Background: Care homes are the institutional providers of long-term care for older people. The OPTIMAL study argued that it is probable that there are key activities within different models of health-care provision that are important for residents’ health care.

Objectives: To understand ‘what works, for whom, why and in what circumstances?’. Study questions focused on how different mechanisms within the various models of service delivery act as the ‘active ingredients’ associated with positive health-related outcomes for care home residents.

Methods: Using realist methods we focused on five outcomes: (1) medication use and review; (2) use of out-of-hours services; (3) hospital admissions, including emergency department attendances and length of hospital stay; (4) resource use; and (5) user satisfaction. Phase 1: interviewed stakeholders and reviewed the evidence to develop an explanatory theory of what supported good health-care provision for further testing in phase 2. Phase 2 developed a minimum data set of resident characteristics and tracked their care for 12 months. We also interviewed residents, family and staff receiving and providing health care to residents. The 12 study care homes were located on the south coast, the Midlands and the east of England. Health-care provision to care homes was distinctive in each site.

Findings: Phase 1 found that health-care provision to care homes is reactive and inequitable. The realist review argued that incentives or sanctions, agreed protocols, clinical expertise and structured approaches to assessment and care planning could support improved health-related outcomes; however, to achieve change NHS professionals and care home staff needed to work together from the outset to identify, co-design and implement agreed approaches to health care. Phase 2 tested this further and found that, although there were few differences between the sites in residents’ use of resources, the differences in service integration between the NHS and care homes did reflect how these institutions approached activities that supported relational working. Key to this was how much time NHS staff and care home staff had had to learn how to work together and if the work was seen as legitimate, requiring ongoing investment by commissioners and engagement from practitioners. Residents appreciated the general practitioner (GP) input and, when supported by other care home-specific NHS services, GPs reported that it was sustainable and valued work. Access to dementia expertise, ongoing training and support was essential to ensure that both NHS and care home staff were equipped to provide appropriate care.

Limitations: Findings were constrained by the numbers of residents recruited and retained in phase 2 for the 12 months of data collection.

Conclusions: NHS services work well with care homes when payments and role specification endorse the importance of this work at an institutional level as well as with individual residents. GP involvement is important but needs additional support from other services to be sustainable. A focus on strategies that promote co-design-based approaches between the NHS and care homes has the potential to improve residents’ access to and experience of health care.

Funding: The National Institute for Health Research Health Services and Delivery Research programme.