Explore the vast range of titles available.

Background: Approximately 40% of all antibiotics related to an acute-care hospital stay are prescribed at the time of hospital discharge. However, there is no metric to compare hospitals on their antibiotic-prescribing at this transition of care. In this study, we sought to build a risk-adjusted metric for comparing hospitals on their overall post-discharge antibiotic use. Methods: This was a retrospective study across all acute-care admissions within the Veterans Health Administration during 2018-2021. For patients discharged to home or self-care, data on antibiotics administered while inpatient and those prescribed at discharge were collected. To predict post-discharge antibiotic use (days of therapy, DOT), we built a zero-inflated negative binomial model with a random intercept for each VA medical center. Data were split into training and testing sets to measure model performance and absolute error. Covariates included patient demographics, medical specialty at discharge, comorbidities, discharge diagnoses of infection, and the length of inpatient antibiotic therapy. Outliers, defined as DOT ≥ 30, were excluded, and the predicted random intercept was used to determine hospital performance. To compare hospitals with a positive versus negative random intercept in our model (i.e. higher vs. lower than expected overall post-discharge use, respectively), we calculated mean total antibiotic duration (inpatient + post-discharge) for two uncomplicated infection types: community-acquired pneumonia (CAP) and skin and soft tissue infections (SSTI). Results: 1,804,400 patients were discharged to home or self-care across 130 hospitals. The mean age was 67.8 (SD 12.9), and 93.7% were male. Antibiotics were prescribed to 41.5% while hospitalized and 19.5% at discharge. The median number of post-discharge DOT among those prescribed post-discharge antibiotics was 7 (IQR 4-12). The predictive model detected post-discharge antibiotic use with fidelity, including accurate identification of any post-discharge antibiotic exposure (area under the precision-recall curve=0.97 ) and reliable prediction of the number of post-discharge DOT in those who were exposed (mean absolute error = 1.65; Figure 1). At negative versus positive random intercept hospitals (Figure 2), antibiotic duration for CAP and SSTIs was 7.3 versus 8.1 days (p < 0 .001) and 9.4 vs. 10.2 days (p < 0 .001), respectively. Conclusion: A model using electronically available data was able to accurately predict antibiotic use prescribed at hospital discharge. Hospitals with lower than expected overall post-discharge antibiotic use also prescribed shorter courses of antibiotic therapy for uncomplicated cases of CAP and SSTI, which may reflect more robust processes at these sites to reduce antibiotic overuse at discharge. Disclosure: Michi Goto: Contracted Research – Merck

Background: Chlorhexidine gluconate bathing (CHGB) prevents healthcare associated infections (HAIs). CHGB quality is rarely assessed; prior studies identified that concentrations of CHG can be suboptimal, particularly at the neck, and if rinsed after application. In the setting of increased HAI rates on 3 high-risk units, we evaluated CHG skin concentrations, comparing results to bathing documentation and patient reports as part of a quality improvement initiative. Methods: All patients admitted to 3 high-risk units were swabbed for CHG concentration testing at the neck, bilateral upper arms, and groin. Swabs were processed using a semi-quantitative colorimetric CHG assay. A threshold of 0.001875% CHG was used to determine adequacy based on prior studies. Adequacy was assessed by body site, timing of bath, and patient-reported skin care activities using Chi-square tests in SAS 9.4. Per hospital policy, all admitted patients are bathed daily with 2% CHG pre-packed wipes. Patients without a documented CHGB for the duration of the admission were excluded. Results: CHG testing was completed on 63 patients: 23 on medical ICU, 18 surgical ICU, 22 oncology ward, yielding 249 samples. Only ward patients could report the time of last CHGB, which agreed with nursing documentation for 12/21(57%) Adequacy by sample was no different across units: 59/88(67%) Oncology, 68/90(76%) MICU, 56/71(79%) SICU, p=0.2091. Site adequacy was different by site: neck 36/63(57%), left arm 49/62(79%), right arm 50/62(81%), groin 48/62(77%), p=0.0083. Samples taken from the 11 patients with > = 24 hours since last CHGB were more likely to be below threshold concentration: 19/47(40%) versus 47/202(23%) not adequate in the recent treatment grouping. Three patients reported showering soon after the CHGB and 8 patients used moisturizing lotion. The percent of samples below threshold for the showering patients (6/12, 50%) and lotion-users (11/32, 34%) were not significantly different from the non-showering or non-lotion using patient samples (p=0.0588 and 0.2800 respectively). Conclusion: In a facility with longstanding daily CHGB policies in place, 66/249 samples from 63 patients lacked adequate concentrations of CHG for optimal HAI prevention. Even in patients with recent CHGB, 23% of sites tested revealed inadequate levels of CHG, while 60% of those overdue for CHGB kept adequate concentrations. Reliable implementation strategies are required for CHGB so as to ensure maximal infection prevention impact.

Oral presentation is a popular type of assessment in undergraduate degree programs. However, presentation delivery and grading pose considerable challenges to students and faculty alike. For the former, many students who learn English as an additional language may fear giving oral presentations in English due to a lack of confidence. For the latter, faculty who teach multiple classes and have many students may find it difficult to spend adequate time helping students refine their communication skills. This study examines an AI-assisted presentation platform that was built to offer students more opportunities for presentation training without the need for faculty intervention. Surveys with students and teachers were conducted to inform the design of the platform. After a preliminary platform was developed, two methods were employed to evaluate its reliability: a beta test with 24 students and a comparison of AI and human scoring of the presentation performance of 36 students. It was found that students are highly receptive to the platform, but there are noticeable differences between AI and human scoring abilities. The results reveal some limitations of AI and human raters, and emphasize the potential benefit of exploring collaborative AI–human intelligence.

Background: Inappropriate antibiotic use impacts patient safety and antimicrobial resistance patterns. In 2013, general dentists in the U.S. prescribed nearly 10% of all outpatient oral antibiotics (24.5 million prescriptions). The American Dental Association (ADA) published guidelines in 2019 recommending limited antibiotic prescribing for the treatment of dental pain and swelling. We characterized dental prescribing during 2018–2022 to assess whether antibiotic use decreased after the guideline’s release. In addition, we examined access to dental care. Methods: All antibiotic prescriptions dispensed during 2018–2022 were extracted from the IQVIA Xponent database, which captured ≥92% of all U.S. outpatient prescriptions and projected to 100% coverage. Prescriptions by general dentists were compared to total outpatient oral antibiotic prescriptions and summarized by patient sex, patient age, and prescriber geographic region. Census denominators were used to calculate prescribing rates per 1,000 persons. IQVIA general dentist counts were used to calculate dentists per 100,000 persons. Results: General dentists prescribed 24.7 million antibiotic prescriptions in 2018 (75 prescriptions per 1,000 persons) compared with 25.2 million (76 prescriptions per 1,000 persons) in 2022. During 2020–2022, general dentists prescribed >10% of all outpatient antibiotic prescriptions (range 10.7%–12.1%). In each year, prescription rates were higher for females, patients > 65 years, and among prescribers in the Northeast. In 2022, there were 58 general dentists per 100,000 persons in the United States. The highest general dentist rate was in District of Columbia (100 per 100,000 persons) and the lowest rate was in Delaware (41 per 100,000 persons). Conclusions: Despite the ADA’s 2019 guidelines, prescribing by general dentists remained stable during 2018–2022. Because the total number of antibiotic prescriptions overall decreased, general dentists’ share of all outpatient antibiotic prescriptions increased to >10% in recent years. Rate variation by patient characteristics and prescriber region may reflect differences in dental disease burden or may represent unnecessary antibiotic use. Dental antibiotic stewardship is needed, including dissemination and implementation of current prescribing guidelines. Further evaluation of prescribing indications and access to dental care is needed to inform dental stewardship priorities.

Background: Community Acquired Pneumonia (CAP) is the most common reason for antibiotic treatment in hospitalized adults. Some prior studies have found treatment differences by race/ethnicity but research on the topic is limited, results are mixed, and it is unclear if clinical outcomes are affected. We sought to examine whether guideline-concordant CAP care and patient outcomes varied by race/ethnicity. Methods: Using the Vizient clinical database, we conducted a cross sectional analysis of all hospitalized patients > = 18 years of age with a primary diagnosis of pneumonia (ICD10 codes: J12-J18) from 2018-2021. Univariate and bivariate analyses examined the distribution of demographic, clinical and hospital characteristics across race/ethnicity. The primary outcome was receipt of therapy concordant with ATS/IDSA Clinical Practice Guideline for CAP. Final models included only patients with bacterial pneumonia and examined the relationship between race/ethnicity and guideline-concordant antibiotic treatment. Secondary analysis examined the interaction between race/ethnicity and concordant antibiotic treatment with length of stay >7 days, 30-day hospital readmission, adverse events or complications in separate models. We used hierarchical multivariable regression models accounting for clustering within patients and among patients hospitalized at the same facility. Due to sample size, significance was assessed with an OR > = 1.2 and p≤ 0.05. All analyses used SAS (v.9.4, SAS Institute Inc. Cary, NC). Results: There were 1,277,770 admissions with a primary diagnosis of bacterial CAP. Sixty-nine percent of the sample was White, 18% Black, 8% Hispanic, 2% Asian and 3% identified as other. 56% of the sample received concordant care. In adjusted models Black patients had greater odds of overall concordant care (OR 1.22; p 7 days (OR 0.67 p <.0001), complication or adverse event (OR 0.75 p <.0001), but not readmission within 30 days. Conclusion: We observed differences between Black and White patients in the receipt of concordant treatment. Hospital bed size, CMI and region played an important role in both antibiotic treatment decisions and clinical outcomes, indicating that hospital and regional prescribing cultures may play in role in treatment inequities.

Background: The National Healthcare Safety Network (NHSN) provides detailed surveillance case definitions for healthcare-associated infections (HAI), including central line-associated bloodstream infections (CLABSI). CLABSI data are used for several purposes, including improving patient safety, value-based purchasing, and comparing hospitals’ performance. Our Infection Prevention (IP) team conducts house-wide HAI surveillance. To ensure that our hospital CLABSI reporting is accurate and that staff are implementing case definitions consistently and systematically, we conducted an internal validation of CLABSI. This undertaking allowed us to identify educational opportunities for IPs and improve surveillance data consistency. Methods: At UNC Hospitals, data on all positive blood cultures collected in the inpatient setting from July 2022 – June 2023 were obtained from electronic medical records. A random number generator was used to select 16 records per quarter. Each record was then randomly assigned to two different IPs (out of 8 total inpatient IPs) for review. Concordance of CLABSI classification was summarized across the two reviews and compared to the initial review. Discordant cases were then reviewed by the Associate IP Director (a certified IP with 15 years of experience) for final adjudication. A summary of findings and discordant cases details were discussed at regular IP educational meetings. Results: From July 2022-June 2023, there were 1658 positive blood cultures collected in the inpatient setting. Of the 64 randomly selected blood cultures, total concordance amongst all reviewers occurred 65.6% of the time. Concordance improved in the 2nd half of FY23 compared to the 1st half (72% vs, 59%, p>0.05). Amongst the 33% of blood culture results with reviewer discrepancy, the most common reasons were related to distinction of a bloodstream infection secondary to another infection site (32%) and application of the repeat infection timeframe (18%). Importantly, there was only one instance where a blood culture result was categorized by all 3 reviewers as present on admission, but upon Associate Director review, actually represented a CLABSI (i.e., false negative). Conclusions: Standardized case definitions remain open to interpretation. At our hospital, we experienced discordance in approximately one-third of instances during review of blood culture data amongst trained infection preventionists. Reviewing all blood culture data is key for validation so that both false positives and false negative CLABSIs can be identified. Identifying the most common reasons for discordance and using specific examples when case disagreement occurred for educational purposes may lead to improved reliability and accuracy of application of the NHSN surveillance defintions.

Background: The number of advanced practice providers (APPs)—nurse practitioners (NPs) and physician assistants (PAs)—continues to expand across the United States. Several studies suggest differences in antibiotic prescribing rates and appropriateness by APPs compared to physicians. The objective of this analysis is to characterize population- and provider-specific outpatient antibiotic prescribing rates among physicians and APPs nationally, by state, and within urban versus rural counties. Methods: We estimated outpatient oral antibiotic prescription rates for 2011 and 2022 using county-level prescription dispensing data from IQVIA Xponent® (numerator) and population census estimates (denominator). Provider specialty denominators were provided by IQVIA, based on data from the American Medical Association. Counties were classified as urban or rural per the 2013 National Center for Health Statistics classification. National and state-level prescription volume, rates per 1000 population, and average number of prescriptions per provider were calculated for physicians, NPs, and PAs. We assessed the degree to which provider-specific rates explained the variance of the overall rate by state, using the coefficient of determination (r2) from Pearson’s correlation. Results: Between 2011 and 2022, overall U.S. antibiotic prescribing declined from 877 to 709 per 1000 population, a 19.2% relative reduction. The provider-specific proportion of the overall prescribing rate relatively decreased by 32% for physicians but increased by 157% for APPs (NPs 229%, PAs 86%; Figure 1). State-level antibiotic prescribing rates varied by provider type for both years, shifting towards proportionally greater APP prescribing in 2022 (Figure 2). For 2011 and 2022, physician prescribing rate strongly correlated with the overall state rate (r2 = 0.83 in 2011 versus 0.80 in 2022), whereas the correlation of the NP prescribing rate increased (r2 = 0.20 in 2011 versus 0.76 in 2022). A total of 60,327 (7.2%) physicians practiced in rural settings in contrast to 42,876 (12%) NPs and 14,495 (9.4%) PAs in 2022. Providers in rural counties prescribed more antibiotics per provider on average compared to urban counties; rural physicians prescribed 57% more antibiotics per provider (207 vs 132 antibiotics per provider), rural NPs prescribed 115% more (284 vs 132), and rural PAs prescribed 53% more (289 vs 189). Conclusions: The relative contribution of APPs to outpatient antibiotic prescriptions more than doubled over the past decade, accounting for 1 in 3 prescriptions in 2022. This contribution was especially prominent among NPs in rural counties. Further evaluation of antibiotic prescribing appropriateness among APPs and integration of APPs into antibiotic stewardship efforts in various settings.

Background: Artificial intelligence (AI) tools have demonstrated success in US medical licensing examinations; however, their utility in infection prevention and control (IPC) remains unknown. Methods: The program of hospital epidemiology handles consultation calls and records each question and answer. Using 2022 data, we selected 31 frequently asked questions. We utilized four AI tools, including Chat GPT-3.5 and 4.0, Bing AI, and OpenEvidence, to generate answers. We predefined scales (Table 1) to capture responses by three reviewers, including two hospital epidemiologists and one infection preventionist. The mean score of ≥ 3 and ≥ 4 was considered acceptable in accuracy and completeness, respectively. We reported the percentage of responses with acceptable accuracy and completeness out of assessed questions for each category. Results: Among 31 questions, 16 were associated with isolation duration, 9 with healthcare personnel (HCP) exposure, 4 with cleaning contaminated rooms, and 2 with patient exposure. Regarding accuracy, most AI tools performed worse in questions about isolation duration, ranging between 75% and 93.8%. All AI tools, except OpenEvidence, had a 100% accuracy rate for HCP and patient exposure. All AI tools had a 100% accuracy rate for contaminated room handling. The highest overall acceptable accuracy rate was observed in Chat GPT-3.5. Regarding completeness, most AI tools performed worse in questions about isolation duration, ranging between 44% and 75%. All AI tools, except OpenEvidence, had a 100% completeness rate for contaminated rooms and patient exposure. The highest overall acceptable completeness rate was observed in Bing AI (Table 2). Conclusions: All AI tools provided reasonable answers to commonly asked IPC-related questions, although, there were variations among different tools used. AI could be used to supplement the infection control program, especially if resources are limited.

Background: External comparisons of antimicrobial use (AU) may be more informative if adjusted for encounter characteristics. Optimal methods to define input variables for encounter-level risk-adjustment models of AU are not established. Methods: This retrospective analysis of electronic health record data included 50 US hospitals in 2020-2021. We used NHSN definitions for all antibacterials days of therapy (DOT), including adult and pediatric encounters with at least 1 day present in inpatient locations. We assessed 4 methods to define input variables: 1) diagnosis-related group (DRG) categories by Yu et al., 2) adjudicated Elixhauser comorbidity categories by Goodman et al., 3) all Clinical Classification Software Refined (CCSR) diagnosis and procedure categories, and 4) adjudicated CCSR categories where codes not appropriate for AU risk-adjustment were excluded by expert consensus, requiring review of 867 codes over 4 months to attain consensus. Data were split randomly, stratified by bed size as follows: 1) training dataset including two-thirds of encounters among two-thirds of hospitals; 2) internal testing set including one-third of encounters within training hospitals, and 3) external testing set including the remaining one-third of hospitals. We used a gradient-boosted machine (GBM) tree-based model and two-staged approach to first identify encounters with zero DOT, then estimate DOT among those with >0.5 probability of receiving antibiotics. Accuracy was assessed using mean absolute error (MAE) in testing datasets. Correlation plots compared model estimates and observed DOT among testing datasets. The top 20 most influential variables were defined using modeled variable importance. Results: Our datasets included 629,445 training, 314,971 internal testing, and 419,109 external testing encounters. Demographic data included 41% male, 59% non-Hispanic White, 25% non-Hispanic Black, 9% Hispanic, and 5% pediatric encounters. DRG was missing in 29% of encounters. MAE was lower in pediatrics as compared to adults, and lowest for models incorporating CCSR inputs (Figure 1). Performance in internal and external testing was similar, though Goodman/Elixhauser variable strategies were less accurate in external testing and underestimated long DOT outliers (Figure 2). Agnostic and adjudicated CCSR model estimates were highly correlated; their influential variables lists were similar (Figure 3). Conclusion: Larger numbers of CCSR diagnosis and procedure inputs improved risk-adjustment model accuracy compared with prior strategies. Variable importance and accuracy were similar for agnostic and adjudicated approaches. However, maintaining adjudications by experts would require significant time and potentially introduce personal bias. If findings are confirmed, the need for expert adjudication of input variables should be reconsidered. Disclosure: Elizabeth Dodds Ashley: Advisor- HealthTrackRx. David J Weber: Consultant on vaccines: Pfizer; DSMB chair: GSK; Consultant on disinfection: BD, GAMA, PDI, Germitec