Despite increasing use, limited data has been published comparing safety of different outpatient parenteral antimicrobial therapy (OPAT) models. Potential risks of self-administration at home include venous access device infection and other line complications. This study aims to investigate rates and predictors of intravenous access device complications in a large OPAT cohort. This is a retrospective cohort study of all uses of midlines, peripherally inserted central catheters (PICCs) and tunnelled central venous catheters (TCVCs) with univariate and multivariate (logistic regression) analysis of factors associated with line infections (LIs) and with other line events (OLEs). On univariate analysis, line infections were associated with length of line use, female sex and TCVC lines (compared to midlines). Patients self-administering OPAT in the home had a non-significantly lower rate of LIs. On multivariate analysis only duration of line use was a significant predictor of LIs-OR 1.012 (95%CI 1.001-1.023). For OLEs, multivariate analysis suggested that only line type and use of flucloxacillin were significant explanatory variables. In this cohort, there is no evidence that self-administration of OPAT is associated with higher rates of venous access device complications after controlling for confounding variables.