To explore hospital staff experiences and perceptions of patient-perpetrated violence. Descriptive qualitative study. Twelve semi-structured interviews (June-August 2022) were held with a diverse sample of hospital nurses, doctors, allied health professionals, security and a non-clinical manager. The framework approach was used to organise and analyse data, using Attribution Theory as a theoretical lens. Three themes were identified: violence as (un)predictable, violence as (un)preventable and the cumulative toll of violence. In making sense of why patients become violent, participants described different 'types' of aggressive patients and variably attributed behaviours to situation, disposition or a combination of both. Regardless of perceived causal factors, staff overwhelmingly appeared to view violence as predictable. Participants also reflected on the wider structural problems underpinning violence, frequently alluding to their sense of relative powerlessness to initiate change. The cumulative toll of violence was a common thread, with staff describing their acquisition of 'resilience' and reflecting on its role in their responses to escalating situations. Many hospital staff are resigned to the inevitability of violence. The concept of staff 'resilience' following violence is not unproblematic, having the potential to serve as a guise for acceptance and as an additional variable for which staff are held accountable. When designing strategies, organisations should ensure that accountability for violence reduction is distributed across multiple levels. This study makes a novel contribution by exploring the perspectives of multiple staff groups working across diverse hospital settings, and adds to a sparse literature on this subject in the UK. Efforts to address violence against healthcare staff need to be power-conscious, ensuring that accountability is distributed across multiple levels. This study is reported in line with the Consolidated Criteria for Reporting Qualitative Studies (COREQ). No patient or public contribution.