134. Impact of an Antibiotic Stewardship Treatment and Management Algorithm for Liver Abscesses

Abstract Background Antibiotic prescribing for pyogenic liver abscess(es) (PLA) is highly variable with literature primarily aimed at assessing surgical intervention with a scarcity of data for antibiotic selection and duration of therapy. Given the lack of data, there is no clear consensus for treatment options or length of treatment. Our Antimicrobial Support Network (ASN) in collaboration with the hepatopancreatobiliary (HPB) team created a treatment and management algorithm to guide duration of therapy and antibiotic selection. Methods A retrospective, quasi-experimental cohort study was performed at Carolinas Medical Center in hospitalized patients with PLA with an HPB and/or infectious diseases consult. The primary outcome was antipseudomonal beta-lactam days of therapy (DOT) per 1000 patient days (PD) in the pre-versus post-intervention group. Secondary outcomes included rates of treatment failure at 90 days, 90-day all-cause and abscess-related hospital readmission, C. difficile and multi-drug resistant organism (MDRO) colonization at 90 days from diagnosis, and hospital length of stay (LOS). Additional a priori subgroup analyses of duration of therapy, treatment failure, all-cause and abscess-related readmissions were also conducted based on surgical intervention. Results A total of 93 patients were included, 49 patients in the pre-intervention group and 44 patients in the post-intervention group. Baseline characteristics were similar between the groups. The majority of liver abscesses were unilocular and monomicrobial. Anti-pseudomonal beta-lactam DOT per 1000 PD decreased by 13.8% (507.4 versus 437.5 DOT/1000 PD). Treatment failure occurred in 30.6% of pre-intervention patients and 18.2% of post-intervention patients (p = 0.165). Patients in the post-intervention group were discharged a median of 2.4 days sooner than the pre-intervention period (12.2 days vs. 9.8 days, p = 0.159). No significant differences resulted in 90-day readmission rates or 90-day C. difficile or MDRO rates. Table 1. Primary Outcome for Patients with Pyogenic Liver Abscesses Treated Pre- and Post-Antibiotic Stewardship Algorithm Table 2. Secondary Outcomes for Patients with Pyogenic Liver Abscesses Treated Pre- and Post-Antibiotic Stewardship Algorithm Conclusion The implementation of a PLA treatment and management algorithm led to a decrease in anti-pseudomonal beta-lactams without impacting clinical outcomes and a trend towards decreased LOS. Disclosures All Authors: No reported disclosures

for aminoglycosides and all neonatal antibacterial agents. SAARs were compared using the NHSN Statistics Calculator.
Results. For third generation cephalosporins, there were 385 observed antimicrobial days (OAD) and 115 expected antimicrobial days (EAD) in the pre-implementation period compared to 597 OAD and 228 EAD in the post implementation period. This resulted in a SAAR of 3.34 and 2.62, respectively; a reduction of 22% (p < 0.001). For aminoglycosides, there were 713 OAD and 584 EAD compared to 1617 OAD and 1155 EAD. This resulted in a SAAR of 1.22 and 1.4; an increase of 15% (p = 0.002). For all neonatal antibacterial agents, there were 2716 OAD and 1739 EAD compared to 5321 OAD and 3438 EAD. This resulted in a SAAR of 1.56 and 1.55; indicating no change in use (p = 0.70). See Table 1 for results. Background. Management of a hospital's antimicrobial formulary is an important aspect of antimicrobial stewardship and cost containment strategies. Ensuring that essential medications for clinical care are available and excluding therapeutic duplicates and unnecessary antimicrobials is time and resource intensive. Comparisons of antimicrobial formularies across multiple rural hospitals have not been evaluated in the literature. We hypothesized that a comprehensive formulary evaluation would reveal important opportunities for antimicrobial stewardship efforts and could help smaller hospitals optimize available medications.
Methods. The University of Washington Tele-Antimicrobial Stewardship Program (UW-TASP) is comprised of 68 hospitals of varying sizes, most of which are rural and critical access, in Washington, Oregon, Arizona, Idaho, and Utah. We surveyed UW-TASP participating hospitals and other networked rural hospitals in multiple Western states using REDCap, a HIPAA-compliant, electronic data management program. Respondents reported which antimicrobials are on their hospital formulary as well as basic information about hospital size and inpatient units. Data were reviewed by a panel of infectious diseases trained physicians and pharmacists at UW-TASP.
Results. Surveys from 49 hospitals were received; two were excluded from the data analysis (Table 1) -one submission was incomplete, and one was a large inpatient psychiatric hospital. Select antimicrobials and proportion of hospitals carrying these agents is shown in Table 2. Several antimicrobials are on the formulary at all hospitals, regardless of size. In some critical access hospitals (< 25 beds), empiric first-line bacterial meningitis and viral encephalitis coverage (Table 3) was lacking. Six hospitals (12.7%) lacked ampicillin for Listeria coverage and only one had a suitable alternative agent (meropenem). Seven hospitals (14.9%) lacked intravenous acyclovir, although three had oral valacyclovir. Formulary inclusion of agents for multi-drug resistant organisms was rare.

Conclusion.
In critical access hospitals in the Western USA, lack of essential empiric antimicrobials may be more of a concern than inclusion of agents with unnecessarily broad spectra.
Disclosures. Chloe Bryson-Cahn, MD, Alaska Airlines (Other Financial or Material Support, Co-Medical Director, position is through the University of Washington) Background. Antibiotic prescribing for pyogenic liver abscess(es) (PLA) is highly variable with literature primarily aimed at assessing surgical intervention with a scarcity of data for antibiotic selection and duration of therapy. Given the lack of data, there is no clear consensus for treatment options or length of treatment. Our Antimicrobial Support Network (ASN) in collaboration with the hepatopancreatobiliary (HPB) team created a treatment and management algorithm to guide duration of therapy and antibiotic selection. S180 • OFID 2021:8 (Suppl 1) • Abstracts Methods. A retrospective, quasi-experimental cohort study was performed at Carolinas Medical Center in hospitalized patients with PLA with an HPB and/or infectious diseases consult. The primary outcome was antipseudomonal beta-lactam days of therapy (DOT) per 1000 patient days (PD) in the pre-versus post-intervention group. Secondary outcomes included rates of treatment failure at 90 days, 90-day all-cause and abscess-related hospital readmission, C. difficile and multi-drug resistant organism (MDRO) colonization at 90 days from diagnosis, and hospital length of stay (LOS). Additional a priori subgroup analyses of duration of therapy, treatment failure, all-cause and abscess-related readmissions were also conducted based on surgical intervention.

Results.
A total of 93 patients were included, 49 patients in the pre-intervention group and 44 patients in the post-intervention group. Baseline characteristics were similar between the groups. The majority of liver abscesses were unilocular and monomicrobial. Anti-pseudomonal beta-lactam DOT per 1000 PD decreased by 13.8% (507.4 versus 437.5 DOT/1000 PD). Treatment failure occurred in 30.6% of pre-intervention patients and 18.2% of post-intervention patients (p = 0.165). Patients in the post-intervention group were discharged a median of 2.4 days sooner than the pre-intervention period (12.2 days vs. 9.8 days, p = 0.159). No significant differences resulted in 90-day readmission rates or 90-day C. difficile or MDRO rates. Background. In nursing homes, federal mandates call for more judicious use of antibiotics and antipsychotics. Previous research indicates that practice patterns of nursing home practitioners, rather than resident's signs and symptoms or overall medical conditions, drive antibiotic use. We hypothesized that nursing home practitioners who prescribe antibiotics more frequently than their peers may display a similar practice pattern for antipsychotics. Here, we examine similarities in prescribing patterns for antibiotics and antipsychotics among practitioners at 29 U.S. nursing homes.
Methods. Prescription data came from 2016 invoices from a pharmacy common to all 29 nursing homes. We defined practitioners as individuals who prescribed ≥1% of systemic medications at a nursing home and excluded practitioners without no prescriptions for anti-hypertensive drugs assuming they were not treating a general nursing home population (i.e. treating hospice or dementia patients). Using anti-hypertensive starts for standardization, we calculated the expected number of starts for both antibiotics and antipsychotics. Using funnel plots with Poisson 99% control limits for the observed-to-expected ratio, we identified practitioners whose use of either class of drugs exceeded these control limits. Practitioners were classified as high, average, or low prescribers for each class of drugs.