115. Variable Use of Diagnostic Codes for Acute Respiratory Infections Across Emergency Departments and Urgent Care Clinics in an Integrated Healthcare System: Implications for Accuracy of Antibiotic Stewardship Metrics

Abstract Background Antibiotic stewardship initiatives can leverage metrics that make peer-peer comparisons. A commonly used metric measures how frequently a clinician prescribes antibiotics for acute respiratory infections (ARIs), as defined by diagnostic codes. However, it is unclear if clinicians differ in their use of ARI diagnostic codes. In this study, we evaluated differences in how frequently clinicians code for ARIs and factors that are associated with the use of ARI diagnostic codes in Emergency Department (ED) and Urgent Care (UC) visits across an integrated healthcare system. Methods We analyzed a retrospective cohort of all ED and UC patient-visits across 129 Veterans Affairs medical centers during 2016-2018. ARI visits were identified using ICD-10 codes for acute bronchitis, influenza, pharyngitis, sinusitis, and upper respiratory tract infections for clinicians with 100 or more visits. A generalized linear mixed model with a link logit function that accounted for clustering at the clinician and facility-level was used to calculate median odds ratios (OR) and to identify factors associated with increased likelihood of entering an ARI code. Results There were 6,016,499 patient-visits, and 519,389 (8.6%) were coded as an ARI (Table 1). The mean rate of ARI diagnoses across all visits was 8.9% (SD 2.5%) at the facility-level and 7.4% (SD 4.5%) at the clinician-level (Table 2). The median OR was 2.19 (95% CI 2.18, 2.22), suggesting there was between-clinician variation in coding for ARI diagnoses. Visits were significantly more likely to be coded as ARIs if seen by an advanced practice provider (OR=2.36, 95% CI 2.19, 2.54), if a fever was recorded (OR=4.20, 95% CI 4.18, 4.29), and if the visit occurred between December-March (OR=1.97, 95% CI 1.96, 1.98). Approximately 2/5th of the variability (41.4%) in assigning an ARI diagnostic code was explained by differences across individual clinicians. Conclusion There was substantial variability in how frequently ED and UC clinicians coded a visit as an ARI, and a large proportion of the variability was explained by differences across clinicians. Unmeasured factors could include different approaches to using diagnostic codes. ARI metrics based on diagnostic codes may need to account for differences in clinicians’ coding behavior. Disclosures All Authors: No reported disclosures

Background. Providing effective transitions-of-care (TOC) services improves outcomes for patients discharged on high-risk medications. Literature has shown that successful TOC for certain antimicrobials reduces hospital readmissions, medication errors, and improves post-discharge follow-up and laboratory monitoring. Prior to this quality improvement (QI) initiative, there was no formal TOC process for patients discharged on high-risk antimicrobial therapy (HAT) at our institution. Without standardization, only 55.1% of patients discharged on HAT had successful TOC. The aim of this initiative was to develop and implement a TOC protocol in at least 90% of patients discharged on HAT.
Methods. This QI initiative utilized the Institute of Healthcare Improvement model for improvement. A workgroup of key stakeholders developed a protocol to identify and standardize TOC services provided to patients discharged on HAT. Successful protocol completion was achieved if the following process metrics were evaluated, obtained, and documented prior to discharge: baseline laboratory values, pharmacokinetic monitoring, appropriate intravenous access, drug-drug interactions, medication availability, discharge medication counseling, and formal pharmacist documentation in a discharge note. Outcome metrics included referral to outpatient infectious disease (ID) follow-up, 90-day readmissions, and successful TOC. Balancing metrics included pharmacist time and protocol initiation for patients not discharged on HAT.
Results. Between October 2020 and May 2021, 218 patients met protocol inclusion criteria. Of these, 203/218 (93.1%) were appropriately identified with the new TOC process. The protocol was successfully followed in 78.9% of patients identified. Readmission rates were 42.8%, which was unchanged from baseline. Inpatient ID involvement increased from 80.9% to 95.7% and referral to outpatient ID follow-up from 59% to 94.8%.
Conclusion. This newly developed TOC protocol successfully identifies patients discharged on HAT, improves provision of TOC services to these high-risk patients, and significantly improves the rate of infectious disease involvement while inpatient and after discharge.
Disclosures. All Authors: No reported disclosures

Optimization of Inpatient Antibiotic Use via an Electronic Antimicrobial Stewardship Module and an Infectious Diseases Pharmacy Resident
Victor Chen, PharmD 1 ; Lauren Allen, PharmD 2 ; Hongkai Bao, Background. Antibiotic resistance is a public health crisis and antimicrobial stewardship (AMS) pharmacists serve a crucial role in preventing inappropriate use. At Montefiore Medical Center (1,500-bed hospital), a new electronic medical record AMS module was implemented with assistance from an infectious diseases (ID) pharmacy resident in October 2020. The module utilizes a dynamic scoring system to assist in prioritizing interventions, including bug-drug mismatches, insufficient coverage, or de-escalation. The AMS module is operationalized by ID pharmacists during the week and an ID pharmacy resident every other weekend. The objective of this study was to assess the impact of an ID pharmacy resident performing AMS module interventions on broad spectrum antibiotic use.
Methods. An observational study of AMS module interventions on antibiotic use (AU) in days of therapy per 1,000 days present and standardized antimicrobial administration ratio (SAAR) was performed. AU data for piperacillin-tazobactam (P/T) and SAAR prior to (October 2019-December 2019) and after (October 2020 -December 2020) the integration of an ID pharmacy resident and the AMS module was compared. Additional data collected included total number and type of interventions.
Results. A total of 539 interventions were made by AMS pharmacists and 36.5% of these were completed by the ID pharmacy resident. Across 6 different units, there was a statistically significant decrease in the SAAR for broad spectrum antibacterial agents (Figure 1), and a decrease of at least 10% in P/T use during the two different time periods (Table 1). An estimated P/T cost reduction of 26% of ($48,708 to $36,235.80) was observed. AMS pharmacists made 63 interventions in respective units. The top three intervention types were dose/frequency/duration recommendations, pharmacokinetic vancomycin dosing/monitoring, and de-escalation. The acceptance rate of interventions was 99% (534 accepted interventions/539 total interventions).