A nationwide Clostridium difficile infection (CDI) prevention initiative was implemented in Veterans Health Administration (VA) acute-care facilities in July 2012 because C. difficile had replaced methicillin-resistant Staphylococcus aureus (MRSA) as the most common healthcare-associated infection.Reference Evans, Kralovic, Simbartl, Jain and Roselle 1 This prevention initiative was based on Society for Healthcare Epidemiology of American (SHEA)/Infectious Disease Society of America (IDSA) guidelinesReference Cohen, Gerding and Johnson 2 and was patterned after the VA’s successful efforts to reduce MRSA healthcare-associated infections. This initiative included a 4-part bundle emphasizing (1) environmental management, (2) hand hygiene, (3) contact precautions for suspected or documented CDI cases, and (4) an institutional culture change in which infection control becomes everyone’s business. Analysis of data from the first 33 months of this program showed that the rate of hospital-acquired cases had decreased 15%.Reference Evans, Kralovic, Simbartl, Jain and Roselle 1
A nationwide guideline for the prevention of C. difficile infections in VA long-term care facilities (LTCFs) was implemented in February 2014 because VA LTCFs are often closely linked to acute-care facilities. Additionally, CDI has been associated with increased morbidity and mortality, and it disproportionally affects individuals ≥65 years old,Reference Karanika, Grigoras and Flokas 3 which constitute the major proportion of residents in VA LTCFs. These facilities typically provide skilled nursing services ranging from short-term rehabilitation to long-term care for dementia. Recommendations for the prevention of CDI in LTCFs were similar to those for acute care but included approaches to meeting challenges inherent in providing robust infection control while maintaining a homelike environment (see Supplemental Materials). Here, we report CDI cases over the first 33 months of the LTCF initiative.
METHODS
Clostridium difficile infection case definitions, data collection, data evaluation, and reporting were performed exactly as described previously.Reference Reeves, Evans and Simbartl 4 CDI cases were identified by a positive diagnostic laboratory test (LabID event) as recommended by the National Healthcare Safety Network (NHSN). 5 Trends in clinically confirmed LTCF-onset cases, defined as a resident with clinical evidence of illness (ie, diarrhea or histopathologic or colonoscopic evidence of pseudomembranous colitis) and a nonduplicate, nonrecurrent positive diagnostic laboratory test collected >48 hours after admission, were evaluated for this report. Diarrhea was defined as stool that took the shape of its container, and our guideline recommended that specimens be submitted for testing only if the resident had ≥3 liquid stools within 24 hours. Approximately 90% of LTCF-onset cases met the clinically confirmed definition.
We analyzed prospective data that had been entered each month from April 1, 2014, through December 31, 2016, by multidrug-resistant organism (MDRO) prevention coordinators at each of the 132 VA LTCF reporting sites into a centralized database housed at the Inpatient Evaluation Center in Cincinnati, Ohio.Reference Reeves, Evans and Simbartl 4 Data were aggregated quarterly, beginning in April 2014, the first complete quarter following guideline implementation. Quarterly CDI trends were evaluated using a negative binomial regression model, which accounted for resident days as an offset variable and included admission prevalence and diagnostic test type in the model. These trends were compared to the 24-month baseline period before implementation of the LTCF Initiative.Reference Reeves, Evans and Simbartl 4 The quarterly CDI admission prevalence rate was calculated as the number of nonduplicate (including recurring) positive LabID events collected ≤24 hours before to ≤48 hours after admission divided by the number of admissions to the LTCF for the quarter ×100. All statistical analyses were performed using SAS version 9.3 software (SAS Institute, Cary, NC). The analysis of these data was approved by the Cincinnati VA Medical Center Institutional Review Board.
RESULTS
During the 33-month analysis period, there were 137,289 admissions, 9,288,098 resident days, and 1,373 clinically confirmed LTCF-onset CDI cases. The LTCF CDI admission prevalence rate was ~0.38 cases per 100 admissions at the beginning of the analysis period and did not significantly change during the study period (P=.38, linear regression). The percent of facility laboratories using a nucleic acid amplification test for CDI diagnostic testing was 81.6% in 2015 and 83.5% in 2017, which was not significantly different (P=.95, χ2 test). The nationwide number of clinically confirmed LTCF-onset CDI cases did not change in the 2 years prior to implementation of the LTCF CDI prevention initiative in February 2014 but decreased 36.1% in the subsequent 33-month analysis period (P=.0009 for trend) (Figure 1).
FIGURE 1 Quarterly number of Veterans Affairs clinically-confirmed long-term care facility onset Clostridium difficile infections comparing the baseline period (April 2012 through March 2014) to the postimplementation period beginning in April 2014, which was the first complete quarter following guideline implementation. Trends with P values for both periods were evaluated using negative binomial regression accounting for resident days, admission prevalence, and diagnostic test type.
DISCUSSION
This report, using a negative binomial model that accounted for admission prevalence, resident-days, and diagnostic test type, shows that VA clinically confirmed LTCF-onset CDI cases declined significantly over 33 months coincident with the implementation of a prevention initiative based on SHEA/NHSN guidelines and tailored to long-term care. This result mirrors the experience in VA acute-care facilities using a similar prevention initiative.Reference Evans, Kralovic, Simbartl, Jain and Roselle 1
Although others have shown that LTCF residents have an increased attributable burden of CDI compared to those who live in the community,Reference Karanika, Grigoras and Flokas 3 data on actual rates are sparse. Initial case rates ranging from 0.8 to 2.9 per 10,000 resident days have been reported.Reference Campbell, Giljahn and Machesky 6 , 7 The pooled VA rate of 1.98 clinically confirmed LTCF-onset CDI cases per 10,000 resident days in April 2014 (when the initiative was implemented) is consistent with these numbers, and the decreased pooled rate of 1.26 cases per 10,000 resident days at the end of our analysis period shows significant improvement based on our regression model.
Strengths of this report include (1) the large database from 132 LTCFs over a 33-month period where a standardized prevention initiative was implemented, (2) the use of a positive LabID event for identification of CDI cases thereby limiting bias in case inclusion, and (3) the use of a preformatted spreadsheet provided by the MDRO Prevention Office, which automatically categorizes cases as clinically confirmed LTCF-onset based on resident demographics and timing of the CDI LabID event.
The definition of LTCF-onset as a LabID event occurring >48 hours after admission may be questioned because there is controversy over whether LTCF CDI cases might actually originate in acute care. The NHSN adopted a surveillance category of “acute-care transfer to long-term-care facility onset” defined as a CDI case diagnosed ≤4 weeks after transfer to account for this possibility. 5 However, recent VA data based on culture surveillance suggest that most initial CDI cases occurring in residents within 1 month of transfer are due to organisms acquired in the LTCF.Reference Ponnada, Guerrero and Jury 8
The exact reason for the decrease in cases within the VA LTCFs is not known. Given the large number of facilities involved and the long observation period, we were not able to collect data on individual facility activities or sustainability of activities; hence, we cannot report a “magic bullet” responsible for the declining trend. As we outlined in our report detailing a significant decrease in CDI rates in the acute-care setting,Reference Evans, Kralovic, Simbartl, Jain and Roselle 1 the decline may be multifactorial. In the VA’s case, these factors most likely include (1) strong leadership from the VA Central Office, (2) individual facility accountability through the collection of monthly CDI case rates, (3) the national dissemination and rigorous application of SHEA/NHSN recommendations tailored to long-term care, (4) continuous education to the field provided by the MRSA/MDRO Office, (5) strong antimicrobial stewardship programs,Reference Kelly, Jones and Echevarria 9 and (6) the application of consistent practices in different components of the healthcare system that VA residents frequent. This hypothesis is consistent with the concept that facilities that work together achieve better infection control than those that act alone.Reference Lee, Bartsch and Wong 10 Thus, decreases in the LTCFs may reflect decreases in acute care. Although this report included data from veterans who were mainly men, the general approach to the prevention of CDI recommended by SHEA/NHSN tailored to long-term care may be applicable to the general US nursing home population.Reference Reeves, Evans and Simbartl 4
ACKNOWLEDGMENTS
We thank the VA Under Secretary for Health, the Deputy Under Secretary for Health for Policy and Services, the CDI Data Work Group, the VA MRSA/MDRO Task Force, the MRSA/MDRO Prevention Coordinators, infection prevention and control professionals, infectious diseases specialists, and clinical laboratory personnel at each facility for support of the MDRO prevention initiatives and their hard work and dedication toward improving the health care of America’s veterans. Thanks are also due to Linda Flarida of the MDRO Prevention Office for help with data validation and cleaning.
Financial support: No financial support was provided relevant to this article.
Potential conflicts of interest: All authors report no conflicts of interest relevant to this article.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2017.305
