Background: Empiric extended-spectrum antibiotics are routinely given to patients with cancer despite low risk of infection with multidrug-resistant organisms (MDROs). This secondary analysis of the four INSPIRE (INtelligent Stewardship Prompts to Improve Real-time Empiric Antibiotic Selection) trials evaluated how computerized physician order entry (CPOE) prompts providing patient- and pathogen-specific MDRO risk estimates affected empiric extended-spectrum antibiotic use in patients with cancer. Methods: We identified non-critically ill hospitalized adults (> Results: Including all trials, 36,861 (18,272 baseline; 18,589 intervention) patients had cancer. Mean age was 69.0 (13.6); 48.0% (17,675) were male. Extended-spectrum antibiotic days-of-therapy decreased by 27% (95% CI:20-34%, PPPP Conclusions and Relevance: An antibiotic stewardship bundle that included CPOE prompts recommending standard-spectrum antibiotics for patients at low risk for antimicrobial-resistant infections reduced extended-spectrum antibiotic use in non-critically ill patients with cancer who were hospitalized with community-acquired pneumonia, UTI, abdominal infection, or SST, without observed differences in safety outcomes.

Background: Skin changes can precede central line infections by days to weeks, making early identification essential for prevention. Routine nursing assessments help prevent CLABSIs in hospitals, but comparable surveillance is lacking for patients managing central lines at home, leading to preventable infections and readmissions. Hypothesizing that artificial intelligence (AI) can be trained to reliably detect visible inflammation/infection from a photo, we developed algorithms for auto-detection of early signs of localized skin inflammation or infection. Methods: We conducted a secondary analysis of photos collected from 3 cohorts of adult patients with central venous catheters (CVCs) at a large academic hospital, an outpatient oncology clinic, and 6 nursing homes between January 2014-June 2017. Skin tone was recorded using the Fitzpatrick Scale, as erythema may be harder to detect in darker skin tones (types IV–VI). Insertion-site inflammation or infection was assessed in each photo using the Central Line Insertion Site Assessment (CLISA) score (Table 1) by trained research staff, with two-physician review. A convolutional neural network model was trained de novo from random weights (binary cross-entropy loss) to predict binarized CLISA score, defined as normal/low risk (non-actionable) scores of 0 or 1 vs high risk (actionable) score of 2 or 3. Five-fold cross-validation estimated model performance for accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the curve receiver operating characteristic (AUC/ROC) with 95% confidence intervals. Results: Among 5,551 photographs of 964 CVCs in 874 adults, there were 1,191 (21%) photos from hospitalized patients, 2,539 (46%) from outpatient oncology patients, and 1,821 (33%) from nursing home residents. Mean (SD) age was 60 (17), 54% were male, and 183 (21%) patients had darker skin tones (Fitzpatrick skin type IV-VI), Table 2. Among all photos, 1,585 (29%) had localized inflammation/infection (CLISA 2 or 3) and 298 (0.1%) had purulence. Compared to gold standard clinician assessment, AI model performance optimized for accuracy was able to distinguish between normal/low risk (CLISA 0-1) and inflamed/infected catheters (CLISA 2 or 3): accuracy 0.82, sensitivity 0.52, specificity 0.94, PPV 0.83, NPV 0.78, AUC 0.87 (95% CI 0.86-0.88); results were similar for models detecting purulence, Figure 1A-B. Conclusion: AI algorithms can identify CVCs with photographic signs of inflammation or infection and signal the need for intervention to prevent CLABSI. AI-autodetection of high-risk catheters can extend hospital CLABSI prevention to settings with limited monitoring, using photographs alone and minimizing reliance on patient self-assessment or care escalation.

This commentary informs on key pragmatic contributors to strategic and sustainable surgical IPC and AMS initiatives. Three key recommendations to strengthen and sustain surgical IPC and initiatives are (1) institutional leadership support, (2) a programmatic multidisciplinary implementation plan, and (3) effective communication strategies using motivational interview.

Background: Microbiome restoration through fecal microbiota transplantation (FMT) is a promising decolonization strategy that reduces multidrug-resistant organism (MDRO) colonization and infection. However, the strains mediating MDRO decolonization remain poorly defined, and long-term follow-up (LTF) data after FMT for MDRO decolonization are limited. We enrolled participants from a completed randomized trial of FMT for MDRO decolonization (PREMIX, NCT02922816) into an LTF study to characterize long-term engraftment dynamics of donor-derived anaerobic taxa and their association with sustained MDRO decolonization. Methods: Participants provided stool samples every six months for up to five years following completion of the trial. Samples underwent selective culturing for MDROs using chromogenic agars, with organism identification and susceptibility confirmation by MALDI-TOF and Vitek2. MDRO isolates underwent whole-genome sequencing (WGS), and stool samples underwent metagenomic sequencing. To determine similarity between baseline and LTF MDROs, pairwise comparison of baseline and LTF MDRO isolates was performed with a unique strain defined as < 99.5% average nucleotide identify difference (ANI) to other strains. To determine engraftment of donor derived anaerobic taxa construction of metagenome-assembled genomes (MAG) from donor and participants samples was performed. Donor MAGs detected at <50% breadth in LTF metagenome samples were compared to corresponding species-level MAGs from participant LTF samples, with unique MAGs defined as <99.5% ANI like strains above Results Six PREMIX participants were enrolled to the LTF study, five of whom received FMT and one control. 80% (4/5) of FMT treated patients have remained decolonized (culture negative) while 0% (0/1) of control patients have remained decolonized at most recent follow up (median[IQR] 3.6 [3.4,4.0] years). The single FMT-treated participant with MDRO detection in LTF samples had evidence of a new acquisition compared to the baseline MDRO, whereas the MDRO detected in the control participant was identical to the baseline strain based on ANI cut offs. Among FMT-treated participants, metagenomic analysis demonstrated persistent detection of donor-derived anaerobic strains at <50% breadth, with high ANI between donor MAGs and participant LTF MAGs for multiple key taxa, including Acetatifactor sp., Akkermansia muciniphila, Alistipes putredinis, Barnesiella intestinihominis, Eubacterium siraeum, Mesosutterella massiliensis, Odoribacter splanchnicus, and Phascolarctobacterium faecium, consistent with long-term engraftment (Figure 1). No donor-derived MAGs were detected in control participants at <99.5% ANI. Conclusions: FMT was associated with long-term durable decolonization potentially mediated by the engraftment and persistence of key donor derived anaerobic strains. Our findings of persistence of discrete taxa highlights the potential for developing rationally constructed strain-based microbiome therapeutics for decolonization and infection prevention.

In this cross-sectional study of 136 antimicrobial prescriptions at a rehabilitation hospital with recent implementation of electronic medication order entry and no mandatory requirements, we found incomplete documentation of both indication (37%) and duration (75%). An accompanying survey identified potential solutions including reminders, improved choice architecture, and forced functions.

Background: Social deprivation may affect both infection severity and the quality of antibiotic prescribing, but its influence on downstream outcomes and the role of stewardship processes in this pathway remain uncertain. Methods: We conducted a retrospective cohort study of 1,365,957 emergency department (ED) encounters among older adults across multiple U.S. hospitals, 2013–2022. We assessed the impact of community social deprivation on clinical outcomes and identified mediation through antibiotic prescribing (guideline-concordant empiric use). Social deprivation was quantified using a z-scored Social Deprivation Index (SDI). Outcomes were log-transformed length of stay (LOS), 30-day ED revisit, 30-day mortality, 30-day C. difficile infection, and DOOR (Desirability of Outcome Ranking; 1=alive with LOS at or below the median and no 30-day events, 2=alive with LOS above the median and no events, 3=30-day ED revisit, 4<=i>C. difficile infection, 5=death). We fitted multivariable ordinary least squares models with state- and month-fixed effects and clustered standard errors, controlling for demographics, comorbidities, clinical severity, travel time, hospital characteristics, and rurality. We applied regression-based mediation analysis to partition the SDI–outcome relationship into direct effects and indirect effects mediated by prescribing concordance. Secondary models substituted antibiotic overuse or underuse for concordance. Results: Concordant prescribing occurred in 83% of encounters; overuse and underuse occurred in 9.5% and 7.4%, respectively. Higher SDI was associated with slightly worse outcomes: each 1-SD increase corresponded to modestly longer LOS, higher risk of 30-day ED revisit and mortality, and worse DOOR scores, with minimal association with C. difficile infection (Figure 1). Concordance was strongly associated with better outcomes, including shorter LOS, fewer ED revisits, lower mortality, and lower DOOR scores. Mediation analyses (Figure 2) indicated that concordance explained about 11% of the SDI–LOS association and <5% of the associations with ED revisit, mortality, and DOOR. In secondary models (Figure 3), overuse was consistently associated with worse outcomes—longer LOS, higher mortality, and higher DOOR scores—independent of SDI. Underuse showed inverse associations with LOS and DOOR but only tiny, clinically negligible increases in ED revisit and mortality, patterns likely reflecting residual confounding and selection of lower-acuity patients. Conclusion: We found that social deprivation was associated with worse health outcomes, with antibiotic concordance serving as a partial pathway linking deprivation to harm. Stewardship efforts that improve concordance and curb overuse—particularly in socially deprived communities—may help reduce outcome disparities without encouraging under-treatment.

Background: The National Healthcare Safety Network Antimicrobial Resistance Option (NHSN AR) is the nation’s primary tracking system for antimicrobial resistance. As of December 2025, 89% of Washington (WA) acute care hospitals are reporting antimicrobial susceptibility data to the AR option. The AR option provides a feature to generate antibiograms, a clinical tool to guide empiric antibiotic selection. However, the NHSN generated antibiograms do not adhere to the Clinical and Laboratory Standards Institute’s widely used M39 (CLSI M39) guidelines, which are best practices for antibiogram creation. Due to concerns of limited clinical applicability, we sought to create a CLSIM39 compliant antibiogram dashboard using NHSN AR data to support hospitals to make informed clinical decisions and track local resistance data. Methods: We downloaded the line list for All Antimicrobial Resistance Events from NHSN, transformed the data in R, and imported it into PowerBI to generate interactive antibiogram reports. The initial dashboard focused on the most commonly reported Gram-negative pathogens in WA: Enterococcus faecalis, Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis, and Pseudomonas aeruginosa. We included isolates from blood, cerebrospinal fluid, lower respiratory tract, skin and soft tissue, wound, musculoskeletal, and urine specimens. Following CLSI M39 guidelines, we retained only the first isolate per organism per patient per calendar year. For specimen-specific reporting, we applied specimen filters by source prior to isolate selection. We selected antimicrobials and final susceptibility interpretations to calculate percent susceptible based on CLSIM39 and M100 guidelines. We suppressed organism-agent combinations with fewer than 30 isolates tested or lack of clinical relevance. To improve interpretability, we included the median and range of isolates tested for each organism. Three antibiograms were available: all-specimen sources, blood only, and urine only. To ensure data security, we assigned hospital alias codes and hosted the dashboard on a secure PowerBI platform. Results: WA DOH launched the NHSN AR Antibiogram Dashboard in December 2025. The dashboard displays susceptibility data based on analysis of 209,836 isolates collected between 1/1/2020 and 11/30/2025. We generated 74 all-specimen, 43 blood, and 74 urine-specific facility antibiograms. Currently, 93 antimicrobial stewards representing 80 hospitals have dashboard access, enabling them to track antimicrobial resistance trends year-to-year and compare resistance patterns by specimen source. Conclusion: The dashboard equips antimicrobial stewards with actionable data to monitor local resistance trends and inform facility-specific antimicrobial use guidelines. We plan to expand the dashboard to include additional organisms and specimen-specific reports.

While Centers for Medicare and Medicaid Services-mandated Antimicrobial Stewardship Programs are now in place across U.S. skilled nursing facilities, reported high rates of compliance may mask persistent gaps in clinical effectiveness. This review summarizes evidence-based antimicrobial stewardship interventions, their impact, and practical tools to support implementation. Resources from Centers for Disease Control and Prevention, Agency for Healthcare Research and Quality, other frameworks, and proposed metrics are highlighted to guide effective stewardship in resource-variable nursing home settings.

Background: Intravenous penicillin G is the preferred first-line agent for intrapartum Group B streptococcus prophylaxis; however, many pregnant patients carry unverified penicillin allergy labels. Although approximately 10% of the general population reports a penicillin allergy, only about 1% have a true allergy, and these inaccurate labels often lead to unnecessary use of broader-spectrum antibiotics, highlighting the need for improved allergy verification and de-labeling in obstetric care. Methods: A retrospective chart review was conducted from January 2022 to December 2024 evaluating documented penicillin allergies in pregnant patients who received care and delivered within a state-wide health system. Patients were excluded if they had a penicillin allergy documented during or after the encounter, or if the allergy was identified as an erroneous entry. A study-specific allergy risk assessment was used to stratify patients into no allergy, low risk, moderate-to-high risk, and severe risk categories, using a conservative approach as the true severity of historical reactions could not be definitively verified through retrospective chart review alone. Patients categorized as no allergy were considered eligible for de-labeling. Results: Of 303 birth encounters screened, 285 met inclusion criteria; ten patients with previously removed penicillin allergy labels were excluded from the primary outcome analysis. Among the remaining included encounters, 51 patients (18.5%) were identified as eligible for penicillin allergy de-labeling based on chart review alone. Eligibility was driven by documentation consistent with non-allergic reactions, including intolerance symptoms (5.8%), history of prior penicillin tolerance (10.5%), and reported familial allergy without personal reaction history (2.2%). Referral patterns for allergy evaluation were assessed to identify additional de-labeling opportunities. Only 12 patients were referred to an allergist during pregnancy; however, 11 of these patients underwent penicillin challenge testing and were successfully de-labeled. The high success rate among referred patients highlights frequent mislabeling of penicillin allergies and suggests substantial missed opportunities for broader de-labeling in the obstetric population. Conclusion: Nearly one in five pregnant patients (18.5%) with documented penicillin allergies were eligible for de-labeling based on chart review alone. Low referral rates for allergy evaluation also indicate significant opportunities to expand penicillin allergy testing in obstetric care, further improving antibiotic selection and antimicrobial stewardship.

A ceftriaxone safety alert prompted a review of rapid response and cardiac arrest events in relation to the timing of intravenous cephalosporin administrations across a large health system. Despite high utilization, we found a low rate of significant ceftriaxone-related adverse events with a similar incidence as other intravenous cephalosporins.

Background: Pediatric solid organ transplant recipients receive broad-spectrum antibiotics. The CDC NHSN Antimicrobial Use and Resistance (AUR) Module reports days of therapy (DOT) by pediatric Standardized Antimicrobial Administration Ratio (SAAR) categories, but those unit-level summaries do not identify the contexts that drive broad-spectrum DOT. We used a procedure-anchored framework to quantify post-prophylaxis broad-spectrum exposure and attribute it to blood-culture evaluation, surgical site infection (SSI), and subsequent procedures. Methods: We analyzed 164 pediatric transplant procedures (heart 59, liver 58, kidney 29, intestinal 8, multivisceral 10) with follow-up through postoperative day (POD) 365. Medication administration records were linked to procedures and transformed into an agent-day dataset with POD assignment. Organ-specific protocols defined a prophylaxis-window end day; post-prophylaxis analyses included antibiotic-days when POD exceeded end day. Broad-spectrum categories included Ped_SAAR_BSHO plus HighestRisk [Ped_SAAR_XDR (extensively drug-resistant) plus carbapenems]. SSI was chart-reviewed. A hierarchy assigned each post-prophylaxis broad-spectrum agent-day a primary context (culture, SSI, subsequent procedure, or residual unclassified). Antibiotic episodes (courses) used a locked gap rule (split if exposed-day gap ?3 days) and were linked to culture-days, SSIs or subsequent procedures to classify episode outcomes. Results: Over POD 0–365, patients contributed 10,064 hospital-days and received 9,774 antibacterial DOT (971.2/1,000 hospital-days); Ped_BSHO contributed 2,529 DOT (251.3/1,000 hospital-days). Ped_BSHO rates varied by organ (DOT/1,000 hospital-days: 76.3 heart, 273.0 liver, 105.4 kidney, 355.5 intestinal, 494.2 multivisceral). Following completion of index procedure prophylaxis, broad-spectrum exposure totaled 2,379 DOT-days; HighestRisk accounted for 667 DOT-days (435 multivisceral). Primary attribution classified post-prophylaxis broad-spectrum DOT-days as unclassified (982; 41.3%) or blood-culture rule-out (979; 41.1%); SSI treatment windows contributed 215 DOT-days (9.0%), and subsequent procedures contributed 66 DOT-days (2.8%). Unclassified DOT was concentrated (top 10 patients 64.8%) and occurred largely in intensive care units; piperacillin–tazobactam and meropenem predominated, and 24.5% of unclassified DOT-days were HighestRisk. From 1,390 blood culture accessions we defined 1119 sets, of which 94.7% (n=1060) were no growth, 3.4% (n=38) positive and 0.5% (n=6) contaminant. In the episode-based analysis, antibiotic courses that overlapped a blood-culture day were usually rule-out rather than culture-confirmed, and rule-out episodes contributed substantially more antibiotic burden than culture-positive episodes (DOT 1,510 vs 248). Culture intensity was high in multivisceral/intestinal PICU patients and PCICU heart patients. Conclusions: After prophylaxis ends, broad-spectrum exposure is dominated by evaluation-driven care and a large residual unclassified burden, while SSI-associated exposure remains a smaller, safety-anchored signal. This procedure-anchored, episode-based framework quantifies organ-specific burden and prioritizes targets for paired diagnostic and antibiotic stewardship.

Introduction: Clinicians often order a variety of tests for hospitalized patients, including urinalysis with urine cultures. However, fever has limited diagnostic value for urinary tract infection (UTI); pooled estimates indicate that fever has likelihood ratios near 1, offering minimal discriminatory value for ruling UTI in or out. We evaluated the diagnostic yield of inpatient urine cultures by fever status in an acute care safety-net hospital in Denver, Colorado. Methods: We analyzed inpatient urine culture events from April 2016 through December 2025 using an automated electronic medical record report. Fever was captured within the National Healthcare Safety Network (NHSN) infection window period (+/- 3 calendar days). Culture positivity was defined per NHSN UTI criteria as ?10 colony forming units of one or two organisms; cultures with no growth or reported as yeast/Candida spp. or mixed flora were classified as negative. Antibiotic exposure was extracted from medications given up to 3 days after culture collection, and urinary infection-related diagnoses from problem/diagnosis entries up to 7 days after culture collection. To scale adjudication, we used deterministic text classification of medication and diagnosis fields. Medication strings were normalized and parsed with regular expressions to extract generic drug names (token preceding dose/formulation), then matched to a curated antibiotic dictionary. Urinary diagnoses were identified by regular-expression matching to UTI-related terms (e.g., UTI/urinary tract infection, cystitis, pyelonephritis, urosepsis). Results: Across 17,629 inpatient urine cultures, fever within the NHSN infection window period was documented in 37.1% of episodes and UTI diagnoses were documented in 29.6%. Overall, 7,983 cultures (42.1%) were positive for a bacterial pathogen, and antibiotics were prescribed in 77.8% of episodes. Culture positivity was similar to, or higher than, that among afebrile vs febrile episodes (44.1% vs 38.8%) and was higher with documented UTI diagnoses (58.5% vs 35.2% without). Culture positivity varied by strata: UTI diagnoses were more common among culture-positive episodes (60.9% with fever; 57.1% without fever), while febrile episodes without UTI diagnoses had the lowest positivity (29.4%) (Figure 1). Antibiotic exposure remained common across the fever–culture result combinations, including 22.5% of all episodes occurring in afebrile, culture-negative patients who nevertheless received antibiotics and 18.9% in febrile, culture-negative patients who received antibiotics (Figure 2). Conclusion: Fever was not reliably associated with urine culture positivity, whereas UTI diagnosis (clinical suspicion) was. Antibiotic use remained common after negative urine cultures, supporting opportunities to refine inpatient urine culture ordering and antimicrobial stewardship.

Background: Hospital acquired infections are known to increase hospital stays and increase morbidity and mortality. Historically, Methicillin-resistance Staphylococcus aureus (MRSA) screening and contact precautions have been the primary strategy to prevent the spread of MRSA in healthcare systems. Increasing evidence supports the role for MRSA decolonization to decrease the spread of this important pathogen. The COVID-19 pandemic and related personal protective equipment conservation has led to many programs re-examining contact precautions for MRSA. Methods: In February of 2024 MRSA screening was discontinued for new admissions and an MRSA decolonization protocol was implemented for patients at high risk for MRSA infection as defined in the SHEA compendium for reducing MRSA infections. Decolonization included nasal Mupirocin 2% ointment and 2% Chlorhexidine Gluconate (CHG) bathing for five days in the adult inpatient population. High risk patients targeted for this intervention included patients transferred from another healthcare facility, admitted from home health, on dialysis, prior history of MRSA infection or those with central venous catheters. Contact isolation precautions for MRSA colonization were discontinued but maintained for active infection with MRSA. Results: After implementation, MRSA nares screening tests declined by 24.4% while utilization of mupirocin increased 5.85 times the pre-implementation average. In the twelve months prior to the intervention there were 221.1 new MRSA infections per 1000 inpatient admissions and in the 12 months following intervention there were 155.4 new MRSA infections per 1000 inpatient admissions. The rate of MRSA infections had been increasing prior to intervention and the rate of increase decreased by 54.8% after implementation. We estimated a cost savings of $194,745 if all patients that were previously screened were instead decolonized. We additionally noted an approximate 18% reduction in isolation gown use resulting an additional estimated annual cost savings of $115,369 per year if sustained. Conclusions: The implementation of a MRSA decolonization program resulted in significant cost savings without any associated increase in clinical MRSA cultures. The program appeared to have far greater buy-in than the previous screen and isolate program despite having similar patient selection criteria. This may reflect a greater perceived importance of screening relative to a treatment strategy. A decolonization strategy was found to be effective at decreasing the rate of new MRSA infections even with the removal of isolation precautions for colonization. Low compliance with isolation precautions for colonization are thought to be a contributing factor in the limited efficacy of a screen and isolate strategy.

Background: Staphylococcus aureus (SA) is a leading cause of infection-related mortality worldwide and remains one of the most common pathogens in healthcare settings, particularly methicillin-resistant SA (MRSA). Approximately 30% of patients are colonized with SA in the nares, and these asymptomatic carriers serve as a reservoir for invasive infections, including bloodstream infections, pneumonia, and surgical site infections. Targeted strategies, such as nasal decolonization, are essential to reduce colonization and prevent subsequent infection. We evaluated the impact of SA nasal decolonization on overall SA infection rates and healthcare facility-onset (HO) MRSA bacteremia, a laboratory-identified (LabID) event. Methods: In this pre–post quasi-experimental retrospective study, we compared overall SA, MRSA, and methicillin-susceptible SA (MSSA) infection rates per 1,000 patient days and the HO-MRSA bacteremia Standardized Infection Ratio (SIR) between the preintervention period (January 2020–June 2021) and the intervention period (October 2023–September 2025) at a 5-hospital healthcare system in Michigan. HO-MRSA bacteremia was determined using National Healthcare Safety Network (NHSN) criteria. SA infections included positive clinical cultures in hospitalized patients. Previously, our protocol consisted of daily chlorhexidine (CHG) bathing only, and HO-MRSA bacteremia SIR remained above goal. In a two-phased approach, we implemented nasal decolonization with an alcohol-based nasal antiseptic twice daily in addition to CHG for all ICU patients in August 2021, followed by all general practice unit patients with central lines in August 2023. Results: Following implementation, the cumulative HO-MRSA bacteremia SIR decreased 50% from 1.148 to 0.571 (95% CI, 0.336–0.73; p. Conclusion: Adding nasal decolonization to daily CHG bathing significantly reduced SA infection rates and HO-MRSA bacteremia. These findings confirm the effectiveness of a bundled decolonization strategy in mitigating both MRSA and MSSA infections. Broader adoption of nasal antiseptic protocols, alongside CHG bathing, should be considered a critical component of infection prevention strategy for high-risk hospital populations.