A mixed-methods multi-site case study of a person-centred intervention for constant observation in hospitals with people living with dementia

Introduction Constant observation is widely used with people living with dementia admitted to hospital when identified at risk of harm to themselves or others. Staff allocated to closely monitor individual or small groups of patients intervene when there are safety concerns and may engage with patients’ psychosocial needs. However, care is inconsistent and dependent upon individual and organisational factors. This study aimed to understand the work of implementing a co-designed intervention for facilitating person-centred approaches during constant observation practices. Methods A convergent parallel mixed-methods multi-site case study was adopted to explore implementation over 12 weeks in three English hospitals. The study recruited participants from with six wards and one hospital-wide team. Qualitative and quantitative data involved: i) observations of staff-patient interactions (four time points), ii) in depth interviews with hospital staff (one time point) iii) staff surveys (two time points). Qualitative data analysis was organised using Normalisation Process Theory to map and understand the implementation process. NOrmalization MeAsure Development (NoMAD) survey data were analysed using descriptive statistics. Findings We recruited 163 participants - staff (n=88), people living with dementia (n=71), family supporters (n=4). The intervention was well received and considered useful by staff. Incremental changes, such as staff initiating non-task related conversations with patients and using tools to inform actions for reducing distress, were observed. However, establishing the importance of psychosocial, alongside physical and medical, needs was not achieved. Staff found it difficult to challenge the dominance of medical management and organisations’ priorities to minimise risk. Fears that discussions about constant observation with family supporters might upset them or result in accusations of inadequate care inhibited work to collect and share potentially useful information. Conclusion The intervention endorsed and supported staff to focus on the quality of their care work; this was not usual practice. Routine use was impacted by prior knowledge of dementia, how the intervention aligned with ward practice and competing priorities. Additional work is required to support the shift from work organised as a reaction to urgent, risky situations to work that supports prevention and enhances care.