2017 with a polymerase chain reaction (PCR) assay (Simplexa C. difficile Direct Kit, Focus Diagnostics, Cypress, CA). A 2-step testing protocol began July 2, 2018, with all positive PCR tests for the C. difficile toxin gene reflexing to an enzyme immunoassay (EIA) for the toxin antigen (C. DIFF QUIK CHEK COMPLETE, TechLab, Blacksburg, VA).

      Interventions in effect as of July 2018 and associated educational messaging

      Information about the following was disseminated to the medical staff, housestaff, advanced practice clinicians, administrators, and nursing leaders on Monday, July 2, 2018:
      • (1)
        A nursing protocol to identify CDI cases present on admission (POA): As of May 18, 2018, all admissions >1 year of age were screened by asking: "Has the patient had 1 or more liquid or watery stools in the last 24 hours?" An affirmative response triggers an order for C. difficile stool testing and contact precautions. Educational efforts stressed that this order should not be overridden, even if a clinician deemed the diarrhea to be from another cause.
      • (2)
        Introduction of a 2-step CDI testing algorithm: As of July 2, 2018, all samples positive by PCR for the toxin gene reflex to an EIA for the antigen. Educational efforts stated that patients with diarrhea who are PCR(+)/EIA(-) are likely carriers
        • Polage CR
        • Gyorke CE
        • Kennedy MA
        • et al.
        Overdiagnosis of Clostridium difficile infection in the molecular test era.
        and may not need treatment but still require isolation because they are shedding C. difficile in their diarrhea.
      • (3)
        A new order set for treating adults with C. difficile: Available June 27, 2018, based on newly published guidelines.
        • McDonald LC
        • Gerding DN
        • Johnson S
        • et al.
        Clinical practice guidelines for Clostridium difficile infection in adults and children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA).
      • (4)
        Fidaxomicin added to formulary: Available June 19, 2018, with education that it may lead to fewer relapses than vancomycin
        • Louie TJ
        • Miller MA
        • Mullane KM
        • et al.
        Fidaxomicin versus vancomycin for Clostridium difficile infection.
        ,
        • Cornely OA
        • Crook DW
        • Esposito R
        • et al.
        Fidaxomicin versus vancomycin for infection with Clostridium difficile in Europe, Canada, and the USA: a double-blind, non-inferiority, randomised controlled trial.
        and would be appropriate for certain patients.
      • (5)
        Other interventions: We now encouraged stewardship of acid-suppressing drugs,
        • Janarthanan S
        • Ditah I
        • Adler DG
        • Ehrinpreis MN.
        Clostridium difficile-associated diarrhea and proton pump inhibitor therapy: a meta-analysis.
        probiotics for some patients on antibiotics,
        • Goldenberg JZ
        • Yap C
        • Lytvyn L
        • et al.
        Probiotics for the prevention of Clostridium difficile-associated diarrhea in adults and children.
        and secondary prophylaxis with vancomycin for patients receiving antibiotics who had CDI in the previous 12 months.
        • Carignan A
        • Poulin S
        • Martin P
        • et al.
        Efficacy of secondary prophylaxis with vancomycin for preventing recurrent Clostridium difficile infections.
        ,
        • Van Hise NW
        • Bryant AM
        • Hennessey EK
        • Crannage AJ
        • Khoury JA
        • Manian FA.
        Efficacy of oral vancomycin in preventing recurrent Clostridium difficile infection in patients treated with systemic antimicrobial agents.

      Study design

      We conducted a quasi-experimental study comparing the 3 years before and 3 years after the above interventions were in effect, with detailed analysis of CDI cases in the 12 months following implementation. For quality review purposes, our ASP had designed an MPage, effective July 1, 2018, to capture data from hospitalized patients with a positive C. difficile PCR test. For the current study, which was submitted to our Institutional Review Board and deemed to meet the criteria for exempt research according to applicable regulations, the following data were analyzed: admission date, date/time and HD of sample collection for the first positive PCR test that admission, date/time the result was reported, result of the corresponding EIA test, date/time that was reported, number of days with an antimicrobial in the 60 days before the date of the sample, number of days with a proton-pump inhibitor or histamine-2 receptor blocker in the 7 days before the collection date, and number of days with a laxative in the 3 days before the collection date.

      Terminology and statistical analysis

      Explanations of the terms HAI-CDI and HO-CDI are shown in Table 1. Data from the 3 years before and 3 years after the interventions were compared using the 2-tailed t test in Microsoft Excel (Redmond, WA), and percentages were compared by chi-square tests in SAS software version 9.4 (Cary, NC). P values < .05 were considered significant.
      Table 1Terminology for cases of C. difficile infection
      TermAbbreviation used in this reportRelevant criteria for incident cases in hospitalized patients
      The date of admission is considered hospital day 1.
      Usage
      Healthcare-associated infection due to CDI
      Also referred to by the NHSN as C. difficile gastrointestinal system infection or GI-CDI.57
      HAI-CDIPositive test on HD 3 or later and assessment by infection preventionistReported to NHSN MDRO/CDI module for infection surveillance purposes
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].
      ,
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Identifying Healthcare-associated Infections (HAI) for NHSN Surveillance [Chapter 2].
      Hospital-onset (or healthcare facility-onset) CDI
      The two terms are used for the same entity and share this abbreviation.45
      HO-CDIPositive test on HD 4 or laterReported to NHSN MDRO/CDI module as a LabID Event
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].
      and used by CDC to calculate SIR
      Centers for Disease Control and Prevention
      2017 National and State Healthcare-Associated Infections Progress Report.
      CDC, Centers for Disease Control and Prevention; CDI, C. difficile infection; HD, hospital day; LabID, laboratory-identified; MDRO, multi-drug resistant organism; NHSN, National Healthcare Safety Network; SIR, standardized infection ratio.
      low asterisk The date of admission is considered hospital day 1.
      Also referred to by the NHSN as C. difficile gastrointestinal system infection or GI-CDI.
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      CDC/NHSN Surveillance Definitions for Specific Types of Infections [Chapter 17].
      The two terms are used for the same entity and share this abbreviation.
      Centers for Disease Control and Prevention
      2017 National and State Healthcare-Associated Infections Progress Report.
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].

      Results

      Analysis of baseline data

      The number of C. difficile cases that met the NHSN surveillance definition for HAI-CDI (ie, laboratory-defined and clinically-consistent on or after HD 3) increased from an average of 4.5/month in the first half of 2009 to 15.5/month in the second half of 2017 (Fig 1A). However, the number of patient-days had grown from a mean of 11,757/month in 2009 to 16,229/month in 2017, and normalizing the monthly case count for every 10,000 patient-days reduced the upward slope of the trendline (as described in the legend to Fig 1B).
      Fig 1
      Fig 1Baseline C. difficile data. (A) The number of cases of C. difficile infection (CDI) per month that met the NHSN surveillance definition for a healthcare-associated infection (HAI)-CDI (occurring on hospital day 3 or later, as in ) is shown from 2009 through 2017. The slope for the linear trendline indicates an increase each month of 0.101 cases. (B) The same data, but normalized per 10,000 patient-days (the y-axis is not to the same scale). The trendline is not shown but its slope indicates an increase each month of 0.049 cases per 10,000 patient-days. Four time periods (1, 2, 3, 4) are indicated, as defined by changes in testing methods or reporting requirements. (C) The quarterly Standardized Infection Ratios for hospital-onset (HO)-CDI cases (occurring on hospital day 4 or later, as in ), has been reported by the CDC since 2013. The heavy line indicates 1.0, where the observed number equals the expected number.
      Additional analysis in early 2018 sought to correlate the residual increases with changes in testing methodology or reporting requirements. The first major change occurred in April 2011, when our laboratory moved from a tissue culture-based toxin assay to LAMP technology. This molecular assay was reported to be 22% more sensitive than the tissue culture assay,
      • Lalande V
      • Barrault L
      • Wadel S
      • Eckert C
      • Petit JC
      • Barbut F.
      Evaluation of a loop-mediated isothermal amplification assay for diagnosis of Clostridium difficile infections.
      and indeed, our average monthly case count/10,000 patient-days increased by 14%, from 4.4 before this change to 5.0 in the period immediately afterwards (Fig 1B, periods 1 and 2, respectively).
      The next major change occurred in January 2015, when the NHSN surveillance definition of HAI-CDI no longer permitted excluding cases as POA when there was evidence in the chart that the patient had diarrhea in the 2 days before admission or during the first 2 hospital days. Such cases would now be deemed HAI-CDI if the first positive C. difficile test was from HD 3 or later, regardless of chart documentation of earlier symptoms.
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      NHSN HAI Surveillance Changes for 2015.
      In fact, the average monthly case count/10,000 patient-days increased by 60%, from 5.0 before this reporting change to 8.0 afterwards (Fig 1B, periods 2 and 3, respectively).
      The final relevant change occurred in June 2017, when the laboratory moved from a LAMP to a PCR assay; internal comparisons showed the latter test could increase positive results by 16%.
      • England M
      • Erdman P
      • Greene W.
      Evaluation of two molecular assays for the detection of Clostridium difficile.
      Indeed, this change was associated with an increase in the average monthly case count/10,000 patient-days of 20%, from 8.0 before the change to 9.6 afterwards (Fig 1B, periods 3 and 4, respectively). Moreover, focusing on HO-CDI cases (i.e., laboratory-identified cases on or after HD 4) showed the SIR for our hospital had been ≤1 since the last quarter of 2013 (Fig 1C).

      Conceptual model and hypotheses derived from baseline data

      Though the risk-adjusted SIR had been ≤1, we were still experiencing ∼16 cases categorized as HAI-CDI/month in the second half of 2017 (Fig 1A). Figure 2 shows a theoretical breakdown of those 16 cases, forming the framework for the hypotheses that drove our subsequent interventions. Because internal analysis found that 28% of cases in the second half of 2017 were diagnosed on stool from HD 3 or 4, many cases after January 2015 may have been POA but were included in the more-stringent reporting requirements. Thus, our first hypothesis was that ∼25% of the 16 cases were actually POA (Fig 2, left). Our second hypothesis, based on the work of Polage et al.,
      • Polage CR
      • Gyorke CE
      • Kennedy MA
      • et al.
      Overdiagnosis of Clostridium difficile infection in the molecular test era.
      was that half the cases identified by PCR reflected colonization, rather than CDI that might respond to treatment. Thus, colonization might account for 2 of the 4 cases postulated to be POA and 6 of the 12 cases not POA. Finally, based on genomic sequencing and molecular subtyping studies,
      • Eyre DW
      • Cule ML
      • Wilson DJ
      • et al.
      Diverse sources of C. difficile infection identified on whole-genome sequencing.
      ,
      • Curry SR
      • Muto CA
      • Schlackman JL
      • et al.
      Use of multilocus variable number of tandem repeats analysis genotyping to determine the role of asymptomatic carriers in Clostridium difficile transmission.
      we presumed that as a rough approximation, 1/3 of cases might be attributable to spread from another symptomatic CDI case, 1/3 might be acquired from an asymptomatic carrier, and 1/3 might be from the patient's endogenous flora. Thus, of our hypothesized 6 cases of true HAI-CDI, 2 might come from each of these 3 sources (Fig 2, bottom right).
      Fig 2
      Fig 2Conceptual framework for 16 incident cases of C. difficile per month in late 2017. The main assumptions and hypotheses of this model are that: (i) 1/4 of cases that met the NHSN definition for a healthcare-associated infection due to C. difficile infection (HAI-CDI, as in ) were present on admission but not identified as such; (ii) diarrhea in 1/2 of cases with a positive polymerase chain reaction result involves colonization rather than being due to C. difficile; and (iii) three potential sources for true incident cases, with comparable 1/3 probability, are another active hospitalized case, an asymptomatic colonized patient, or the patient's own bowel flora. All assumptions are approximate but provide the framework for the main interventions in this report.
      Figure 2 also underscores key aspects of this model: (1) the 6 true cases not POA might be preventable via antimicrobial stewardship, (2) the 4 true cases not related to the patient's preexisting flora were potentially preventable via infection control measures, and (3) all 16 patients with diarrhea and a positive PCR are excreting C. difficile organisms or spores into their environment and represented potential sources of spread.

      Analysis of post-intervention data

      In the first year following the interventions, the number of HAI-CDI cases, which had been ∼160 for 3 consecutive years before July 2018, fell from 161 to 63 (Fig 3, the column graph). Importantly, the improvement was durable, with 52 cases in the most recent complete year of 2020-2021. The decrease in case counts in the 3 years after, compared to the 3 years before the interventions was statistically significant (P < .001). Additionally, the SIR values for HO-CDI cases, which had hovered around 0.75 in the 3 years before July 2018, also fell in the first year following the interventions, from 0.714 to 0.459, and now hovers around 0.5 (Fig 3, the line graph). The decrease in the SIR values in the 3 years after, compared to the 3 years before the interventions was also statistically significant (P < .02).
      Fig 3
      Fig 3C. difficile infection (CDI) in the 3 years before and after the interventions. The column graph shows yearly cases that met the NHSN surveillance definition for a healthcare-associated infection (HAI)-CDI (occurring on hospital day 3 or later, as in ), and the line graph shows the yearly Standardized Infection Ratio (SIR) for hospital-onset (HO)-CDI cases (occurring on hospital day 4 or later, as in ). Data reflect academic years (July to June), and the vertical arrow indicates July 2018 (when the interventions in this report were all in effect). The P values reported by the CDC indicate the SIRs were significantly different from 1 for 2 of the first 3 years (P < .001 for 2015-16 and 2017-18) and all 3 later years (P < .0001 for 2018-19, 2019-20, and 2020-21). The case count and SIR values in the 3 later years were significantly lower than the 3 earlier years (P < .001 and P < .02, respectively).
      We also undertook a detailed analysis (see Methods) for all positive PCR tests in the first year of the interventions. From July 2018 to June 2019, there were 3,330 PCR tests for the toxin gene performed on stool samples from inpatients, and 395 (11.9%) of these tests were positive (Table 2). Approximately half of the tests performed, as well as half of the positive tests, were obtained on HD 1 or 2, and the percentage of tests that were positive was similar for HD 1 or 2 compared to HD 3 or later. All PCR(+) stools then reflexed to the EIA, and 118 (29.9%) were positive for the toxin antigen. Of the 118 PCR(+)/EIA(+) samples, 55 (46.6%) were from HD 1 or 2, and 63 (53.4%) were from HD 3 or later. The percentage of PCR(+) tests that were EIA(+) also appeared to be slightly lower on HD 1 or 2 compared to HD 3 or later, though none of these differences were statistically significant. The complete data set is shown in Table 2.
      Table 2Results of C. difficile assays in first year after interventions
      TestTotalHospital day 1 or 2
      The date of admission is considered hospital day 1.
      Hospital day 3 or later
      The date of admission is considered hospital day 1.
      PCR for Toxin Gene
       Number of tests performed3,3301,6921,638
       Percentage of tests performed50.8%49.2%
       Number of tests that were (+)395198197
       Percentage of all (+) tests50.1%49.9%
       Percentage of tests performed that were (+)11.9%11.7%12.0%
      Subsequent EIA for Toxin Antigen
       Number of tests performed395198197
       Number of tests that were (+)1185563
       Percentage of PCR(+)/EIA(+) tests46.6%53.4%
       Percentage of PCR(+) tests that were EIA(+)29.9%27.8%32.0%
      PCR, polymerase chain reaction; EAI, enzyme immunoassay; (+), positive.
      low asterisk The date of admission is considered hospital day 1.
      The detailed data for these 12 months also provided important process measures. The mean time from stool collection to report of the PCR result was ∼7.5 hours with a median of ∼6 hours (data not shown). Moreover, for tests that were PCR(+), the reflexed EIA took only a mean of 68 additional minutes (median 39 minutes) to be reported. Finally, comparing PCR(+)/EIA(+) and PCR(+)/EIA(-) patients revealed no significant differences in the percentage who received inpatient antibiotics in the previous 60 days (78.0% vs 77.3%), proton-pump inhibitors or histamine-2 receptor blockers in the previous 7 days (47.5% vs 58.1%), or laxatives in the previous 3 days (33.1% vs 38.6%).

      Discussion

      Although national data indicated that the number of CDI cases in hospitals had been decreasing,
      • Guh AY
      • Mu Y
      • Winston LG
      • et al.
      Trends in U.S. burden of Clostridioides difficile infection and outcomes.
      our medical center was seeing an increasing number of cases that met the NHSN definition for HAI-CDI (Fig 1A). Though some of the rise could be attributed to increasing patient-days and there was an association with testing methodology, the greatest increase was associated with the new reporting requirements in January 2015, when cases could no longer be excluded as POA based on documentation that symptoms were present just before or soon after admission (Fig 1B). Despite these explanations and the knowledge that our SIR for HO-CDI had been consistently ≤1 (Fig 1C), we took several new actions aimed at minimizing the case count (see Methods).
      The first intervention was a nurse-driven protocol to identify CDI that was POA by asking a simple question of all admissions. Others have reported that empowering nurses to order CDI testing upon documentation of diarrhea during the first 3 hospital days can reduce the number of HO-CDI cases and improve the SIR.
      • Bruno-Murtha LA
      • Osgood RA
      • Alexandre CE
      A successful strategy to decrease hospital-onset Clostridium difficile.
      However, rather than merely shifting some patients out of the nosocomial classifications, we emphasized two clinical benefits of our admission protocol: (1) the patient-care value of expediting identification of someone who would benefit from treatment, and (2) the infection-control value implicit in our presumption that patients with diarrhea who have a positive PCR test require isolation even if their diarrhea was not due to CDI. Thus, we set a low threshold by asking “Has the patient had 1 or more liquid or watery stools in the last 24 hours?” Given the intended high sensitivity, clinicians were asked not to override the automatic order for CDI testing, even if they thought their patient had diarrhea from another cause. We recognize that this approach runs counter to diagnostic stewardship which, for example, would defer CDI testing in someone who recently ingested laxatives.
      • McDonald LC
      • Gerding DN
      • Johnson S
      • et al.
      Clinical practice guidelines for Clostridium difficile infection in adults and children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA).
      ,
      • Howard-Anderson JR
      • Sexton ME
      • Robichaux C
      • et al.
      The impact of an electronic medical record nudge on reducing testing for hospital-onset Clostridioides difficile infection.
      ,
      • Rock C
      • Maragakis LL.
      Diagnostic stewardship for Clostridiodes difficile testing: from laxatives to diarrhea and beyond.
      However, even patients who have taken laxatives may have active CDI.
      • White NC
      • Mendo-Lopez R
      • Papamichael K
      • et al.
      Laxative use does not preclude diagnosis or reduce disease severity in Clostridiodes difficile infection.
      Moreover, we decided that judicious testing overlooked the potential infection control value discussed above. We should underscore that this approach does not mean that we screen for asymptomatic carriers who do not have diarrhea.
      • Collison M
      • Murillo C
      • Marrs R
      • et al.
      Universal screening for Clostridioides difficile at an urban academic medical center.
      ,
      • Barker AK
      • Alagoz O
      • Safdar N.
      Interventions to reduce the incidence of hospital-onset Clostridium difficile infection: An agent-based modeling approach to evaluate clinical effectiveness in adult acute care hospitals.
      Finding that 46.6% of all PCR(+)/EIA(+) cases were from stool collected on HD 1 or 2 (Table 2) supports our hypothesis that many cases were POA but not being identified as such after the reporting rules became more stringent in 2015. This percentage exceeds the 25% hypothesized in our model (Fig 2), likely because it includes overt cases that would have been captured as POA even before this admission tool. An internal assessment also revealed that the admission question accounted for 16 (29.1%) of the 55 patients who were PCR(+)/EIA(+) on HD 1 or 2 (data not shown), further supporting continued use of this tool.
      Because testing everyone who had diarrhea in the 24 hours before admission could lead to overdiagnosis, we also instituted a 2-step testing algorithm. A large study had reported that only 45% of patients who were PCR(+) were also EIA(+), with PCR(+)/EIA(-) patients — despite most getting no treatment — having outcomes similar to PCR(-) patients.
      • Polage CR
      • Gyorke CE
      • Kennedy MA
      • et al.
      Overdiagnosis of Clostridium difficile infection in the molecular test era.
      Thus, although our policy is to isolate PCR(+)/EIA(-) patients, many of them do not need treatment. However, the decision to treat is up to the clinician, given that some patients in the cited report were treated, the EIA is less sensitive than PCR, and a later study suggested these conclusions may not apply to patients with hematologic malignancies.
      • Ziegler M
      • Landsburg D
      • Pegues D
      • et al.
      Clinical characteristics and outcomes of hematologic malignancy patients with positive Clostridium difficile toxin immunoassay versus polymerase chain reaction test results.
      That only ∼30% of PCR(+) stools in our study were EIA(+), similar to the percentage suggested by another report
      • McCauley BP
      • Evans ME
      • Simbartl LA
      • Gamage SD
      • Kralovic SM
      • Roselle GA.
      Effect of testing methods on incidence of Clostridioides difficile infection rates in Veterans' Affairs medical centers.
      and lower than the 50% hypothesized in our model, bolsters our practice of not referring to the EIA as a confirmatory test. Nonetheless, this 2-step algorithm has reduced the number of people unnecessarily treated for CDI. It also has the benefit that as of January 2018, facilities performing a multi-step algorithm must use the final test in the sequence to decide whether a patient meets the NHSN's laboratory definition for a CDI.
      National Healthcare Safety Network, Centers for Disease Control and Prevention
      2018 Changes to CDI LabID Event Reporting & Analysis.
      It has not escaped our notice that had we not instituted the 2-step algorithm in July 2018, the 197 PCR(+) samples from HD 3 or later would have been counted as HAI-CDI, along with some of the PCR(+) cases from HD 1 or 2 that might not have been tested as early without the admission question. Thus, the number of HAI-CDI cases might have exceeded the 161 cases of 2017-2018 (Fig 3). However, clinicians may have lowered their threshold for testing on later hospital days due to messaging that even 1 diarrheal stool should trigger testing on admission.
      • Bruno-Murtha LA
      • Osgood RA
      • Alexandre CE
      A successful strategy to decrease hospital-onset Clostridium difficile.
      They might also have been more willing to test knowing the EIA could help guide their treatment decisions. This reasoning is consistent with internal data showing that the number of inpatient PCR tests increased from 2,254 in 2017-2018 to 3,330 in 2018-2019, while the percentage that were positive fell from ∼15% (extrapolated from a contemporaneous analysis of 2 of those months, data not shown) to 11.9% (Table 2). Overall, these data underscore the value of linking the 2 initiatives of an admission question about diarrhea and a 2-step testing algorithm.
      Although our intent was to reduce the number of HAI-CDI cases, we also observed a decrease in the risk-adjusted SIR reported by the CDC from the HO-CDI data (Fig 3). While it is possible that other interventions described in the Methods contributed to the marked and durable decreases in the HAI-CDI case count and the calculated HO-CDI SIR, none of those other changes had the systematic implementation or careful data collection as the 2 major interventions detailed in this report, and they were not expected to have as large an impact. Moreover, in-depth analysis of the 2 major interventions demonstrated that the number of incident cases of CDI in hospitalized patients can be dramatically decreased by implementing an admission screening protocol that inquires about diarrhea and the use of a 2-step testing algorithm.

      Author contributions

      MK was involved in the initial concept, accessing and interpreting the data, creating the figures and tables, and drafting the manuscript. ACC was involved in data acquisition and interpretation for Table 2, and the statistical analyses. PEH was involved in data acquisition and analysis for Figures 1 and 3, and the classification and terminology in Table 1. WHG was involved in laboratory methodology. SMR was involved in implementing the nursing admission protocol. MAW, FDG, and MFL were involved in automating acquisition of the data. CMH was involved in the study design and supervision. All authors reviewed drafts and approved a final version of the manuscript.

      Acknowledgments

      We thank Scott Mincemoyer, Jeri Anderson, and Benjamin von Bredow for help with hospital data, and Mark Lesher, Desiree Albright, and Lauren Allison for support of this project. We also appreciate all the nurses and physicians without whose cooperation these interventions would not have been successful.

      References

        • Guh AY
        • Mu Y
        • Winston LG
        • et al.
        Trends in U.S. burden of Clostridioides difficile infection and outcomes.
        N Engl J Med. 2020; 382: 1320-1330
        • Dellit TH
        • Owens RC
        • McGowan Jr., JE
        • et al.
        Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship.
        Clin Infect Dis. 2007; 44: 159-177
        • Barlam TF
        • Cosgrove SE
        • Abbo LM
        • et al.
        Implementing an antibiotic stewardship program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America.
        Clin Infect Dis. 2016; 62: e51-e77
        • Centers for Disease Control and Prevention
        2017 National and State Healthcare-Associated Infections Progress Report.
        CDC, Atlanta2021
        • National Healthcare Safety Network, Centers for Disease Control and Prevention
        Multidrug-Resistant Organism & Clostridioides difficile Infection (MDRO/CDI) Module [Chapter 12].
        National Healthcare Safety Network (NHSN) Patient Safety Component Manual, Atlanta2022 (Accessed December 8, 2021)
        • Katzman M
        • Kim J
        • Lesher MD
        • et al.
        Customizing an electronic medical record to automate the workflow and tracking of an antimicrobial stewardship program.
        Open Forum Infect Dis. 2019; 6: 1-9
        • National Healthcare Safety Network, Centers for Disease Control and Prevention
        CDC/NHSN Surveillance Definitions for Specific Types of Infections [Chapter 17].
        National Healthcare Safety Network (NHSN) Patient Safety Component Manual, Atlanta2022 (Accessed December 8, 2021)
        • National Healthcare Safety Network, Centers for Disease Control and Prevention
        Identifying Healthcare-associated Infections (HAI) for NHSN Surveillance [Chapter 2].
        National Healthcare Safety Network (NHSN) Patient Safety Component Manual, Atlanta2022 (Accessed December 8, 2021)
        • Kelly CP
        • LaMont JT.
        Clostridium difficile infection.
        Annu Rev Med. 1998; 49: 375-390
        • Polage CR
        • Gyorke CE
        • Kennedy MA
        • et al.
        Overdiagnosis of Clostridium difficile infection in the molecular test era.
        JAMA Intern Med. 2015; 175: 1792-1801
        • McDonald LC
        • Gerding DN
        • Johnson S
        • et al.
        Clinical practice guidelines for Clostridium difficile infection in adults and children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA).
        Clin Infect Dis. 2018; 66: 987-994
        • Louie TJ
        • Miller MA
        • Mullane KM
        • et al.
        Fidaxomicin versus vancomycin for Clostridium difficile infection.
        N Engl J Med. 2011; 364: 422-431
        • Cornely OA
        • Crook DW
        • Esposito R
        • et al.
        Fidaxomicin versus vancomycin for infection with Clostridium difficile in Europe, Canada, and the USA: a double-blind, non-inferiority, randomised controlled trial.
        Lancet Infect Dis. 2012; 12: 281-289
        • Janarthanan S
        • Ditah I
        • Adler DG
        • Ehrinpreis MN.
        Clostridium difficile-associated diarrhea and proton pump inhibitor therapy: a meta-analysis.
        Am J Gastroenterol. 2012; 107: 1001-1010
        • Goldenberg JZ
        • Yap C
        • Lytvyn L
        • et al.
        Probiotics for the prevention of Clostridium difficile-associated diarrhea in adults and children.
        Cochrane Database Syst Rev. 2017; 12CD006095
        • Carignan A
        • Poulin S
        • Martin P
        • et al.
        Efficacy of secondary prophylaxis with vancomycin for preventing recurrent Clostridium difficile infections.
        Am J Gastroenterol. 2016; 111: 1834-1840
        • Van Hise NW
        • Bryant AM
        • Hennessey EK
        • Crannage AJ
        • Khoury JA
        • Manian FA.
        Efficacy of oral vancomycin in preventing recurrent Clostridium difficile infection in patients treated with systemic antimicrobial agents.
        Clin Infect Dis. 2016; 63: 651-653
        • Lalande V
        • Barrault L
        • Wadel S
        • Eckert C
        • Petit JC
        • Barbut F.
        Evaluation of a loop-mediated isothermal amplification assay for diagnosis of Clostridium difficile infections.
        J Clin Microbiol. 2011; 49: 2714-2716
        • National Healthcare Safety Network, Centers for Disease Control and Prevention
        NHSN HAI Surveillance Changes for 2015.
        NHSN e-News. 9. CDC, Atlanta2014 (Accessed February 11, 2022)
        • England M
        • Erdman P
        • Greene W.
        Evaluation of two molecular assays for the detection of Clostridium difficile.
        in: ASM Microbe 2017, the Annual Meeting of the American Society for Microbiology. New Orleans, LA. 2017: 5362 (Abstract)
        • Eyre DW
        • Cule ML
        • Wilson DJ
        • et al.
        Diverse sources of C. difficile infection identified on whole-genome sequencing.
        N Engl J Med. 2013; 369: 1195-1205
        • Curry SR
        • Muto CA
        • Schlackman JL
        • et al.
        Use of multilocus variable number of tandem repeats analysis genotyping to determine the role of asymptomatic carriers in Clostridium difficile transmission.
        Clin Infect Dis. 2013; 57: 1094-1102
        • Bruno-Murtha LA
        • Osgood RA
        • Alexandre CE
        A successful strategy to decrease hospital-onset Clostridium difficile.
        Infect Control Hosp Epidemiol. 2018; 39: 234-236
        • Howard-Anderson JR
        • Sexton ME
        • Robichaux C
        • et al.
        The impact of an electronic medical record nudge on reducing testing for hospital-onset Clostridioides difficile infection.
        Infect Control Hosp Epidemiol. 2020; 41: 411-417
        • Rock C
        • Maragakis LL.
        Diagnostic stewardship for Clostridiodes difficile testing: from laxatives to diarrhea and beyond.
        Clin Infect Dis. 2020; 71: 1479-1480
        • White NC
        • Mendo-Lopez R
        • Papamichael K
        • et al.
        Laxative use does not preclude diagnosis or reduce disease severity in Clostridiodes difficile infection.
        Clin Infect Dis. 2020; 71: 1472-1478
        • Collison M
        • Murillo C
        • Marrs R
        • et al.
        Universal screening for Clostridioides difficile at an urban academic medical center.
        Infect Control Hosp Epidemiol. 2021; 42: 351-352
        • Barker AK
        • Alagoz O
        • Safdar N.
        Interventions to reduce the incidence of hospital-onset Clostridium difficile infection: An agent-based modeling approach to evaluate clinical effectiveness in adult acute care hospitals.
        Clin Infect Dis. 2018; 66: 1192-1203
        • Ziegler M
        • Landsburg D
        • Pegues D
        • et al.
        Clinical characteristics and outcomes of hematologic malignancy patients with positive Clostridium difficile toxin immunoassay versus polymerase chain reaction test results.
        Infect Control Hosp Epidemiol. 2018; 39: 863-866
        • McCauley BP
        • Evans ME
        • Simbartl LA
        • Gamage SD
        • Kralovic SM
        • Roselle GA.
        Effect of testing methods on incidence of Clostridioides difficile infection rates in Veterans' Affairs medical centers.
        Infect Control Hosp Epidemiol. 2021; 42: 461-463
        • National Healthcare Safety Network, Centers for Disease Control and Prevention
        2018 Changes to CDI LabID Event Reporting & Analysis.
        NHSN e-News. 12. CDC, Atlanta2017 (Accessed February 22, 2022)