Using a systematic review to uncover theory and outcomes for a complex intervention in health and social care: a worked example using life story work for people with dementia.

Objectives: Life story work involves gathering information about a person, their history and interests, and producing a tangible output, usually a book. It is used increasingly in dementia care. However, theory about if, how, and why it affects outcomes is poorly developed, making the choice of evaluation methods and appropriate outcomes difficult. This article uses a systematic review to uncover theories of change and outcomes for life story work in dementia care. We describe the methods and their use in identifying appropriate outcomes for evaluative research, and their potential for improving evaluation of 'theory poor' interventions.

Methods: A systematic review, carried out using Centre for Reviews and Dissemination guidelines, searched for evidence on underlying theory, good practice, and effectiveness of life story work to inform a feasibility study for the formal evaluation of life story work in dementia care. For the theory element, a 'landscaping review' analysed the extracted text, using qualitative techniques and mind maps to uncover both explicit and implicit causal links between life story work and outcomes. We triangulated the review findings with qualitative research, which used focus groups with people with dementia, caregivers, and professionals that explored the outcomes that they would like to see from life story work.

Results: Of the 56 publications reviewed, only 16 were useful for the theory analysis. Six overarching outcomes were identified: the self-worth and empowerment of people with dementia, individual psychological outcomes, improved relationships between care staff and the individual with dementia, better care, more effective engagement of family members/carers within the care setting, and helping carers to cope. Twelve final theories linking these outcomes to life story work via a causal pathway were elicited. There was substantial overlap in the outcomes identified by the review and by carers, people with dementia, and professionals. Together, the results informed our choice of outcome measures for the pilot evaluation.

Conclusions: This approach may enable researchers to identify and develop the theory necessary before evaluation of a complex intervention in other under- or un-theorized areas. It has the potential both to shorten development stages (and thereby costs) in intervention research and improve the intervention itself.