The optimal timing of blood culture (BCx) sets collection has not been evaluated with continuous BCx detection systems. The yield of BCx was similar between short intervals (median, 3 minutes) and longer intervals (median, 16 or 43 minutes) among 5,856 BCx, except for improved polymicrobial bacteremia detection with long-interval BCx.

Background:Clostridioides difficile infection (CDI) is a leading cause of healthcare-associated infection and is associated with increased morbidity and mortality. Multiple strategies have been implemented to optimize the diagnostic accuracy of CDI testing algorithms. However, overdiagnosis of C. difficile colonization remains a challenge especially in the era of highly sensitive Nucleic acid amplification testing (NAAT). We implemented 2 interventions to reduce the rates of inappropriate C. difficile orders and tests. Methods: We performed a quasi-experimental retrospective study to examine the rates of all inpatient C. difficile test orders and results relative to 2 interventions between January 2018 and February 2021. We defined 3 periods: preintervention, after the first intervention, and after the second intervention. The first intervention, implemented May 2019, was a clinical decision support system (CDSS) tool to guide clinicians to order testing only if CDI criteria were met. The second intervention, implemented July 2020, was the requirement of mandatory antimicrobial team approval of PCR reflex testing for indeterminate toxin or antigen test results. This intervention included a discussion between clinicians and members of stewardship team prior to approval. To evaluate the impact of interventions on ordering appropriateness, chart review was conducted on a random subsample of 100 orders from each period. Hospital-onset CDI (HO-CDI) was calculated using CDC criteria. Results: In total, 3,004 C. difficile test orders were placed during the study period. The rates of reportable HO-CDI were significantly reduced by 57.1% (P = .003). We detected a significant reduction in the number of tests ordered over time from 11.6 to 7.51 per 1,000 patient days. (p Conclusions: CDSS tools target patients at high pretest probability of CDI. The restriction of PCR-reflex testing to the antimicrobial stewardship team is a novel effective measure to minimize the misdiagnosis of CDI. The incorporation of multiple strategies is necessary to improve the diagnostic accuracy of C. difficile testing.

Funding: None

Disclosures: None

Background: CDI is the single most common cause of nosocomial diarrhea in both adults and children. Available data regarding treatment outcomes in hospitalized children remain limited. CDI recurrence in children has been reported in 20%–30% of cases. Consensus regarding the best testing method for CDI is lacking. The 2018 IDSA guideline recommends a multistep algorithm with detection of glutamate dehydrogenase antigen plus toxin, followed by detection of toxigenic C. difficle with nucleic acid amplification test (NAAT) if results are discordant. Methods: We included patients aged 1–26 years admitted from July 2020 through June 2021 with CDI symptoms and positive toxin or NAAT. Healthcare facility-onset CDI (HO-CDI) was defined as positive specimen collected >3 days after admission. Community-onset CDI (CO-CDI) was defined as positive specimen collected ≤3 days after admission. Community-onset healthcare facility-associated CDI (CO-HCFA-CDI) was defined as positive specimen from a patient who was discharged from the facility ≤4 weeks prior. Recurrence was defined as an episode of CDI occurring within 60 days after onset of a previous infection. Results: Mean age of the 63 patients meeting inclusion criteria was 11.2 years (range, 1–21 years). Most patients (n = 37; 58.7%) were male, tested negative for C. difficile toxins (n = 39; 61.9%), and had mild-to-moderate disease (n = 61; 96.8%). Patients with immunocompromising conditions were common, including malignancy (n = 38; 60.3%), inflammatory bowel disorder (n = 8; 12.7%), and history of solid organ transplant (n = 5; 7.9%). Previously healthy without chronic medical conditions were uncommon (n = 4; 6.3%). CO-CDI was most common (n = 26; 41.3%) followed by HO-CDI (n = 23; 36.5%). Also, 34 patients (53.9%) were exposed to antibiotics within the previous 30 days, 16 (47.0%) of whom received 2 or more antibiotics. Sulfamethoxazole–trimethoprim was the most prescribed agent (13; 38%), most (12; 92.3%) as prophylaxis for Pneumocystis jirovecii pneumonia. Furthermore, 42 patients (66.7%) were receiving gastric acid suppressant agents. Laxatives were given to 14 patients (22.2%) within 72 hours of testing, despite electronic reminders. Most were treated with oral vancomycin (n = 46; 73.0%). In addition, 5 patients (7.9%) did not receive CDI treatment at the discretion of the treating physician; all were toxin negative. CDI was cured in 58 patients (92.1%) with only 5 (7.9%) experiencing recurrence infection. Patients testing positive for C. difficile toxin were more likely to experience infection recurrence compared to those with a negative toxin screen: 4 of 24 (16.7%) versus 1 of 39 (2.6%) (P = .044). Conclusions: Most patients with CDI were treated with oral vancomycin at our institution. We observed significantly lower rate of recurrence than previously reported. Toxin-positive patients experienced higher recurrence rate. Prospective studies are needed to confirm our findings.

Funding: None

Disclosures: None

Background:Clostridioides difficile infection (CDI) is the most common healthcare-associated infection (HAI) in the United States. Healthcare facility-onset (HO) CDI reporting is a laboratory-identified (LabID) event and does not rely on symptoms. Inappropriate testing can lead to overdiagnosis in patients who are colonized, especially in those receiving promotility agents. Approximately 45% of HO-CDI cases at our institution occurred in the setting of laxative use in 2019. We assessed the effectiveness of an electronic medical record (EMR) “hard stop” in reducing inappropriate CDI testing and its impact on HO-CDI rates. Methods: We conducted a pre–post quasi-experimental retrospective study comparing test order rates per 1,000 patient days, CDI rate per 1,000 patient days, and standardized infection ratio (SIR) in the preintervention period (January 2018–December 2019) to the intervention period (April 2020–September 2021), at a 5-hospital healthcare system in southeastern Michigan. In February 2020, we implemented a hard stop in Epic that was triggered >3 days after admission for the following criteria: patients <1 year of age; repeated testing within 7 days, and receipt of promotility agents within 48 hours. After discontinuing the promotility agents for at least 48 hours, providers were allowed to place an order if diarrhea persisted. The medical director of infection prevention and control or designee had the ability to override the hard stop when deemed necessary after reviewing the case upon provider request. All orders expired after 24 hours if a specimen was not collected. We retrospectively reviewed the number of overrides after the intervention to determine the positivity rate. Results: Our CDI rates per 1,000 patient days were 3.21 in the preintervention period and 1.48 in the postintervention period, a 54% reduction (Fig. 1). The test order rates were 119.4 in the preintervention period and 87.7 in the postintervention period, a 26.5% reduction (Fig. 2). The SIR decreased from 0.542 in the preintervention period to 0.361 in the postintervention period, a 33% reduction (95% CI, 0.54–0.82; P = .0001). After the intervention, 299 patients had an override. Of these, samples from 218 patients (72.9%) were negative, 50 orders (16.7%) were cancelled, and 28 samples (9%) were positive. Conclusions: Diagnostic stewardship, utilizing an electronic hard stop, was effective in reducing inappropriate C. difficile testing in the setting of promotility agents without delaying diagnosis of HO-CDI. This strategy combined with standard best practices can significantly reduce HO-CDI rates.

Funding: None

Disclosures: None

Antimicrobial use during the coronavirus disease 2019 (COVID-19) pandemic at a tertiary-care center was analyzed using interrupted time-series analysis. Among intravenous antimicrobials, the use of azithromycin and third-generation cephalosporins significantly decreased during the current pandemic. Similarly, the use of oral antimicrobials, including azithromycin and fluoroquinolones, also decreased.

Background: Over-testing for Clostridioides difficile infection outside acute diarrheal illness without a clear alternative cause can lead to inappropriate diagnosis and treatment with antibiotic therapy. Preanalytical interventions such as education, order restriction, and electronic order assistance are common but are limited in effectiveness. As an alternative approach, our antibiotic stewardship program (ASP) implemented prospective audit and feedback (PAF) on C. difficile PCR orders to reduce inappropriate testing.

Methods: The study was conducted at a 399-bed, tertiary-care, Veterans’ Affairs Medical Center and included adult inpatients and outpatients for whom C. difficile PCR testing was ordered. In the preintervention period from June through September 2019, the ASP was alerted to C. difficile PCR tests and collected data but did not intervene. From October 2019 to January 2020, the ASP performed real-time PAF at the time of ordering. Appropriateness of testing was determined based on whether there was a negative result in the prior 7 days and a 3-step review of clinical factors (Fig. 1). When possible, a direct conversation took place with the ordering provider. If not possible, a general note delineating appropriate clinical criteria for testing was generated. No PAF was done outside standard hours. The ASP recommended cancelling tests deemed inappropriate. Monthly test rates during the pre- and postintervention periods were compared using the Student t test with α = .05, and test appropriateness was compared using the χ2 test. All analyses were conducted using Microsoft Excel software. Results: During the preintervention period, a total of 418 tests were ordered (104.5 per month). This number decreased to 276 (69 per month) during the intervention period. (p Conclusions: Direct PAF at the time of C. difficile PCR ordering may increase test appropriateness and is associated with a reduction in overall testing, primarily by reducing the number of tests that are considered not appropriate on clinical grounds. PAF is effective but requires significant time investment by ASP staff and may not be a sustainable intervention over time.

Funding: None

Disclosures: None

Changes in antimicrobial use during the pandemic in relation to long-term trends in utilization among different antimicrobial stewardship program models have not been fully characterized. We analyzed data from an integrated health system using joinpoint regression and found temporal fluctuations in prescribing as well as continuation of existing trends.

Background:Clostridioides difficile infection (CDI) is a major cause of morbidity and healthcare costs in the United States. The epidemiology of CDI has recently shifted, with healthcare-associated (HCA) CDI trending downward and community-associated (CA)-CDI becoming more prominent. The cause of this shift is not well understood but may be related to changing genomic epidemiology. We assessed C. difficile strains across a CDC Emerging Infections Program (EIP) site in Western New York, including strains from both HCA-CDI and CA-CDI cases to characterize predominating strains and putative transmission across epidemiological classifications and between index and recurrent cases. Methods: In total, 535 isolates of C. difficile were collected over a 6-month period in 2018 from the Monroe Country, New York, EIP site and were analyzed using whole-genome sequencing (WGS). Standard epidemiological definitions were used to classify cases as hospital onset (HO-CDI); community associated (CA-CDI); community onset, healthcare associated (CO-HCFA-CDI); or long-term care onset (LTCO-CDI). Recurrent cases were defined as those diagnosed within 8 weeks of an initial positive test. Multilocus sequence types (MLSTs) were assigned according to PUBMLST and single-nucleotide polymorphisms (SNPs) were determined using a modified CFSAN analytical pipeline. Cases resulting from putative transmission were defined as those separated by 0–1 core SNPs. Results: Of 535 isolates, 454 were from index and 81 were from recurrent cases. The index cases were comprised of CA-CDI (47.4%), CO-HCFA-CDI (24%), LTCO-CDI (8.1%), and HO-CDI (19.3%). Cases with recurrent disease mirrored the epidemiological distribution of the larger set. Common MLSTs included ST2 (12.3%), ST8 (10.5%), ST42 (7.9%), ST58 (4.9%), ST43 (4.5%), and ST11 (4.3%). The previously widespread epidemic strain, NAP1/ST1/RT027 accounted for Conclusions: The genomic epidemiology of C. difficile across this large community cohort demonstrated a diverse group of strain types that was similarly distributed across epidemiological classifications and between index and recurrent cases. SNP analysis indicated that direct transmission between cases was uncommon.

Funding: None

Disclosures: None

Background: Prompt removal of unnecessary central venous catheters (CVC) may reduce central-line–associated bloodstream infection (CLABSI). Primary non–central-line–associated hospital-acquired bloodstream infection (BSI), including peripheral intravascular (PIV) catheter-associated bloodstream infection (PIVABSI) remains a problem. Hospitals use CLABSI surveillance data to measure patient safety, yet this measure alone fails to describe the burden of total intravascular device-related infection. We described non-CLABSI primary BSI due to PIV in our medical–surgical ICU population. Methods: Hospital-wide surveillance for primary hospital-acquired BSI, including CLABSI, was conducted in accordance with NHSN protocol. We measured PIV catheter days and central-line days using a database including nursing device documentation and patient census data to count the number of patients with 1 or more devices in place in each location, counted at the same time each day. By substituting the role of the CVC with short or midline PIV in NHSN CLABSI surveillance protocols, we performed surveillance for PIVABSI. We defined PIVABSI as a patient without CVC and either a short or midline catheter in place for >2 calendar days on the date of BSI. Patients with BSI and both CVC and PIV were counted as CLABSI. We compared CVC and PIV utilization and the incidence density of CLABSI and PIVABSI in 8 medical and surgical ICUs at our large teaching hospital. We used OpenEpi version 3.01 software to test the hypothesis that the incidence density of CLABSI would be significantly different from that of PIVABSI. Results: From January to September 2021, there were 16 CLABSIs and 12 primary non–central-line–associated hospital-acquired BSIs, all 12 were PIVABSIs. Of these 12, 8 had >1 PIV in place and none were midlines. There were 13,418 central-line days, 10,897 short and midline peripheral IV days, and 22,415 patient days, resulting in device utilization ratios of 0.60 and 0.49, respectively. The incidence density of CLABSI was 1.2 per 1,000 central-line days, although the incidence density of PIVABSI was 1.1 per 1,000 peripheral IV days (P = .84). There was no difference in pathogens between the 2 groups. Conclusions: PIVABSI represented more than one-third of the total primary hospital-acquired BSIs in our medical and surgical ICUs. Total BSI surveillance is feasible. Efforts to reduce CLABSI should be part of a broader strategy to decrease total hospital-acquired BSI from all vascular access devices.

Funding: None

Disclosures: None

Background: The IDSA has a clinical definition for catheter-related bloodstream infection (CRBSI) that requires ≥1 set of blood cultures from the catheter and ≥1 set from a peripheral vein. However, because blood cultures obtained from a central line may represent contamination rather than true infection, many institutions discourage blood cultures from central lines. We describe blood culture ordering practices in patients with a central line. Methods: The University of Iowa Hospitals & Clinics is an academic medical center with 860 hospital beds. We retrospectively collected data for blood cultures obtained from adult patients (aged ≥18 years) in the emergency department or an inpatient unit during 2020. We focused on the first blood cultures obtained during each admission because they are usually obtained before antibiotic initiation and are the most important opportunity to diagnose bacteremia. We classified blood-culture orders as follows: CRBSI workup, non-CRBSI sepsis workup, or incomplete workup. We defined CRBSI workup as ≥1 blood culture from a central line and ≥1 peripheral blood culture (IDSA guidelines). We defined non-CRBSI sepsis workup as ≥2 peripheral blood cultures without cultures from a central line because providers might have suspected secondary bacteremia rather than CRBSI. We defined incomplete workup as any order that did not meet the CRBSI or non-CRBSI sepsis workup. This occurred when only 1 peripheral culture was obtained or when ≥1 central-line culture was obtained without peripheral cultures. Results: We included 1,150 patient admissions with 4,071 blood cultures. In total, 349 patient admissions with blood culture orders (30.4%) met CRBSI workup. 62.8% were deemed non-CRBSI sepsis workup, and 6.9% were deemed an incomplete workup. Stratified by location, ICUs had the highest percentage of orders with incomplete workups (8.8%), followed by wards (7.2%) and the emergency department (5.1%). In total, 204 patient admissions had ≥1 positive blood culture (17.7%). The most frequently isolated organisms were Staphylococcus epidermidis (n = 33, 16.2%), Staphylococcus aureus (n = 16, 7.8%), and Escherichia coli (n = 15, 7.4%) Conclusions: Analysis of blood culture data allowed us to identify units at our institute that were underperforming in terms of ordering the necessary blood cultures to diagnose CRBSI. Being familiar with CRBSI guidelines as well as decreasing inappropriate ordering will help lead to early and proper diagnosis of CRBSI which can reduce its morbidity, mortality, and cost.

Funding: None

Disclosures: None

A coronavirus disease 2019 (COVID-19) outbreak in a psychiatry hospital revealed specific challenges in its response such as difficulty in isolation, transfer, and identification of close contacts, suboptimal infection control practices, and shortage of personal protective equipment, which were overcome by support from the public health center and a neighboring university hospital.

Background: In 2008, the hospital-acquired conditions (HACs) initiative labeled central-line–associated bloodstream infections (CLABSIs) as preventable “never events” that could no longer be reimbursed by Medicare. However, some patients have inherent unpreventable etiologies for bacteremia, such as obstructive biliary malignancies. We assessed the number of CLABSIs that were reasonably preventable. Methods: We examined all CLABSI cases at 2 academic medical centers over a 2-year period (2019–2021). We established 3 categories of CLABSIs: (1) preventable CLABSI (pCLABSI); (2) end-of-life CLABSI (EOL-CLABSI), which were CLABSIs that were caused by underlying disease processes in patients who were nearing the end of their lives due to a debilitating comorbidity; and (3) definition-based (dCLABSI), which met NHSN criteria for a CLABSI but, based on the pathogen and the clinical situation, likely occurred as a consequence of a patient’s comorbidities. Two experienced infectious diseases physicians (D.U. and A.S.M.) reviewed the charts of each patient with a CLABSI and, based on expert opinion, determined the category for each CLABSI. Results: Over the 2-year period, 147 CLABSIs were identified among the 2 hospitals, 66 (44.9%) of which occurred in an ICU. Most CLABSIs were pCLABSIs, making up 99 CLABSIs (67.3%). In comparison, 20 cases were categorized EOL-CLABSIs (13.6%), although 26 cases were dCLABSIs (17.7%), and 2 cases could not be classified. There was no difference in the distribution of CLABSI types in an ICU versus a non-ICU setting (χ2P = .265). However, we detected microbiologic differences between pCLABSIs, EOL-CLABSIs, and dCLABSIs (χ2P < .001), with gram-positive cocci making up the large majority of pCLABSIs (62.6%), followed by Candida spp (24.2%). Gram-negative bacilli (GNR) made up 11.1% of pCLABSIs. In comparison, GNRs were more prevalent in EOL-CLABSIs and dCLABSIs, making up 30.0% and 38.5% of each CLABSI type, respectively. Conclusions: Two-thirds of CLABSIs were deemed preventable. Central lines are important for managing critically ill patients, many of whom have inherent risk factors for bloodstream infections. EOL-CLABSIs highlight the potential for early care discussions to avoid CLABSIs at the end of a patient’s life and to avoid unnecessary blood cultures for patients on comfort care. Additionally, the pCLABSI distinction allows hospital epidemiology teams to focus on the CLABSI cases that can realistically be prevented with appropriate central-line care, techniques, and hand hygiene. Creating these categories allows hospital systems to use more targeted approaches for improving CLABSI rates.

Funding: None

Disclosures: None

Background: Catheter-related bloodstream infection (CRBSI) rates remain high despite the use of an insertion bundle. We hypothesized that line care and maintenance by a dedicated team would help decrease CRBSI rates. This study was conducted in The Medical City (TMC), is a 526-bed, private, tertiary-care center in Pasig City, Philippines. Methods: All adult hospitalized patients from October 1, 2020, to October 31, 2021, with a newly inserted temporary central venous catheter (CVC) were eligible for inclusion. CRBSI rates before the intervention (October 2019 to March 2020) and after the intervention (April to October 2021) were compared. The intervention arm consisted of a dedicated central venous access team (CVAT) who provided education and performed daily line care and dressing changes per protocol. A series of χ2 and Wilcoxon rank-sum tests were performed to compare characteristics between exposure groups. Incidence rates of CRBSI before and after the intervention were compared using an incidence rate ratio approach. Results: In total, 209 CVCs were enrolled in the study, with 103 CVCs (49.28%) in the preintervention arm and 106 CVCs (50.72%) in the postintervention arm. Baseline patient characteristics were similar. CRBSIs were more frequent in the preintervention arm than the postintervention arm (39 of 103 vs 28 of 106; P = .08). The CRBSI incidence density rate was higher in the preintervention arm than the postintervention arm, but the difference was not statistically significant (37.46 per 1,000 patient days vs 25.97 per 1,000 patient days; P = .14). Median time to CRBSI was similar in both groups (9 vs 8 days). Conclusions: Baseline CRBSI rates were high and risk of infection increased by day 8 after line insertion. We detected a decreasing trend in rates of CRBSI with a dedicated CVAT, but multiple interventions are likely needed to influence overall rates.

Funding: None

Disclosures: None

A specific, clinical-epidemiology, month-long rotation for all infectious disease fellows as well as a 1-year subspecialty track provides education in clinical epidemiology during infectious disease fellowship training. We describe the educational process created at our institution to provide this training.

Background: The NHSN parameter estimate for predicted number of central-line–associated bloodstream infection (CLABSI) is the same for gastroenterology wards as other specialty wards, such as behavioral health and gerontology. We conducted this study to contribute to the body of knowledge surrounding the risk for hospital-acquired bloodstream infection (HABSI) in patients with and without hepatic failure. The Cleveland Clinic is a 1,200-bed, multispecialty hospital with a solid-organ transplant service. Patients with hepatic failure who do not require critical care are housed on 36-bed unit A. On unit A, 43% of patients are under hepatology or gastroenterology service, although 51% of patients are under general internal medicine. Overall, unit A has a high incidence of HABSI. Methods: Surveillance for HABSI and CLABSI is performed at the Cleveland Clinic per NHSN protocol. All patients with a midnight stay on unit A from January 2019 through September 2021 were dichotomized as having hepatic failure (yes or no) if they ever received the International Classification of Diseases Tenth Revision code for “hepatic failure, not elsewhere classified.” We joined the diagnostic code to patient days and central-line-days databases and summarized the data using Microsoft Excel software. We stratified the number of patients, patient days, device days, infection classification, and hospital length of stay by whether the patient had hepatic failure, and we compared the incidence of HABSI and CLABSI between the 2 groups using OpenEpi version 3.01 software. Results: We identified 72 HABSIs among 4,285 patients who stayed on unit A for 30,910 patient days during the study period. The incidences of HABSI in patients with and without hepatic failure were 39.0 and 13.9 per 10,000 patient days, respectively (P < .001). The incidence of CLABSI was 5.4 and 1.9 per 1,000 line days, respectively (P = .01). Patients with hepatic failure stayed longer (11.5 vs 5.9 days), yet the central-line utilization ratios were not substantially different (0.25 vs 0.24). Enterococcus was the most common pathogen involved in CLABSI in both groups (Table 2). Conclusions: Patients with hepatic failure experienced CLABSI more frequently than patients without hepatic failure, stayed longer in the hospital, and were less likely have HABSI attributed to another primary focus of infection according to NHSN definitions. Although hepatic failure may be among the most severe conditions among patients in a gastroenterology ward, we have demonstrated that these units house a population uniquely susceptible to HABSI and CLABSI.

Funding: None

Disclosures: None