BACKGROUND:Health services in high-income countries increasingly recognise the challenge of effectively serving and engaging with marginalised people. Effective engagement with marginalised people is essential to reduce health disparities these populations face. One solution is by tapping into the phenomenon of boundary-spanning people in the community-those who facilitate the flow of ideas, information, activities and relationships across organisation and socio-cultural boundaries. METHODS:A scoping review methodology was applied to peer-reviewed articles to answer the question: "How do health services identify, recruit and use boundary spanners and what are the outcomes?" The review was conducted in seven databases with search terms based on community-based boundary spanning, marginalised people and health services. FINDINGS:We identified 422 articles with the screening process resulting in a final set of 30 articles. We identified five types of community-based boundary spanning: navigators, community health workers, lay workers, peer supporters and community entities. These range from strong alignment to the organisation through to those embedded in the community. We found success in four domains for the organisation, the boundary spanner, the marginalised individuals and the broader community. Quantifiable outcomes related to cost-savings, improved disease management and high levels of clinical care. Outcomes for marginalised individuals related to improved health knowledge and behaviours, improved health, social benefits, reduced barriers to accessing services and increased participation in services. We identified potential organisational barriers to using boundary spanners based on organisational culture and staff beliefs. CONCLUSIONS:Community boundary spanners are a valuable adjunct to the health workforce. They enable access to hard to reach populations with beneficial health outcomes. Maintaining the balance of organisational and community alignment is key to ongoing success and diffusion of this approach.