Background. In healthcare, bridging the research-to-practice gap is a top priority. In 2008 the National Institute for Health Research (NIHR) funded nine Collaborations for Leadership in Applied Health Research and Care (CLAHRC); NHS-university partnerships seeking to accelerate the uptake of research into practice, a process referred to in this thesis as implementation. Evidence suggests that implementation might occur more readily when there is collaboration across various stakeholder and organisational boundaries. Boundary objects are shared things and ideas that are thought to enable communication across boundaries and create an opportunity for stakeholders to work together productively. Despite being studied across a range of settings in which collaboration is key, the role and potential of boundary objects remains understudied in relation to implementation. This thesis fills this gap. Methods. A case study of three CLAHRCs was conducted to explore the role of boundary objects in implementation. Phase 1, a document analysis, identified potential boundary objects (i.e. on paper) across the three cases. In Phase 2, in-depth interviews with people employed in boundary spanning roles in 3 CLAHRCs were conducted to investigate whether and how things and ideas were developed and used as boundary objects during implementation. Findings. Despite high numbers of potential boundary objects identified on paper through the document analysis (defined in this study as boundary objects-in-theory), including care pathways, assessment tools, and disease registers, in practice participants reported that some of these operated to reinforce boundaries. The study showed that there were things and ideas that were shared between stakeholders and enable them to collaborate to varying degrees (defined as boundary objects-in-use), including shared ideas around implementation, clinical topics, and some tools and guidelines. However some of these were perceived as prescriptive and imposed, requiring extensive adaptation to become meaningful to stakeholders. A process of creation and/or adaptation sometimes came about through unanticipated rather than planned processes. The most effective boundary objects-in-use were those which were co-produced in partnership with stakeholders. These were generated through discussions during which boundaries were clarified and solutions were sought to meet stakeholders’ needs, a process of collective endeavour identified as a type of bricolage. Boundary objects-in-use developed through bricolage possessed properties which were found to be lacking from those things that failed to make the transition from boundary objects-in-theory. Successful boundary objects-in-use were symbolically meaningful, resonant, and perceived as authentic by stakeholders. Discussion. An understanding of boundary objects defined by action-based properties rather than structural features is proposed, updating the classic typology. The study showed that for boundary objects-in-theory to make the transition to boundary objects-in-use, all relevant stakeholders must be engaged throughout the development process. Individuals working in implementation, such as boundary spanners, were more likely to deploy boundary objects effectively by using the skills of the bricoleur, initiating the collective creation and use of such objects. Implications. The findings from this study suggest that accepting and encouraging adaptation of those things that could in theory be boundary objects through a process of collective bricolage, instigated by credible boundary spanners, encourages the co-production of useful boundary-objects-in-use. These can represent an effective mechanism to enhance the appeal and relevance of outputs of research by providing a catalyst to align, engage, and accommodate multiple stakeholder perspectives.