Infection Control and Hospital Epidemiology最新文献

Background: Surveillance activities are emerging as exemplar use cases for large language models (LLMs) in health care. The aim of this study was to evaluate the potential for LLMs to support the expansion of surveillance activities to include cardiovascular implantable electronic device (CIED) procedures.

Methods: A validated machine learning-based infection flagging tool was applied to a cohort of VA CIED procedures from 7/1/2021 to 9/30/2023; cases with ≥10% probability of CIED infection underwent manual review. Then, a weighted random sample of 50 infected and 50 uninfected cases was reviewed with generative artificial intelligence (GenAI) assistance. GenAI prompts were iteratively refined to extract and classify all components of infection-related variables from clinical notes. Data extracted by GenAI were compared with manual chart reviews to assess infection status and extraction consistency.

Results: Among 12,927 CIED procedures, 334 (2.58%) had ≥10% probability of CIED infection. Among 100 sampled cases, 50 of 50 uninfected cases were correctly categorized. Among 50 infection cases, GenAI identified all CIED infections, but the timing of events and the attribution to a preceding procedure were incorrect in 7 of 50 cases. The overall specificity of the GenAI-assisted process was 100% and the sensitivity for accurately classifying timing and attribution of CIED infection events was 82%. Errors in timing improved with iterative prompt updates. Manual chart reviews averaged 25 minutes per chart; the GenAI-assisted process averaged 5-7 minutes per chart.

Conclusions: LLMs can help streamline the review process for healthcare-associated infection surveillance, but manual adjudication of output is needed to ensure the correct timeline of events and attribution.

Objectives: To evaluate the implementation and effectiveness of a novel home infusion central line-associated bloodstream infection (CLABSI) and home infusion-onset bloodstream infection (HiOB) dashboard and prevention toolkit.

Design: Mixed methods study.

Setting and participants: Five home infusion agencies participating in the first CLABSI prevention collaborative.

Methods: Agencies uploaded CLABSI and HiOB data to a comparative dashboard. The dashboard started in December 2022 and accepted retrospective data from June 2021. A CLABSI prevention toolkit was made available in June 2024. Using an interrupted time series, we present CLABSI and HiOB rates before and after dashboard and toolkit implementation. We surveyed and interviewed participants about the tools and toolkit, using directed content analysis to analyze the interviews.

Results: After dashboard implementation, there was a decrease in CLABSI (-0.23/10,000 home-CVC days, 95% CI -0.28 to -0.18) and HiOB (-0.25/10,000 home-CVC days, 95% CI: -0.31 to -0.18) over time. With toolkit implementation, there was a further decrease in CLABSI (-0.17/10,000 home-CVC days, 95% CI: -0.30 to -0.044) and HiOB (-0.23/10,000 home-CVC days, 95% CI: -0.37 to -0.089) over time. Themes were associated with use of the tools (accessible, adaptable, patient-centered tools; user-friendly education to enhance understanding; barriers identified; tool mismatches; and strategies for tool delivery) and toolkit implementation (structural barriers, user-centered design, collaborative engagement and communication, toolkit used to enhance workforce competency, and concerns related to consistency).

Conclusions: Implementation of a dashboard and a CLABSI prevention toolkit were each associated with both CLABSI and HiOB reduction in a collaborative of home infusion agencies.

Background: Early identification of patients colonized with multidrug-resistant organisms (MDROs) facilitates infection control interventions. We assessed a Public Health Risk Model's ability to predict carbapenem-resistant Enterobacterales and other MDROs.

Methods: We retrospectively analyzed a medical intensive care unit patient cohort screened at time of admission for MDRO carriage (1/2017-1/2018). Encounters were linked to Illinois Hospital Discharge Data and assigned a public health risk model probability score. We compared the model's performance to traditional screening strategies that use variables locally available to clinicians at time of admission (i.e., transfer from other hospital, tracheostomy, gastrostomy, pressure ulcer). Model discrimination was evaluated by quantifying the area under the curve (AUC). For each approach, we assessed sensitivity, specificity, and number needed to screen (NNS) to detect one MDRO carrier.

Results: Model probability calculation was successful in 1237/1250 (98.9%) admissions. The model identified carbapenem-resistant Enterobacterales colonization well (AUC 0.82) and generalized to predict colonization with other healthcare-associated MDROs, including carbapenem-resistant Pseudomonas aeruginosa (AUC 0.82) and vancomycin-resistant enterococci (AUC 0.76). The model did not predict MDROs with known local community reservoirs, i.e., third-generation cephalosporin-resistant Enterobacterales (AUC 0.61) and methicillin-resistant Staphylococcus aureus (AUC 0.59). At the same NNS, the model had higher sensitivity compared to use of traditional screening strategies (68% versus 41%).

Conclusion: A risk model using patient-level healthcare exposure data from a state public health dataset identified critically ill patients likely to harbor healthcare-associated MDROs at the time of admission.

Objective: We aimed to evaluate the impact of transitioning from a conventional antimicrobial stewardship team (AST) to an interdisciplinary infectious disease support team (IDST) on patient mortality.

Design: Single-center retrospective cohort study.

Setting: Yamanashi Prefectural Central Hospital.

Patients: We included patients with bacteremia during the AST and IDST periods.

Methods: We implemented an interdisciplinary IDST that provided comprehensive support beyond antimicrobial selection and treatment. We compared 28-day mortality outcomes in hospitalized patients with bacteremia during the AST-only period (April 2021-March 2022) and the IDST-intervention period (April 2022-March 2023). Cox proportional hazards modeling was used to assess 28-day mortality; multivariable logistic regression served as a prespecified sensitivity analysis. Between-group differences were tested using two-sided tests (P < .05), and 95% confidence intervals (CIs) were reported.

Results: Among the 854 patients (413 in the AST group and 441 in the IDST group), AST interventions did not significantly decrease 28-day mortality compared with no-intervention cases (20.8% vs 23.3%, P = .630). However, intervention by the IDST significantly reduced 28-day mortality compared with the no-IDST group (16.1% vs 25.2%, P = .019). In the multivariate Cox proportional hazards model, IDST interventions remained independently associated with a reduced 28-day mortality (hazard ratio, 0.518; 95% CI, 0.325-0.826; P = .006).

Conclusions: Transitioning from an AST to an interdisciplinary IDST model was associated with lower 28-day mortality in bacteremia cases, possibly due to sustained diagnostic and therapeutic interventions. Our findings support stewardship as a team-based process and may significantly impact patient outcomes by maintaining involvement until treatment completion and clinical stabilization.

Blood culture contamination (BCC) leads to increased costs and patient harms. We reviewed 525 BCC cases and found 71.2% of BCC cases were ordered for indications with low risk for bacteremia and most received unnecessary tests and antibiotics. Diagnostic stewardship of blood cultures may reduce BCC and its associated costs.

This single center retrospective observational study of serial plasma metagenomic next-generation sequencing testing shows that >95% of serial testing was without meaningful clinical impact. Only 5/173 cases were adjudicated as having significant clinical impact.

Objective: To develop an approach for creating facility-specific urinary antibiograms accounting for the low number of isolates recovered in nursing homes (NHs).

Design: Retrospective analysis of urine culture data collected in NHs in five states.

Setting: Data on 5097 urine culture isolates collected across 59 study NHs from January 1, 2020 to December 31, 2021. Four consulting microbiology laboratories served the study homes.

Methods: We compared a Clinical and Laboratory Standards Institute (CLSI) standard antibiogram model to four weighted-incidence syndromic antibiogram (WISCA) models utilizing alternate formatting rules. Ability to produce a facility-specific antibiogram with at least 30 isolates and the impact on susceptibility predictions were compared.

Results: Only one facility could generate a CLSI standard antibiogram for the three most commonly recovered Gram-negative isolates over a one-year period. Ability to generate an antibiogram increased with each of the four WISCA models trialed (36%, 54%, 85%, 85%) with the most successful models combining all Gram-negative isolates over a two-year period. Shortening the definition of duplicate isolates from 12 to 3 months did not improve performance. Using all Gram-negative isolates, rather than the three most recovered pathogens, resulted in meaningful changes in the predicted activity of ampicillin-sulbactam, cefazolin, ceftriaxone, and trimethoprim-sulfamethoxazole in several study NHs.

Conclusions: These results suggest that WISCAs using 2-years of urinary culture data including all gram-negative isolates and excluding duplicate isolates within twelve months maximizes the number of NHs able to create a valid antibiogram.

Objective: To assess differences in SARS-CoV-2 infection rates between patients receiving hemodialysis in outpatient centers (in-center) and those receiving dialysis in their homes (hemodialysis and peritoneal dialysis) from December 29, 2020, through May 9, 2023.

Design: Retrospective cohort study.

Setting: Outpatient dialysis facilities in the United States reporting to the Centers for Disease Control and Prevention's National Healthcare Safety Network.

Patients: Maintenance dialysis patients that received hemodialysis treatment at or were affiliated with outpatient dialysis facilities.

Methods: SARS-CoV-2 infection rates were assessed by dialysis setting (in-center and home). Weeks were categorized as surge (rate of infection > median) and non-surge (rate of infection ≤ median) and by variant predominance. A negative binomial regression model with generalized estimating equations was constructed to examine differences in rates of infection among patients.

Results: A total of 7,974 dialysis facilities reported 171,338 SARS-CoV-2 infections among patients. In-center hemodialysis patients had higher average rates of SARS-CoV-2 infection at 2.85 infections per 1000 patient-weeks than home patients at 1.69 infections per 1000 patient-weeks. During surge weeks, the differences in rates of infection between in-center and home patients were more pronounced than during non-surge weeks for all variant predominance categories: Delta (relative rate ratio (RRR) = 1.20, CI: 1.09-1.32), B.1 and Other (RRR = 1.11, CI: 1.02-1.22), and Omicron (RRR = 1.07, CI: 1.01-1.12).

Conclusion: Rates of SARS-CoV-2 infection among patients receiving outpatient hemodialysis were persistently higher than rates among patients receiving dialysis treatments at home; these differences were more pronounced during surge weeks.