IntroductionHospitalizations due to severe injection-related infections (SIRIs) and patient-directed discharge (PDD) in people who inject drugs (PWID) are increasing, but research on readmission outcomes at PDD is limited. In this retrospective, matched cohort study we evaluated predictors of 30-day readmission by discharge status among PWID. MethodsAmong patients diagnosed with SIRIs at a tertiary hospital, Fisher's exact tests assessed differences in readmission rates by discharge status. Medications for opioid use disorder (MOUD) at discharge was defined as either having a buprenorphine dose dispensed within 24 h of discharge and buprenorphine being included in the discharge summary as a prescription, or a methadone dose dispensed inpatient within 24 h of discharge. Logistic regression analyses evaluated predictors of readmission outcomes. ResultsAmong 148 PWID with SIRI diagnosis, 30-day readmission rate following PDD was higher than standard discharge (25.7 % vs. 9.5 %, p = 0.016) and MOUD decreased odds of 30-day readmission (OR = 0.32, 95 % CI: 0.12,0.83, p = 0.012). >7 missed days of antibiotic treatment increased odds of 30-day readmission (OR 4.65, 95 % CI: 1.14, 31.72, p = 0.030) within PDD patients. ConclusionsPDD carries higher 30-day readmission rate compared to standard discharge. Strategies to reduce PDD rates and increase MOUD initiation may improve readmission outcomes.