A realist evaluation of the roles of opinion leaders in the diffusion of innovations in primary care in Florianopolis, Brazil

Local opinion leaders are people with credibility and influence within social groups. Therefore, they are often used as part of strategies to promote adoption of innovations in healthcare settings. Evidence from trials show that opinion leaders are an effective strategy to promote behaviour change, but the outcomes of their use are variable and unpredictable across studies. There is a need for better understanding of how and why opinion leaders work or not in different circumstances to improve the design of behaviour change interventions. This study addressed this knowledge gap by analysing the roles of opinion leaders during the implementation of two innovations in a primary care system of Brazil. Using a realist evaluation, I developed, tested and refined programme theories about the roles of the opinion leaders. First, I developed initial theories from documents, literature review, stakeholders’ consultation and my experience in the setting. Second, I tested and refined those initial theories drawing on 18 interviews with managers and practitioners and a reassessment of the literature. The three programme theories focused on how recognising opinion leaders motivates buy-in to innovations; how involving opinion leaders in implementation gives credibility to innovations; and how the practice of opinion leaders with innovations promotes adoption. The analytical framework was based on the programme theories and the Context-Mechanism-Outcome configuration. The causal processes identified in data analysis were compared to the initial theories to generate refined programme theories. The key findings across refined theories were summarised in a middle-range theory. The findings suggested causal processes that might explain some of the variability in opinion leaders’ interventions. Key mechanisms included ownership of innovations, trust, and reinforcement of group norms and modelling. Key contextual factors included interest in the innovations, similarity between opinion leader and peers, and informal relationships. The initial mobilisation of opinion leaders is a separate component of the intervention, leading to contradictory outcomes across system levels. Ultimately, there is a trade-off between harnessing the influence of opinion leaders as a resource for implementation and jeopardising their credibility.