BackgroundElder abuse is a public health problem that is gaining attention due to its serious impacts on people’s health and well-being, and it is predicted to increase along with the world’s rapidly ageing population. Staff-to-resident abuse in nursing homes is a complex and multifaceted phenomenon associated with multiple factors on different levels of the ecological model. This study aimed to explore individual, relational, and institutional characteristics associated with perpetrated staff-to-resident abuse in nursing homes, using a multilevel hierarchical approach.MethodsThis was a cross-sectional exploratory study of 3693 nursing staff (response rate 60.1%) in 100 randomly selected nursing homes in Norway. We explored the characteristics of nursing staff, their relationship with residents, and institutional features associated with three types of abuse: psychological abuse, physical abuse, and neglect. These were modelled using multilevel mixed-effects logistic regression analyses.ResultsIndividual staff factors found to be associated with all three types of abuse were 1) being a registered nurse/social educator (OR 1.77–2.49) or licensed practical nurse (OR 1.64–1.92), 2) reporting symptoms of psychological distress (OR 1.44–1.46), 3) intention to leave the job (OR 1.35–1.40), and 4) reporting poor attitudes towards people with dementia (OR 1.02–1.15). Also, staff who reported poorer quality of childhood were more likely to perpetrate neglect (OR 1.14). Relational factors such as care-related conflicts (OR 1.97–2.33) and resident aggression (OR 1.36–2.09) were associated with all three types of abuse. Of institutional factors, lack of support from a manager was associated with perpetrating psychological abuse (OR 1.56).ConclusionsWe found several predictors of staff-to-resident abuse on different levels of the ecological model, which underlines the importance of using a multifaceted approach to identify risk factors of elder abuse in nursing homes. However, future studies should explore the underlying mechanism and causes with a prospective or qualitative design and target the multifaceted nature of risk factors when designing preventive interventions.