Aim: To evaluate the effects of implementing a geographic cohorting unit for cognitively impaired inpatients with challenging behaviours on staff’s person-centred care processes and patient-to-staff violence.
Methods: A 10-bed secure geographic cohorting unit for frail older cognitively impaired patients was implemented on a medical ward in a large metropolitan hospital. Implementation involved dementia-enabling ward modifications alongside an enhanced multidisciplinary workforce with increased nurse ratios and specialised training. A mixed-methods controlled interrupted time series design was used, with monthly occupational violence incident rates collected from July 2023 to October 2025 across one intervention and three comparison medical wards. Segmented negative binomial regression estimated changes in level and slope, accounting for overdispersion and concurrent trends. Twelve months post-implementation, semi-structured interviews were conducted with a convenience sample of intervention-ward staff.
Results: Following implementation, there was a significant monthly reduction in occupational violence incident counts by 18% per month (IRR 0.82, 95% CI 0.77–0.88; p<0.001), after an initial transient increase. In the controlled model including all wards, monthly incidents decreased by 11% per month, post-intervention (IRR 0.89, 95% CI 0.85–0.92; p<0.001). Staff reported improvements in clinical responsiveness, pain management, behavioural support, team reflexivity, reduced restraint use, enhanced morale, and increased carer confidence.
Conclusion: Implementing a secure, geographic cohorting unit for cognitively impaired inpatients with challenging responsive behaviours was associated with reduced patient-to-staff violence and improvements in person-centred care processes. This demonstrates the value of infrastructure-enabled models of care for older frail cognitively impaired patients in acute-care settings.