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

Background: Interventions targeting urine culture stewardship can improve diagnostic accuracy for urinary tract infections (UTI) and decrease inappropriate antibiotic treatment of asymptomatic bacteriuria. We aimed to determine if a clinical decision support (CDS) tool which provided guidance on and required documentation of the indications would decrease inappropriately ordered urine cultures in an academic healthcare network that already uses conditional (e.g. reflex) urine testing. Methods: In October 2022, four hospitals within one academic healthcare network transitioned to a new electronic health record (EHR). We developed an embedded CDS tool that provided guidance on ordering either a urinalysis (UA) with reflex to urine culture or a non-reflex urine culture (e.g. for pregnant patients) based on the indication for testing (Figure 1). We compared median monthly UA with reflex culture and non-reflex urine culture order rates pre- (8/2017–9/2022) and post- (10/2022–9/2023) intervention using the Wilcoxon rank-sum test. We used interrupted time-series analyses allowing a one-month time window for the intervention effect to assess changes in monthly UA with reflex culture, non-reflex urine culture, and total urine culture order rates associated with the intervention. Using SAS 9.4, we generated Durbin-Watson statistics to assess for autocorrelation and adjusted for this using a stepwise autoregressive model. Result: The median monthly UA with reflex culture order rates per 1000 patient-days were similar pre- and post- intervention at 36.7 (interquartile range [IQR]: 31.0–39.7) and 35.4 (IQR: 32.8–37.0), respectively (Figure 2). Non-reflex and total urine culture rates per 1000 patient-days decreased from 8.5 (IQR: 8.1–9.1) to 4.9 (IQR: 4.7–5.1) and from 20.0 (IQR: 18.9–20.7) to 14.4 (IQR: 14.0–14.6) post-intervention, respectively. Interrupted time-series analyses revealed that the intervention was associated with a decrease in the monthly non-reflex urine culture by 4.8 cultures/1000 patient-days (p< 0.001) and in the total urine culture monthly order rates by 5.0 cultures/ 1000 patient-days (p < 0 .001) [Figures 3a and b]. The UA with reflex order rate did not significantly change with the intervention (not pictured). Conclusion: In an academic healthcare network that already employed conditional urine testing, the implementation of an EHR-based diagnostic stewardship tool led to additional decreases in both non-reflex and total urine cultures ordered.

OBJECTIVES/GOALS: Identify causes for clinical research professional turnover Define data collection methods for exit interviews Provide institutions with resources to collect and analyze exit interviews Employ strategies to maximize the impact of exit interviews on retention METHODS/STUDY POPULATION: The Clinical Research Professional Taskforce (CRPT) exit interview Subgroup has met monthly since January 2023. Action items were agreed to and minutes were kept and reviewed at subsequent virtual working meetings. All members were given opportunity to speak and contribute. After a landscape analysis, conducted via survey, five institutions agreed to provide examples of their exit interview questions. Members spoke at length about goals, methods, collection techniques, institutional involvement, lessons learned and practical applications that could become best practices. RESULTS/ANTICIPATED RESULTS: The Subgroup aggregated all questions into categories and developed sample questions incorporating all data without using any word for word. In order to allow for quantitative assessment and standardized reporting the Subgroup formulated questions to be responded to utilizing a Likert scale with free text fields for select questions where further information is needed. The Subgroup developed best practices describing decision-making metrics, understanding reasons for turnover and reporting data back to leadership. Practical aspects such as method and time of survey collection, anonymity, and training staff are also included. DISCUSSION/SIGNIFICANCE: We are hopeful that sample questions and best practices will be helpful and widely utilized. Understanding the causes and impacts of CRP turnover are critical to meeting the current needs of clinical research. Further work is being done to calculate the cost of turnover to make the business case.

State Medical Boards (SMBs) can take severe disciplinary actions (e.g., license revocation or suspension) against physicians who commit egregious wrongdoing in order to protect the public. However, there is noteworthy variability in the extent to which SMBs impose severe disciplinary action. In this manuscript, we present and synthesize a subset of 11 recommendations based on findings from our team’s larger consensus-building project that identified a list of 56 policies and legal provisions SMBs can use to better protect patients from egregious wrongdoing by physicians.