Background: Candida species are increasingly causing infections and are considered high priority fungal pathogens. Despite this, published data describing the clinical importance of Candida growth in the bile tract is limited to case reports and small cohorts. Our goal is to characterize treatment outcomes of patients who had Candida spp isolated from bile cultures obtained via Endoscopic Retrograde Cholangiopancreatography (ERCP) to determine the value of antifungal use in these cases. Methods: We performed a single-center retrospective cohort study of patients with bile cultures positive for Candida spp collected during ERCPs from January 2010 to December 2023. Patients were identified by cross-matching databases of patients who underwent ERCP and patients with Candida-positive bile cultures. The treatment cohort included patients who received antifungals within seven days of the first ERCP with Candida growth in bile cultures (principal ERCP) compared to a control cohort who did not. Patients with candidemia or deep-seated candida infection prior to the principal ERCP or insufficient chart data were excluded. The primary outcome was a composite of death and/or development of invasive candida infection within one year of the principal ERCP date. Kaplan Meier plots and log-rank tests were used to analyze the primary outcome. Results: A total of 266 patients were included out of 285 with 8 being excluded for insufficient records and 11 being excluded for invasive candidiasis within one year prior. The included patient population was 60.2% male, 79.7% white, 7.9% black, and 12.4% other/unknown race and had a mean age of 63.6 +/- 15.8 years. The most common species of Candida identified were C albicans (65.9%), C glabrata (17.4%), and C tropicalis (7.2%) with 27 patients (9.2%) having 2 isolates in their culture. There were 52 patients (19.5%) who received antifungals—46 fluconazole and 6 micafungin. At one year the primary endpoint occurred in 23 out of 53 patients (43.3%) in the antifungal group and 93 out of 213 patients (43.6%) in the control group. The primary outcome was plotted on a Kaplan Meier curve. The hazard ratio was 1.14 (0.71 to 1.85, 95% CI; p=0.574) which did not reach statistical significance. Additionally, antifungal initiation had no statistically significant impact on rehospitalization rates (p=0.602), relapse of bacterial cholangitis (p=0.230), or recurrent Candida-positive bile cultures (p=0.441) within one year. Conclusions: This retrospective study of the use of antifungals in patients with Candida growing from bile cultures following ERCP found no benefit in starting antifungal treatment.

Background: Following the COVID-19 pandemic, various infection prevention strategies were implemented globally. Due to the challenges of asymptomatic transmission during the presymptomatic period, healthcare facilities in Japan adopted a strict universal masking policy for hospital staff and visitors. However, with the widespread availability of vaccines and therapeutic agents, as well as the reduced pathogenicity of the Omicron variant, particularly in younger populations, the universal masking policy was lifted for patients and visitors at our institution. This study evaluates the incidence of hospital-onset COVID-19 cases in a children’s hospital before and after this policy change. Methods: A retrospective analysis was conducted to assess hospital-onset COVID-19 cases among hospitalized patients at Tokyo Metropolitan Children’s Medical Center, Japan, between March 2020 and December 2024. Hospital-onset infection was defined as symptomatic COVID-19 diagnosed on or after the fifth day of hospitalization with a positive SARS-CoV-2 PCR result. Cases with a Ct value of ≥30, indicative of past infection, were excluded. The infection rate was calculated as the number of hospital-onset COVID-19 cases per 10,000 patient-days. During the universal masking policy period, all adults and children capable of wearing masks were required to wear them at all times in public spaces. From July 2023, patients and visitors were no longer required to wear masks unless otherwise indicated, while hospital staff continued universal masking during patient care but were allowed to remove masks for non-patient-care activities. The study compared two periods: the universal masking policy period (March 2020–June 2023) and the post-universal masking policy period (July 2023–December 2024). Results: During the universal masking policy period, there were 8 hospital-onset COVID-19 cases, compared to 10 cases in the post-universal masking policy period. The median ages of the patients were 106 months (IQR: 48-132) and 94 months (IQR: 56-152), respectively. The hospital-onset COVID-19 infection rates were 0.19 and 0.52 per 10,000 patient-days, respectively (p=0.07). None of the hospital-onset COVID-19 cases progressed to severe disease. Conclusion: The universal masking policy may have contributed to reducing the incidence of hospital-onset COVID-19 among patients and visitors in our children’s hospital to some extent. Further studies are needed to evaluate the long-term impact of masking policies on infection prevention in pediatric settings.

Background: Distinguishing outbreaks from pseudo-outbreaks is essential in healthcare settings. Pseudo-outbreaks are defined by an increase in identified organisms without clinical evidence of infection. Here we report two cases involving Pseudomonas aeruginosa identified in clinical specimens, later determined to represent a pseudo-outbreak. Methods: Patient #1 had vertebral osteomyelitis and epidural abscess; intraoperative and blood cultures grew Streptococcus mitis/oralis. Four days post-surgery, one colony of P. aeruginosa grew from one of three intraoperative aerobic cultures. Patient #2 developed a fracture-related infection of the ankle and underwent debridement and hardware removal; all intraoperative cultures grew methicillin-susceptible Staphylococcus aureus. Four days later, two colonies of P. aeruginosa were detected in one of three intraoperative aerobic cultures. Both these findings were deemed unusual, leading to an outbreak investigation, including chart review and laboratory investigations, to identify a source of contamination. Results: The two cultures were received and set up one day apart by different staff. Subsequently, the WASPLab incubation system’s photographic record of the plates demonstrated no P. aeruginosa within the expected first 48 hours, suggesting contamination during culture collection or processing was unlikely. Further review revealed a heavily inoculated culture of P. aeruginosa was processed by the same laboratory technician on an open bench immediately before handling plates for patients #1 and #2. P. aeruginosa typically grows within 24 hours of incubation, and the colony morphology of the contaminated plates matched those of the heavily inoculated culture. Furthermore, both patients had monomicrobial growth of a likely pathogen causing their infection. Therefore, we concluded that this was cross-contamination, likely via aerosolization or improper plate handling. For patient #1, cefepime was discontinued on post-operative day six and switched to ceftriaxone, completed for six weeks, followed by suppressive therapy with amoxicillin, with no recurrence at three months. Patient #2 completed six weeks of cefazolin without anti-pseudomonal coverage, also without recurrence at three months. Conclusions: The pseudo-outbreak likely stemmed from cross-contamination caused by aerosolization or handling heavily inoculated P. aeruginosa cultures near the time and location of the two patients’ plates on an open bench. Awareness of such rare contamination pathways is critical for microbiology labs and clinicians, especially when handling hazardous isolates such as Brucella spp. Careful record keeping and digital storage of serial plate images can narrow the source of contamination, and active surveillance by trained epidemiology personnel is essential to detecting pseudo-outbreaks. Clinical and microbiological correlation should guide treatment to avoid unnecessary antibiotic treatment.

Background: Globally, Antimicrobial Resistance is a growing threat to global health security and economic development. Due to multidrug resistance, bloodstream infections (BSI) are a growing public health concern and a common cause of morbidity and mortality, especially among non-malarial febrile children Method: This study assessed laboratory BC process outcomes among non-malarial febrile children below five years of age at five AMR surveillance sites in Uganda between 2017 and 2018. Secondary BC testing data was reviewed against established standards. Result: Overall, 959 BC specimens were processed. Of these, 91% were from female patients, neonates, infants, and young children (1-48 months). A total of 37 AMR priority pathogens were identified; Staphylococcus aureus was predominant (54%), followed by Escherichia coli (19%). The diagnostic yield was low (4.9%). Only 6.3% of isolates were identified. AST was performed on 70% (18/26) of identified AMR priority isolates, and only 40% of these tests adhered to recommended standards. Conclusion: Interventions are needed to improve laboratory BC practices for effective patient management through targeted antimicrobial therapy and AMR surveillance in Uganda. Further research on process documentation, diagnostic yield, and a review of patient outcomes for all hospitalized febrile patients is needed.

Accurate mortality forecasting is crucial for actuarial pricing, reserving, and capital planning, yet the traditional Lee-Carter model struggles with non-linear age and cohort patterns, coherent multi-population forecasting, and quantifying prediction uncertainties. Recent advances in deep learning provide a range of tools that can address these limitations, but actuarial surveys have not kept pace. This paper provides the first concise view of deep learning in mortality forecasting. We cover six deep network architectures, namely Recurrent Neural Networks, Convolutional Neural Networks, Transformers, Autoencoders, Locally Connected Networks, and Multi-Task Feed-Forward Networks. We discuss how these architectures tackle cohort effects, population coherence, interpretability, and uncertainty in mortality forecasting. Evidence from the literature shows that carefully calibrated deep learning models can consistently outperform the Lee-Carter baselines; however, no single architecture resolves every challenge, and open issues remain with data scarcity, interpretability, uncertainty quantification, and keeping pace with the advances of deep learning. This review is also intended to provide actuaries with a practical roadmap for adopting deep learning models in mortality forecasting.

Background: The gold standard for surgical site infection (SSI) surveillance is 100% chart review, a practice neither efficient nor pragmatic for most large hospital systems.

Modern infection surveillance software uses indicators – specific data elements within the patient medical record - to report possible SSI for confirmation by infection preventionists (IPs). Using all available indicators has been shown to increase identification of SSIs and may approximate the gold standard but has been called “noisy” for including many patients with no SSI and costing precious surveillance time. Here, we describe our experience with evaluating the performance of our surveillance system. Methods: The setting for this study was the 21-hospital Cleveland Clinic health system with a uniform SSI surveillance plan and shared medical record. Our software, Bugsy (Epic Systems Corporation), employs six indicators for possible SSI: hospital readmission, return to surgery, positive microbiology tests, and chief complaint, physician diagnoses (billing codes), or administration of post-prophylaxis antibiotics suggestive of SSI. We extracted all possible SSIs, indicators, and confirmed infections for seven NHSN procedure code categories. We calculated the sensitivity and specificity of each indicator individually using OpenEpi v3.01 and estimated the cost associated with IP time spent on indicators that do not result in confirmed SSI. Results: From January to December 2023, 12,739 possible SSIs were reported with any indicator out of 26,276 inpatient procedures (48%). The frequency of procedures flagged for review ranged from 25% for CSEC to 78% for HYST. The number of procedures, possible SSI, and confirmed SSI with rates are shown in Table 1. The sensitivity and specificity of each indicator are shown in Table 2. Infection preventionists spent an average of 2 minutes reviewing each of 12,027 patient charts (401 hours) and determined there was no SSI, costing an estimated $18,602 annually. Conclusion: Nearly 50% of surgical patients were flagged for review for possible SSI with any indicator. Post-prophylaxis antibiotic was the most sensitive (97%) but least specific (65%) indicator. There was variability in indicator performance between procedure types. Readmission to the hospital was more sensitive in procedures with implants, e.g. KPRO and HPRO, than in procedures without, such as COLO and HYST. Evaluating the performance of possible SSI indicators enables IP programs to make data-driven and pragmatic decisions related SSI case finding practices. Tuning the indicator criteria within the software build may be necessary for optimization and presents an opportunity for IP time and cost savings.

Background: Blood culture volume is crucial to accurate diagnosis of a bloodstream infection. Underfilling blood culture bottles decreases test sensitivity and has been associated with contaminants. Pediatric blood culture volume recommendations are patient weight-based and difficult to audit. Our objectives were to assess healthcare worker pediatric blood culture volume knowledge and to measure culture volumes during the pre-implementation phase of a blood culture quality improvement program. Methods: Data were collected May 2024-November 2024. Surveys were administered to healthcare personnel who regularly obtained blood cultures. To estimate the collected blood volume, blood culture bottles in the laboratory were weighed, and weights were subtracted from the averaged weight pre-filled bottles. Bottles that were Results: <90% of the recommended weight-based volume were classified as underfilled. Blood culture results were compared between bottle characteristics using chi squared and Wilcoxon rank tests as appropriate. Results were presented to stakeholders to facilitate discussions on blood culture collection. 65 surveys were completed. 59 (90.8%) of respondents reported receiving blood culture training. Of those who received training, 51 (78.5%) reported that they had received weight-based blood culture training. A convenience sample of 1,076 bottles were weighed, representing 38.8% of blood cultures collected. Of those, 816 (75.8%) were underfilled (median percentage of recommended volume -57.8% (interquartile range (IQR) -80.3%, -12.3%)). Only 574 (54.3%) cultures were appropriately inoculated into a pediatric bottle based on patient weight. 83 bottles (7.7%) grew bacteria or fungi, 61 (73.5%) were non-commensals. There was no association between underfilling and positivity (p = 0.47). The median percentage of recommended volume did not differ between positive and negative cultures (-60.6% and -57.3%, respectively; p = 0.92). The median percentage of recommended volume for culture growing non-commensals and commensals was -55.5% and -69.1%, respectively (p = 0.01). Stakeholder groups reported that barriers to appropriate volume included: uncertainty regarding blood culture protocols, technical issues obtaining blood and total blood draw limits. Conclusions: In a large, regional children’s center, the majority of weighed pediatric blood culture bottles were underfilled despite the majority of respondents reporting blood culture volume training. Fill volume was not associated with positivity, which may be due to the large proportion of underfilled bottles in this sample. Non-commensal blood cultures did have a higher median percentage of recommended volume as compared to commensal blood cultures, which is consistent with prior publications. Future quality improvement programs will focus on dissemination of policy and addressing systems and technical barriers.

Background: Antimicrobial resistance (AMR) poses a critical threat to global health, with healthcare-associated infections (HAIs) such as Clostridioides difficile and multidrug-resistant organisms (MDROs) exacerbated by antibiotic misuse. Up to 50% of inpatients receive antibiotics during their hospital stay, nearly 1/3 of which is inappropriate. The Standardized Antimicrobial Administration Ratio (SAAR) at Michael E DeBakey VA Medical Center (MEDVAMC) exceeded 1 from 2023–2024, signaling higher-than-expected antibiotic use. Nurses, as pivotal frontline healthcare providers, are often underutilized in antimicrobial stewardship program (ASP) efforts due to a lack of formal ASP education. Addressing this gap aligns with The Joint Commission standards, CDC guidelines, and ANA recommendations for improving ASP engagement and reducing HAIs. Methods: This quality improvement project utilized the Plan-Do-Study-Act (PDSA) framework to develop, implement, and refine an educational intervention aimed at enhancing RN knowledge and engagement in ASP. Baseline data, including a survey assessing RN ASP knowledge, informed the creation of a tailored training program. The program emphasized the 5D approach (Diagnosis, Drug, Dose, Duration, De-escalation), the role of nurses in ASP, and interdisciplinary collaboration. The initiative was endorsed by leadership and delivered through interactive workshops and case-based learning. Post-intervention surveys and infection rate analyses were conducted to evaluate outcomes. Results: The intervention led to a 92% increase in RN knowledge, with a mean post-intervention scores of 92 out of 100 among 67 participating nurses, compared to preintervention score of 48 out of 100. Improved RN competency in ASP facilitated stronger interdisciplinary communication and adherence to stewardship protocols, such as performing antibiotic time outs. Feedback from participants highlighted increased confidence in ASP roles and improved patient safety practices. Some examples of patient safety practices that improved, included more consistent documentation of allergy checks, antibiotic indications, and treatment plans within the electronic health record. Post-intervention, nurses felt more comfortable providing patient education on the importance of completing antibiotics, recognizing side effects, and infection prevention. Conclusions: Empowering nurses through targeted ASP education not only bridges critical knowledge gaps but also fosters a culture of safety and accountability in antibiotic use. Sustaining these outcomes requires integrating ASP education into routine RN training, continuous monitoring of infection rates, and leveraging interdisciplinary collaboration to maintain compliance with evidence-based stewardship practices. These findings underscore the transformative potential of nurse-led initiatives in combating AMR and improving healthcare outcomes.

Background: The risk of bacterial transmission through gastrointestinal endoscopes remains a critical concern in healthcare-associated infections, driven by the complex design of endoscopes and potential lapses in reprocessing protocols. Contaminated endoscopes can serve as vectors for multidrug-resistant organisms, posing significant threats to patient safety. Current U.S. guidelines for endoscope reprocessing and infection control do not mandate routine surveillance sampling; however, select facilities have successfully adopted routine surveillance cultures to monitor reprocessing efficacy. The Centers for Disease Control and Prevention (CDC) has published protocols to support facilities choosing to implement such practices, emphasizing the importance of identifying persistent transmission risks. Methods: This retrospective study was conducted at the University of Kentucky Healthcare (UKHC), a 1,086-bed academic medical center, from January 1, 2019, to June 30, 2024. UKHC implemented a surveillance program in July 2016 targeting endoscopic retrograde cholangiopancreatography (ERCP), esophagogastroduodenoscopy (EGD), endoscopic ultrasound (EUS), and colonoscopy endoscopes. Weekly cultures were performed on ERCP scopes, while targeted cultures were conducted on all four scope types used for patients colonized with carbapenem-resistant organisms (CRO). Following manufacturer instructions for use (IFU), post-reprocessing cultures were performed, and pathogens were categorized as concerning or non-concerning based on CDC protocols. Manual chart reviews identified CRO-colonized patients, and match rates were calculated by comparing endoscope culture results with patient isolates. Results: A total of 163 ERCP scopes were cultured, comprising 94 from weekly surveillance and 69 from CRO-targeted surveillance (Figure 1). Weekly surveillance yielded a 9.6% (9/94) positivity rate, while CRO-targeted surveillance showed an 8.7% (6/69) positivity rate. Among six positive samples, no matching CRO was identified. Among 189 EGD scopes subjected to CRO-targeted surveillance, the positivity rate was 25.9% (49/189), with a 4.1% (2/49) match rate to patient isolates. For 27 EUS scopes, the positivity rate was 11.1% (3/27), with a 33.3% (1/3) match rate. Among 59 colonoscopy scopes, the positivity rate was 5.1% (3/59), with no matches to patient isolates. Conclusions: The UKHC surveillance program highlights ongoing risks of bacterial transmission despite adherence to manufacturer-recommended reprocessing protocols. Scopes yielding concerning organisms underwent additional reprocessing to mitigate patient risk. All scopes with positive organisms after the second reprocessing were sent back to the manufacturer. Surveillance programs provide valuable insights into disinfection efficacy, helping to identify gaps and guide infection prevention strategies. Further refinement and standardization of surveillance protocols are essential to mitigate transmission risks associated with gastrointestinal endoscopy and improve patient safety on a broader scale.

Background: Variability in outpatient parenteral antimicrobial therapy (OPAT) management and challenges to providing recommended OPAT care can compromise patient safety and care quality. Little is known about how OPAT is currently delivered by healthcare systems across the United States (US), including within the Veteran’s Health Administration (VHA). We sought to understand and compare OPAT delivery at selected Veterans Affairs medical centers. Method: Using a qualitative methodology, we conducted semi-structured interviews with key informants involved in OPAT delivery at 6 VHA medical centers with different complexity levels in the Midwestern US. Facility complexity is determined by patient volume and complexity level along with the amount of teaching and research conducted at the facility. Interviews occurred between February and December 2024 with healthcare personnel (n=30), including primary care and infectious diseases physicians, pharmacists, nursing staff, care coordinators, and vascular access providers. Data collection focused on better understanding OPAT processes within key domains of decision-making, patient education, care coordination, and post-discharge management. We used rapid analysis and a summary matrix to compare practices across sites within each domain. Result: Our findings highlight significant variability among VHA medical centers that provide OPAT to Veteran patients. Three of the 6 medical centers had dedicated OPAT programs as evidenced by a multidisciplinary team with clearly delineated roles and responsibilities, and processes that may help mitigate adverse outcomes and improve communication between providers at all OPAT care points. These processes map to the key elements outlined in the Infectious Diseases Society of America (IDSA) practice guidelines for OPAT programs, and include determination of appropriate therapy, patient education, lab monitoring, and discontinuation of treatment. (Figure 1) Conversely, at the three VHA sites without evidence of a multidisciplinary OPAT team or program, most participants described poor communication and coordination, lack of support, and uncertainty among providers about who is responsible for OPAT care. This confusion extends to follow-up and discontinuation of treatment. OPAT key elements were lacking or poorly defined. A process map helps visualize the contrasts in care between sites with and without defined OPAT programs. (Figure 1) Conclusion: Despite its centralized healthcare system, VHA medical centers demonstrate highly variable processes with respect to OPAT care. In the absence of a clear OPAT policy or program, uncertainty among providers about roles and responsibilities may be greater. The presence of a dedicated multidisciplinary OPAT team may help improve communication and care coordination, thereby minimizing quality and safety concerns.

Background: Healthcare associated infections (HAIs) are important areas of concern as they increase length of hospital stay, increase hospital costs, and have high morbidity and mortality. For instance, central line associated bloodstream infections (CLABSIs) approximately increase length of stay by 13.4 days and increase hospital costs by $43,975. Studies also suggest a 1.5-2.5x increase in mortality in patients who develop CLABSIs. In 2013, the REDUCE MRSA trial compared universal MRSA decolonization to targeted MRSA decolonization in the ICU and found superiority ini reducing positive MRSA cultures and all cause bloodstream infections. We aim to decrease central line associated bloodstream infections at our institution by adopting the REDUCE MRSA trial protocol. Methods and Outcomes: All patients admitted to the medical/surgical ICU and cardiac ICU at St Francis Hospital starting in December of 2023 received daily intranasal mupirocin and chlorhexidine bathing regardless of their MRSA status. The primary outcome assessed was the CLABSI rate per month. The secondary outcomes were the standard infection ratio and CLABSI per central line day. We compared data from 2020-2023 to data after initiation of the protocol in 2024. We used unpaired t testing to assess the CLABSI rate per month and used a negative binomial regression model to calculate the standard infection ratio according to the NHSN 2015 national baseline. Results and Discussion: We had a total of 6 CLABSIs in the ICU this year after initiating universal MRSA decolonization. The number of CLABSIs per month decreased from 0.65 per month from 2020-2023 down to 0.50 per month in 2024. These results, while not statistically significant, are limited by the small sample size since the protocol was just initiated this year. One interesting finding was 5 of the 6 CLABSIs occurred during January through March, which brings up the question if introducing these new changes required time for nursing education and compliance to improve. Conclusions: Our results suggest that universal MRSA decolonization in the ICU may decrease the number of CLABSIs. We will continue to collect more data in the coming years to assess for statistical significance. We recommend further research to assess for potential benefits of universal MRSA decolonization in other areas of the hospital where MRSA infection rates are high like step down units.

Introduction: Although most rhinosinusitis cases are viral, misdiagnosis of an underlying bacterial cause is common, leading to excessive antibiotic utilization. Interventions to improve diagnosis and prescribing for sinusitis may reduce antimicrobial resistance and improve patient outcomes. Methods: Antibiotic prescriptions by 237 URMC Primary Care Network(PCN) clinicians between 9/1–11/31/2022 (baseline n=23,048) and 12/1/2023–2/29/2024 (post-intervention n=18,885) were extracted as part of a network-wide education and prescription feedback intervention focusing on antibiotic utilization rates, and guideline-concordant prescribing for sinusitis defined from local antibiograms and national guidelines. Random subsets of pre- and post-intervention prescriptions (250 each period for 90% power, Type I alpha=0.05) were systematically reviewed by two medical student or resident reviewers, with adjudication of discrepancies by infectious diseases-trained clinicians for 1) appropriate diagnosis of bacterial infection, 2) antibiotic selection, and 3) treatment duration. Charts were also reviewed for treatment failure requiring course extension, urgent care or emergency department utilization, hospital admission, and antibiotic-related toxicity during therapy or within 30 days of conclusion. Statistical analysis was performed in GraphPad and Excel ver. 2408. Results: Correct use of diagnostic criteria increased from 52.8% to 63.6% (p<0.01) post-intervention, a 10.8% absolute and 20.4% relative increase in appropriate diagnosis of bacterial sinusitis. Although rates of guideline-concordant treatment duration remained similar (74% vs. 74.4%, p=0.92), appropriate antibiotic selection increased from 71.6% to 85.6% (p< 0.01) (Fig. 1), and complete concordance of both spectrum and duration increased from 49.6% to 60.8% (p=0.012). Rates of antibiotic-associated adverse events requiring treatment discontinuation were similar pre- and post-intervention (Table 1). There were no Clostridium difficile infections. Health care contact beyond the prescribing office increased post-intervention (p < 0.01) (Table 1) but this was unrelated to sinusitis or antibiotic treatment complications (Fig. 2). Conclusions: A large-scale network intervention significantly improved use of diagnostic criteria and appropriate treatment of acute bacterial sinusitis without negatively impacting the incidence of treatment- or infection-related complications.

Background: Healthcare produces 10% of US carbon emissions. Diagnostic stewardship can improve patient care while reducing emissions from unnecessary testing. In the Bronx, community outbreaks of Legionella pneumophila result in frequent testing, yet, most testing for Legionella and Pneumococcus are negative. The institution is transitioning from separate tests for Streptococcus pneumoniae and L. pneumophila to a single combined test. We evaluated the environmental impact of this change. Methods: To approximate emissions from separate versus combined testing kits, each kit component was weighed and converted to CO2 emission equivalents using the US Environmental Protection Agency (EPA) greenhouse gas equivalencies calculators. Using data from the previous 12 months and projecting similar future testing patterns, we estimated an emissions reduction based on combined testing. By multiplying waste emissions before and after combined testing, we were able to estimate the emissions diverted by this institutional change. Results: From January 2023 to November 2024, 13906 urine Legionella tests were performed and 62 (0.45%) were positive, while 12796 urine Pneumococcal tests were performed and 479 (3.74%) were positive. Separate testing kits generated 218.2 kg of plastic waste and 749.1 kg of paper waste. Projected plastic waste from January 2025 to November of 2026 using the combined test is 48.5 kg of paper waste and 64.2 kg of plastic waste. The total estimated emissions reduction is 4.34 Kg of CO2. Conclusions: Diagnostic stewardship via a combined urinary antigen test can reduce waste from unnecessary testing without altering workflow. This institutional change is projected to reduce the plastic waste equivalent to 8480 water bottles and reduce emissions equivalent to using 11 gallons of gasoline or charging 340 smartphones. Our calculations are likely an underestimate of total emissions diverted since we only estimated emissions produced by waste in the landfill and did not account for other associated emissions such as those produced by transporting the waste. While the impact is modest, diagnostic stewardship applied broadly is a step towards a goal of net zero.

Background: Follow-up blood cultures (BCx) are ordered after an initial positive culture in many instances. The number of follow-up cultures needed is not clear. Obtaining unnecessary BCx may cause unintended consequences. The optimal balance between stewardship and patient safety warrants investigation. We sought to assess the frequency with which a third set is positive after a negative second BCx. Methods: We conducted a retrospective study of BCx submitted to the microbiology laboratory from 1/1/18-11/1/23. We included all patients ≥18 years who had at least two follow-up BCx drawn 24-72 hours after an initial positive culture. Data were collected from electronic medical records. Cultures obtained within two hours of each other were counted as one set. Different strains of an organism were considered to be different organisms. Patients were divided into four groups based on BCx positivity, with a focus on the cohort with a positive culture after a negative follow-up set. Results: 28,875 patients had an initial positive BCx, of which 2,636 had at least two follow-up cultures drawn in the selected timeframe. Within this group, 585 (22.2%) had two positive follow-up sets, 1500 (56.9%) had two negative, 431 (16.4%) had a positive followed by a negative, and 120 (4.6%) had a negative followed by a positive. Of this cohort, 71 (2.7%) grew the same organism in the initial and second follow-up cultures, while 49 (1.9%) did not. In the same-organism subset, the most commonly identified bacteria were coagulase-negative staphylococci (n=21; 0.8%), gram-negative bacteria (n=17; 0.6%), methicillin-sensitive Staphylococcus aureus (n=13; 0.5%), and methicillin-resistant S. aureus (n=7; 0.3%). The most frequently isolated organisms in this subset were S. aureus (n=20; 0.8%), Staphylococcus epidermidis (n=16; 0.6%), and Escherichia coli (n=11; 0.4%). In the different-organism subgroup, 35 (1.3%) of the second follow-up sets had suspected contamination, though true bacteremia from skin/soft tissue (n=4; 0.2%), central line (n=4; 0.2%), unknown (n=3; 0.1%), and other sources was observed, often due to S. aureus (n=4; 0.2%), E. coli (n=2; 0.1%), and Candida (n=2; 0.1%). Conclusion: The number of patients with ongoing bacteremia that would have been missed with one follow-up BCx was small. The skip phenomenon has been described with S. aureus but was seen with gram-negatives as well. The second follow-up cultures were sometimes positive for contaminants. Further data are needed to determine when two follow-up sets should be obtained rather than one.

Background: Following the identification of Legionella pneumophilia by next generation sequencing (NGS) testing in an immunocompromised patient in June 2024, an extensive environmental sampling initiative was launched to determine possible contamination sources. Method: The water management team was immediately notified and performed initial testing on the unit where the positive case was identified, in which six ice and water machines, two showers, and four sinks were all sampled from two inpatient oncology units. Result: Only 4 of the 270 identified ice machines were tested every quarter. Since there was no randomization in the testing, this did not provide an accurate representation of the effectiveness of the hospital’s water management system. The results identified an ice machine located in a low traffic area of the unit which resulted positive for Legionella pneumophilia. Conclusion: Our findings led to the identification of the implicated ice machine in a low traffic area of the unit as our primary source, highlighting the need for more robust water monitoring practices across high-risk areas. When interviewing the nursing team, it was mentioned that the patient frequently requested ice chips to consume due to oral lesions, and nursing flow confirmed the impacted ice machine was the source of the patient’s ice chips. In response, we expanded water testing protocols from 4 (1%) to 22 (8%) randomized ice machines in quarter three testing, resulting in the identification of 10 of the 22 machines testing positive for Legionella, a 45.5% positivity rate. This allowed us to change our water testing program moving forward to include 120% more ice machines in the quarterly testing program. We also implemented a phased approach to address the low flow machines throughout the hospital that are at risk of bacterial production in which we are gradually installing flow meters and monitoring usage for approximately one to two months per machine. Machines that are deemed low flow will be removed from the unit. We aim to reduce low flow ice machines on the inpatient units by 50% by the end of quarter 4 in 2025. These changes have reinforced our commitment to patient safety by limiting pathogen exposure and operating from an infection prevention standpoint. This investigation underscores the critical role of environmental monitoring and ongoing assessment of water-related equipment in healthcare settings.

Background: The diagnostic approach for Clostridioides difficile infection (CDI) significantly influences treatment and resource utilization. This study compares clinical characteristics and treatment choices based on three testing algorithms combining antigen and toxin enzyme immunoassay (EIA) tests and nucleic acid amplification tests (NAAT). Methods: We performed a retrospective study of patients tested for CDI between August 2022 and November 2024 in a large health system where multiple CDI testing algorithms are utilized: (arm 1) antigen and toxin EIA with automatic reflex to NAAT if discrepant results; (arm 2) NAAT with automatic reflex to toxin EIA if NAAT positive; and (arm 3) antigen and toxin EIA with NAAT at provider request with approval by Antimicrobial Stewardship. The last step in the testing algorithms determined whether the result was considered positive. We determined positivity rate by algorithm results and compared clinical variables including fever (temperature > 38.0° C) or abnormal white blood cell (WBC) count (12,000) within 48 hours prior to test order, laxative use within 24 hours prior to test order and treatment rates between those who tested toxin positive by EIA and those who tested toxin positive by NAAT only. Treatment was defined as receiving oral vancomycin or fidaxomicin for more than 5 days OR receiving those medications on the day prior to or day of discharge. Results: A total of 16,555 patients were tested. Overall algorithm positivity rate was highest in the EIA with reflex to NAAT (arm 1) at 13.7% compared to 5.7% for arm 2 (NAAT with reflex to toxin EIA) and 5.1% for arm 3 (EIA with NAAT at Provider Request). Toxin EIA positive patients were 1.2 times more likely than NAAT positive patients to display fever or abnormal WBC in the 48 hours prior to test order (p < 0 .001). Toxin EIA positive patients were less likely to receive laxatives compared to NAAT only positive patients. (p=0.11). Among toxin EIA positive cases, 89% received treatment compared to 57% in toxin NAAT only positive cases (p < 0 .001). 46% of patients who tested NAAT positive with a subsequently negative toxin EIA were treated. Conclusion: Patients with toxin EIA positive tests were more likely to exhibit systemic signs of infection and were treated at higher rates compared to NAAT-positive cases. While NAAT-based testing identified additional cases, many may reflect colonization. Treatment of toxin NAAT positive/toxin EIA negative patients was common highlighting opportunities for diagnostic stewardship.

Background: Central line-associated bloodstream infections (CLABSIs) are preventable infections associated with poor outcomes. Nationally, the CLABSI standardized infection ratio (SIR) decreased from 2018 to 2019, but those positive results were derailed by the COVID-19 pandemic. From FY21 to FY22, the CLABSI SIR in our facility’s medical intensive care unit (MICU) more than doubled. In March 2021, we created a multidisciplinary central venous access device (CVAD) rounding team to decrease CLABSI in the MICU. Method: We conducted a prospective pre-post quality improvement study in an academic, quaternary care hospital with a 30-bed MICU. The decision to implement a multidisciplinary CVAD rounding process was based on a review of published best practices. The study was approved by the UNC Institutional Review Board.

Our team included MICU clinicians, registered nurses, an infection preventionist, a vascular access registered nurse and a “CVAD Liaison”. The CVAD liaison role is a registered nurse trained maintaining aseptic technique during CVAD insertions and educating staff on CVAD maintenance. Each teammate had assigned responsibilities (Fig. 1).

The team rounded weekly on every MICU patient with a CVAD. Components of the rounding process were compliance audits of all CVADs; evaluation of line necessity; targeted education and process improvement. Result: This study evaluated the intervention’s impact on CLABSI, CVAD utilization, and maintenance bundle compliance rates. Data were collected for five fiscal years (FY19 to FY23). Following the intervention, the MICU experienced a 57% decrease in CLABSI rates between FY21 and FY22. This reduction was meaningful for patient care, although not statistically significant. Infection rates rose slightly in FY23 but remained lower than from FY19-21 (Table 1). There was no statistically significant difference in the CVAD utilization rate between FY21 and FY22. There was an improvement in the percentage of intact dressings (Fig.2). Conclusion: Following the implementation of a multidisciplinary CVAD rounding team, there was a 57% decrease in the MICU’s FY22 CLABSI infection rate from the prior fiscal year. This decrease in CLABSI was sustained with a similar CLABSI rate in FY23. One potential explanation for the CLABSI reduction was increased awareness of the importance of maintaining CVAD dressings (Fig. 2). The primary challenge for this team has been sustaining staff availability. Staff shortages and burnout have made it challenging to find coverage for rounds. Our team began addressing staffing challenges by expanding the pool of interested staff. Sustainment of rounding and CLABSI reduction will require continued monitoring and recruitment.

Introduction: Blood culture (BCx) diagnostic stewardship is crucial for optimizing health resources and ensuring appropriate clinical testing while minimizing unnecessary cultures that could lead to increased false positives and subsequent antibiotic overuse. BCx algorithms have effectively lowered BCx rates across various patient populations without compromising patient safety. However, patients with a durable left ventricular assist device (LVAD) represent a unique group where the safety and applicability of these algorithms remain underexplored. Methods: We adapted the BCx algorithm from the DISTRIBUTE study by Fabre et al (Figure 1) and retrospectively applied it to HeartMate 3 LVAD recipients with BCx testing performed between July 1, 2019, and April 30, 2024. Each BCx was reviewed and adjudicated according to the algorithm to determine the appropriateness of BCx indication. We also assessed the incidence of true positives, contaminants, and negative cultures among BCx testing deemed as inappropriate to evaluate the algorithm’s potential impact on clinical decision-making in this specialized patient population. We used the Centers for Disease Control and Prevention’s standard definition of a contaminated BCx. Results: We reviewed 1531 blood cultures in 121 unique LVAD recipients. The most common clinical indications for BCx collection were for documenting bloodstream clearance (363, 23.7%), suspected infective endocarditis or endovascular infection (260, 17.0%), and isolated fever and/or leukocytosis (217, 14.2%). We adjudicated 945 (61.7%) BCx collections as appropriate and 586 (38.3%) as inappropriate. Out of the 586 inappropriate BCx collections, 577 (98.5%) were negative and 8 (1.4%) resulted in a contaminant (Figure 2). Only 1 (0.2%) BCx adjudicated as inappropriate resulted in a true positive, which isolated Streptococcus infantarius in an LVAD patient receiving active chemotherapy for colorectal cancer and was felt to represent gastrointestinal translocation. Discussion: We retrospectively applied a BCx algorithm to LVAD recipients to determine the clinical impact of applying such an algorithm to a high-risk patient population. We found that the BCx algorithm missed only 1 true positive bloodstream infection in a patient with additional risk factors. This study provides preliminary support that a BCx algorithm could reduce BCx testing in LVAD recipients without compromising clinical safety. Future studies on BCx diagnostic stewardship in this population should prospectively collect data and monitor for additional adverse events, such as readmission, mortality, length of stay, and antibiotic days of therapy.

Introduction: Hand hygiene (HH) performance in our facility declined during the COVID-19 pandemic and failed to return to baseline despite a widespread education campaign and increased HH rounding. To better understand provider perceptions and inform future interventions, we conducted a survey examining self-perception of HH performance, factors determining HH practices, barriers to adherence, and burnout. Methods: The survey assessed self-perceived HH performance relative to peers, perceived opportunities for improvement, barriers to HH, factors causing variation in personal HH practice (workload/acuity, peers, time of day, patient characteristics), and self-reported burnout using a validated, single-item burnout scale. Surveys were conducted in-person with clinical providers on one high-performing and three lower-performing intensive care units. All available clinical team members were included; non-clinical staff were excluded. Self-perception of performance was compared by unit, role, years of experience, and burnout rating using the Kruskal-Wallis test. Analyses were completed in SAS 9.4, Cary, NC. Results: One hundred surveys were completed. One person declined. The majority of those surveyed believed themselves to be in the top quartile of HH performers (87%). The actual HH compliance measured on these 4 units for 1/2024-11/2024 was 65% based on 7726 total directly observed opportunities. No one selected bottom or bottom 3rd quartiles. Figure 1 shows responses by percentile self-ranking. There was no difference in perceived performance by unit (p=0.4006), years of experience (p=0.9679), or burnout (p=0.2621). Non-clinical “other” type providers perceived performance to be slightly higher than clinical provider types: mean 91st percentile versus 82nd for prescribing providers, 84th for students, and 81st for nurses/nurse assistants, p=0.0353. Empty HH dispensers was the most frequent barrier cited, by 77%. A point prevalence survey on these 4 units completed the week after the survey ended verified that 22% (25/113) of dispensers were empty, however, in all of the 25 except for 1, there was a filled dispenser within 8 feet of the empty one. Discussion: HH performance was perceived to be better than average by the majority of inpatient healthcare providers (HCP) surveyed, despite data from these units indicating opportunities. Empty dispensers were consistently cited as a barrier, but likely could have been surmounted by a few steps based on the locations of the next available filled dispensers. Further improvement in HH will be difficult without efforts to move perception closer to reality for individual HCPs. Video footage or re-enactments of an observed care episode may help identify opportunities for improvement.

Background: Asymptomatic bacteriuria (ASB) is common in older adults; however, its treatment in this population is not beneficial and can be harmful. We developed an intervention to update the indications on a urinalysis and urine culture order set according to recent Infectious Diseases Society of America (IDSA) guidelines recommending against testing and treatment for bacteriuria in older adults with altered mental status (AMS) or falls. Methods: This retrospective quasi-experimental study included adult patients > 65 years who presented to an Emergency Department (ED) in Southeast Texas from July 2018 to February 2024. The intervention, implemented in June 2020, included a change in the urinalysis (UA) with reflex to urine culture order set that replaced “AMS” as an indication for UA with “AMS and fever or leukocytosis” and provider education. The study’s primary outcome was whether urine studies were obtained within 24 hours of presentation to the ED. Secondary outcomes were antibiotic days of therapy (DOT) for suspected UTI, length of stay (LOS), and 30-day hospital readmission. The difference in differences (DID) technique was used to assess the impact of the intervention on older adults who presented with AMS or a fall (cases) compared to those with other chief complaints (controls). Outcomes were evaluated using multiple regression models (logistic for urinary studies ordered and readmission, linear for log LOS, and zero-inflated negative binomials for DOT) adjusted for patient demographics and comorbidities. Results: The study included 31,626 patients. Almost a third of older adult patients who presented to the ED with a fall or AMS (31.5%) received urinary testing, and 5.9% received antibiotics. 20.2% of control patients received urinary testing in the ED, and 6.3% received antibiotics. After adjusting for confounders, the intervention did not impact the percentage of tests ordered [(DID OR 1.14, 95% CI 0.93 – 1.40), (Figure 1)]. The incident rate ratio (IRR) of not receiving antibiotics was similar after the intervention (IRR 0.78, 95% CI 0.54 – 1.13), while the baseline number of antibiotics a patient received increased (IRR 1.54, 95% CI 1.05 – 2.27). Conclusions: Modifying UA indications on the urinalysis order set did not reduce testing or treatment for bacteriuria in older adults presenting with AMS or a fall. Longer-term or repetitive educational interventions could be an effective way to improve long-term stewardship outcomes in this population, especially for providers with less exposure to guidelines. Further discussion and investigation are needed.