Abstract
OBJECTIVE: Postdischarge antibiotics are often sub-optimal or unnecessary. This study sought to measure the risk of diarrhea in recently hospitalized patients treated with postdischarge antibiotics. DESIGN: Retrospective cohort study. SETTING: 125 acute-care hospitals in the Veterans Health Administration (VHA). PATIENTS: Patients hospitalized within VHA during 2018-2021. METHODS: The primary exposure was postdischarge antibiotics. The primary outcome was time to C. difficile testing, which served as a surrogate marker for clinically significant diarrhea. Only tests that were performed during the 30 days after discharge and before all-cause hospital readmission were captured. We constructed a final Cox proportional hazards model with 27 fixed-effect predictors as well as a random intercept for each hospital. RESULTS: There were 1,686,819 qualifying admissions, and 333,310 (19.8%) received postdischarge antibiotics. There were 13,387 patients (0.8%) who had a test for C. difficile done. Among those tested, the median time to testing was 6.7 days for those tested while on postdischarge antibiotics and 14.1 days for those tested while not on postdischarge antibiotics. Compared to patients not on postdischarge antibiotics, the hazard ratio for testing was 1.40 (95% CI, 1.29-1.51) among patients on low-risk postdischarge antibiotics and 1.56 (95% CI, 1.42-1.71) among those on high-risk postdischarge antibiotics. CONCLUSIONS: In this national VHA hospital cohort, patients prescribed postdischarge antibiotics had a 40-56% increased risk of C. difficile testing compared to those not prescribed postdischarge antibiotics. Efforts to optimize antibiotic-prescribing at hospital discharge, particularly by reducing excessive duration and avoiding high-risk agents, may help mitigate these risks.