Background Older people living with frailty (older people) often receive suboptimal end-of-life care (Lloyd et al., 2016). Frequently under- or over-treated, older people experience inappropriate hospital admissions and procedures (Houben et al., 2014; Hunt et al., 2014). Most die in hospital despite 81% wishing to die at home (Office for National Statistics 2016). Advance care planning (ACP), a structured conversation between professionals and someone nearing the end-of-life, enables people to discuss and document what matters to them regarding their future care (Thomas & Lobo, 2011) Facilitating informed decision-making regarding ACP means people are more likely to receive person-centred end-of-life care (National Council for Palliative Care, 2011). Multiple challenges mean that ACP is relatively uncommon for older people (Musa et al., 2015; Pollock & Wilson, 2015). Consequently, priorities are often not discussed prior to significant deterioration (Sharp et al., 2013). leading to crisis decision-making which older people may not have capacity for (Clegg, 2013). Current initiatives encompass aspects of ACP. However, there is no systematic, agreed format or training for health and social care professionals (professionals) that supports older people to articulate their broader end-of-life wishes on an ongoing basis (Brinkman-Stoppelenburg et al., 2014; Johnson et al., 2015; Weathers et al., 2016).

Aims This four-phase study aims to develop a relevant ACP intervention in collaboration with older people, their carers and professionals. Phase one, presented here, aims to establish current evidence regarding the components and implementation of ACP for community-dwelling older people, and use this to develop a theoretical model which will underpin the remaining study phases.

Methods An integrative review will be conducted using CINAHL, Embase, Ovid Medline, PsycINFO, and NHS Evidence databases, grey literature, and hand-searching. Paper selection will be verified by the research team and critically appraised using relevant systematic tools. Whittemore and Knafl’s (2005) method will be used to develop an integrated, theoretical model. This will include data mapping against COM-B, a model of behaviour change often used in intervention development, to identify relevant target behaviours. (Michie et al., 2011)

Results and Conclusions The study began in January 2017. Phase one is currently underway and its results, conclusions and recommendations will be presented at this conference.