A realist evaluation approach to explaining the role of context in the impact of a complex eHealth intervention for improving prevention of cardiovascular disease.

BACKGROUND: Reduction of cardiovascular disease (CVD) is a worldwide health priority and innovative uses of technology-based interventions may assist patients with improving prevention behaviours. Targeting these interventions to recipients most likely to benefit requires understanding how contexts of use influence responsiveness to the intervention, and how this interaction favours or discourages health behaviour. Using a realist evaluation approach, the aim of this study was to examine the contextual factors influencing behaviour change within a multi-feature eHealth intervention with personalised data integration from the primary care electronic health record (EHR). METHODS: Realist evaluation of qualitative data from the Consumer Navigation of Electronic Cardiovascular Tools (CONNECT) randomised trial (N = 934). Thirty-six participants from the intervention group (N = 486) who had completed 12 months of study follow-up were interviewed. Coding of transcripts was structured around configurations of contexts, mechanisms, and outcomes of intervention use. Contextual narratives were derived from thematic analysis of the interviews. RESULTS: Mechanisms favouring positive health behaviour occurred when participants responded to four interactive features of the intervention. Facilitating mechanisms included greater cognitive engagement whereby participants perceived value and benefit, and felt motivated, confident and incentivised. Participants moved from being unconcerned (or unaware) to more task-oriented engagement with personal CVD risk profile and prevention. Increased personalisation occurred when modifiable CVD risk factors became relatable to lifestyle behaviour; and experiences of feeling greater agency/self-efficacy emerged. Use and non-use of the intervention were influenced by four overarching narratives within the individual's micro-level and meso-level environments: illness experiences; receptiveness to risk and prevention information; history of the doctor-patient relationship; and relationship with technology. CONCLUSIONS: Intervention-context interactions are central to understanding how change mechanisms activate within complex interventions to exert their impact on recipients. Intervention use and non-use were context-dependent, underscoring the need for further research to target eHealth innovations to those most likely to benefit.