Background
Clostridioides difficile infection (CDI) affects 500,000 people and contributes to 30,000 deaths annually.1 Risk factors include prolonged use of antibiotics or gastric acid reducers, advanced age, frequent healthcare encounters, and chronic gastrointestinal (GI) disease.Reference Russo, Kuntz, Yu, Smith, Hauser and Halchenko2,Reference Song and Kim3 These factors also increase the risk for asymptomatic colonization, for which clinical guidelines recommend against treatment.Reference Johnson, Lavergne, Skinner, Gonzales-Luna, Garey and Kelly4 The Department of Veterans Affairs (VA) has issued guidelines to improve CDI diagnostic stewardship. These guidelines discourage repeat testing within 7 days of a previous test for the same episode of diarrheal illness and specifically discourage repeat testing as a “test of cure” after treatment.Reference McDonald, Gerding, Johnson, Bakken, Carroll and Coffin5 The goal of this retrospective cohort study was to examine the rates of 7-day repeat testing within the VA healthcare system.
Methods
Study design and patient population
This was a retrospective cohort of patients who had at least one CDI test ordered between January 1, 2019, and December 31, 2022. Only tests performed in inpatient and long-term care settings were included. Data was collected from eight VA Medical Centers (VAMC) in Illinois, Wisconsin, and Michigan.
Data collection
The VA corporate data warehouse (CDW) is a national repository of clinical and administrative data, which was used to obtain the following patient information: birth date, CDI test date, race, ethnicity, sex, admission date, discharge date, facility complexity, and facility urban/rural status. Comorbidities in the previous 365 days before the CDI test were identified by using ICD-10 codes to calculate the Charlson Comorbidity Score. The degree and title/classification of the ordering provider were collected. Based on their degree and title/classification, providers were categorized as resident physicians, attending physicians, or advanced practice practitioners (APPs).
7-day repeat definition
A two-step testing algorithm was used to diagnose CDI at all included facilities, which involved various combinations of the following tests: nucleic acid amplification tests for Toxin A and B genes, Toxin Enzyme Immunoassays, and Glutamate dehydrogenase tests. An infectious disease clinician defined test results as either positive, negative, or cancelled. If both tests were positive, the test was considered positive. If both tests were negative, the test was considered negative. If one test resulted positive and one was negative, then the test was considered negative. If one or both tests were cancelled, then the test was defined as cancelled. A seven-day repeat test was a CDI test ordered for the same patient within seven days of a previous test, regardless of the result of either test.
Statistical analysis
Unadjusted analyses were conducted using chi-squared and t-tests, or their non-parametric equivalents, to determine factors associated with repeat testing. Binomial logistic regression models included variables that were significant at the p < 0.05 level. To account for procedural differences across facilities, as well as differences in provider demographics, as smaller VAs often do not have residency programs, models were clustered at the facility level. Only variables that remained significant in the adjusted model were retained in the final parsimonious model. All statistical analyses were completed using SAS 9.4.
Results
There were 6584 tests for 4640 unique patients identified for the study period. The overall cohort was primarily White (61.2%), male (95.7%), older (mean = 70.2 years, SD = 11.8), and treated at an urban VAMC (88.9%). The volume of CDI tests ordered at each medical center across the four-year study period varied considerably from 77 to 2116. The frequency of 7-day repeats ranged from 0% to 5% (p-value = 0.0003). There was also a wide variation in the percentage of tests ordered by each provider classification by VAMC. The percentage of tests ordered by residents ranged from 0% to 93%, by attendings from 6% to 82%, and for APPs from 1% to 34% (Figure 1).
CDI Tests ordered by facility: the number of tests ordered by each provider type by facility over the four-year study period.
There were 294 (4.5%) tests defined as 7-day repeats (Table 1). There was no association between repeat testing and patient race/ethnicity, age, sex, Charlson score, or urban/rural geography. Repeat tests had 83% higher odds of being ordered for a patient in long-term care compared to those ordered in acute care (OR = 1.83 (95% CI 1.52,2.22)) (Table 1). Tests that were 7-day repeats were three times more likely to be negative (OR = 3.62 (1.86, 7.04)) and eight times more likely to be cancelled (OR = 8.52 (3.32, 21.8)). Compared to residents, attending physicians (OR = 0.67 (0.52, 0.86)) and APPs (OR = 0.61 (0.34, 1.09)) were less likely to order repeat testing. There was a 2.1% conversion rate from negative to positive between initial and repeat tests and a 6.3% rate from cancelled to positive repeat (Supplemental Table 1).
Factors associated with 7-day repeat testing (n = 6584 tests)
Discussion
In this evaluation of over 4600 patients who had at least one CDI test ordered, we found the overall frequency of 7-day repeat testing was low (4.5%) compared to other studies that have reported repeat testing rates ranging from 11% to 31%.Reference Kelly, Yarrington, Zembower, Sutton, Silkaitis and Postelnick6,Reference Friedland, Brown, Glick, Lusby, Lemkin and Leekha7 This finding suggests strong adherence to VA diagnostic stewardship guidance. Care settings (long-term and intensive) and provider type (resident vs. attending/APP) were found to be significantly associated with repeat testing. Finally, repeat tests were three times more likely to be negative and eight times more likely to be cancelled, strengthening the argument that repeat tests are low-value and should continue being de-implemented in the healthcare setting.
Previous studies have found similar associations between care setting and repeat CDI testing. One evaluation of over 75,000 CDI orders found a 30% increased likelihood of inappropriate testing (for several reasons, including repeat testing) in the intensive care setting compared to general medicine wards.Reference Axenfeld, Greendyke, Li, Green, Whittier and Freedberg8 Providers in long-term and intensive care settings may be more concerned about CDI risks and complications in their patients. They may have greater misperceptions about the benefit of repeat CDI testing. Targeted education tailored to the needs of these care settings could be beneficial in decreasing inappropriate testing. Furthermore, the finding that residents were more likely to order repeat CDI tests compared to attendings or APPs suggests the need for educational interventions for trainees.
This study has several strengths. The large sample size allowed for a robust evaluation of the research question. Clustering helped account for the variation between facilities. Additionally, VA’s extensive data sources enabled the analysis of patient, provider, and facility factors associated with repeat testing.
However, the study has some limitations. First, the retrospective nature of the data could have allowed for some misclassification of factors. Although the missing data was minimal, there may have been bias due to data exclusion from the cohort. Additionally, repeat testing could be clinically appropriate in certain situations; this could not be assessed in this study, as such determinations would require a chart review, which was outside the scope of this project. The reason for admission could be associated with CDI testing, but it was not collected. However, the overall health of patients was assessed via the Charlson comorbidity index and was associated with repeat testing. Any ongoing CDI or diagnostic stewardship efforts occurring at individual facilities were not captured. Finally, the VA population is predominantly older and male; therefore, the results of this study may not be generalizable.
Conclusion
This study found that 7-day repeat testing for CDI was a rare occurrence in the VA (4.5%). Care setting and provider type were associated with an increased likelihood of repeat testing. Future work should continue to identify factors that contribute to repeat testing and design prevention interventions for inappropriate testing.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2025.10267.
Financial support
This work was funded by the Department of Veterans Affairs, Quality Enhancement Research Initiative QUE 20-016, and Research Career Scientist RCS 20-192. The views expressed in this manuscript are the views of the authors and do not reflect those of the U.S. Government or the Department of Veterans Affairs.
Competing of interests
The authors declare that they have no conflicts of interest.
Open access
