Ethnographic process evaluation of a quality improvement project to improve transitions of care for older people

Objectives: Quality improvement projects to address transitions of care across care boundaries are increasingly common but meet with mixed success for reasons that are poorly understood. We aimed to characterise challenges in a project to improve transitions for older people between hospital and care homes. Design: Independent process evaluation, using ethnographic observations and interviews, of a quality improvement project. Setting and participants: An English hospital and two residential care homes for older people. Data: 32 hours of non-participant observations and 12 semistructured interviews with project members, hospital and care home staff. Results: A hospital-based improvement team sought to reduce unplanned readmissions from residential care homes using interventions including a community-based geriatric team that could be accessed directly by care homes and a communication tool intended to facilitate transfer of information between homes and hospital. Only very modest (if any) impacts of these interventions on readmission rates could be detected. The process evaluation identified multiple challenges in implementing interventions and securing improvement. Many of these arose because of lack of consensus on the nature of the problem and the proper solutions: while the hospital team was keen to reduce readmissions and saw the problems as lying in poor communication and lack of community-based support for care homes, the care home staff had different priorities. Care home staff were unconvinced that the improvement interventions were aligned with their needs or addressed their concerns, resulting in compromised implementation. Conclusions: Process evaluations have a valuable role in quality improvement. Our study suggests that a key task for quality improvement projects aimed at transitions of care is that of developing a shared view of the problem to be addressed. A more participatory approach could help to surface assumptions, interpretations and interests and could facilitate the coproduction of solutions. This finding is likely to have broader applicability.