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

Introduction: Vancomycin-associated acute kidney injury (V-AKI) is a common adverse reaction; however, there is currently no method to systematically monitor its incidence. We previously developed and internally validated an electronic algorithm to identify cases of V-AKI using structured electronic health record data at the Johns Hopkins Hospital, which demonstrated excellent agreement with chart review (percent agreement 92.5%; weighted kappa coefficient 0.95), as well as excellent sensitivity (89.7%) and specificity (98.2%) in detecting at least possible V-AKI events. The objective of this study was to evaluate the generalizability of the V-AKI electronic algorithm. Methods: We identified a retrospective cohort of adult and pediatric patients who received ≥1 dose of intravenous vancomycin while admitted to University of Virginia (UVA) Medical Center from 1/2021-1/2023. An increase in creatinine (Cr) of ≥0.3 mg/dL within 48 hours or ≥50% increase in baseline Cr within 7 days, occurring after the first dose and up to 72 hours after the last dose of IV vancomycin, was considered a potential V-AKI event. The electronic algorithm was executed at UVA with only limited contextualization of hospital specific variables (e.g., procedure names). Patients were categorized as excluded/not meeting criteria, or as having an unlikely, possible or probable V-AKI event using a causality framework. A random subset of the cohort underwent chart review by a blinded reviewer for external validation. Percent agreement and a weighted kappa coefficient were calculated. The sensitivity and specificity in identifying at least possible V-AKI events was determined. Results: The electronic algorithm was validated using 200 cases and demonstrated 60.0% percent agreement with chart review (Figure). The weighted kappa coefficient was 0.75. The algorithm was 83.8% sensitive and 71.4% specific in detecting at least possible V-AKI events. Among the 80 discrepant cases, there was only a 1-category difference in 62.5% of cases. The most common reasons for discrepant assessments, which were partly due to inconsistencies in chart review, included disagreement regarding timing of AKI onset (18.6%) and whether renal function returned to baseline (16.3%). Conclusions: An electronic algorithm to identify V-AKI events was successfully implemented at another institution. Although agreement with chart review was only fair, sensitivity in detecting at least possible V-AKI events remained excellent. The electronic algorithm may be useful for systematically and reproducibly identifying V-AKI events across institutions in a scalable manner to inform stewardship interventions. However, further refinement of the algorithm and improvement in consistency of chart review assessments is needed.

Disclosure: Lindsay Donohue: Advisor – Abbvie

Cognitive theories of depression contend that biased cognitive information processing plays a causal role in the development of depression. Extensive research shows that deeper processing of negative and/or shallower processing of positive self-descriptors (i.e., negative and positive self-schemas) predicts current and future depression in adults and children. However, the neural correlates of the development of self-referent encoding are poorly understood. We examined children's self-referential processing using the self-referent encoding task (SRET) collected from 74 children at ages 6, 9, and 12; around age 10, these children also contributed structural magnetic resonance imaging data. From age 6 to age 12, both positive and negative self-referential processing showed mean-level growth, with positive self-schemas increasing relatively faster than negative ones. Further, voxel-based morphometry showed that slower growth in positive self-schemas was associated with lower regional gray matter volume (GMV) in ventrolateral prefrontal cortex (vlPFC). Our results suggest that smaller regional GMV within vlPFC, a critical region for regulatory control in affective processing and emotion development, may have implications for the development of depressogenic self-referential processing in mid-to-late childhood.

Irritability is a transdiagnostic feature of diverse forms of psychopathology and a rapidly growing literature implicates the construct in child maladaptation. However, most irritability measures currently used are drawn from parent-report questionnaires not designed to measure irritability per se; furthermore, parent report methods have several important limitations. We therefore examined the utility of observational ratings of children's irritability in predicting later psychopathology symptoms. Four-hundred and nine 3-year-old children (208 girls) completed observational tasks tapping temperamental emotionality and parents completed questionnaires assessing child irritability and anger. Parent-reported child psychopathology symptoms were assessed concurrently to the irritability assessment and when children were 5 and 8 years old. Children's irritability observed during tasks that did not typically elicit anger predicted their later depressive and hyperactivity symptoms, above and beyond parent-reported irritability and context-appropriate observed anger. Our findings support the use of observational indices of irritability and have implications for the development of observational paradigms designed to assess this construct in childhood.

This SHEA white paper identifies knowledge gaps and challenges in healthcare epidemiology research related to coronavirus disease 2019 (COVID-19) with a focus on core principles of healthcare epidemiology. These gaps, revealed during the worst phases of the COVID-19 pandemic, are described in 10 sections: epidemiology, outbreak investigation, surveillance, isolation precaution practices, personal protective equipment (PPE), environmental contamination and disinfection, drug and supply shortages, antimicrobial stewardship, healthcare personnel (HCP) occupational safety, and return to work policies. Each section highlights three critical healthcare epidemiology research questions with detailed description provided in supplementary materials. This research agenda calls for translational studies from laboratory-based basic science research to well-designed, large-scale studies and health outcomes research. Research gaps and challenges related to nursing homes and social disparities are included. Collaborations across various disciplines, expertise and across diverse geographic locations will be critical.