Change in Provider Specialty Was Associated With Less Fluoroquinolone Use at a Veterans Affairs Long-Term Care Setting.

OBJECTIVE: In July 2021, as part of a planned multiyear broad and long-term organizational realignment, the general medicine service assumed continuous care of residents at a Community Living Center (CLC), which are nursing homes within the Veterans Affairs (VA) health care system. We hypothesized that practitioners accustomed to caring for patients in acute care would be more likely to prescribe antibiotics to long-term care residents.

DESIGN: Retrospective cohort study.

SETTINGS AND PARTICIPANTS: Residents of a 105-bed CLC associated with a large VA medical center.

METHODS: Our cohort included CLC residents between July 1, 2020, and June 30, 2022. We used administrative data to assess resident demographics and medical conditions in the 1 year before and after the change of practitioners. We also compared antibiotics agents prescribed and the following antibiotic use metrics in the year before and after the change: days of therapy (DOT) per 1000 bed days of care (BDOC), antibiotic starts/1000 BDOC, and mean length of therapy in days.

RESULTS: Resident characteristics and overall antibiotic use metrics were similar before and after the change in staffing. The specific agents prescribed differed, with a decrease in fluoroquinolones (14.3 to 5.8 DOT/1000 BDOC; P < .01) and an increase doxycycline (7.4 vs 19.1 DOT/1000 BDOC; P < .01) after the staff change. Rates of Clostridioides difficile infection also decreased, from 6.23 to 3.41 cases/10,000 BDOC after the change in staffing.

CONCLUSIONS AND IMPLICATIONS: The comparable antibiotic use metrics before and after the general medical service assumed care of the CLC residents may be explained by constancy in resident population and other facility-related factors. Differences in the types of agents used suggests that antibiotic stewardship efforts can be tailored not only to the setting and patient population but also to the practitioners' discipline.