We used longitudinal network analysis to assess the effect of participating in the SCOPE study on networks of care aide-led QI teams located in two Canadian provinces. While increasing connectivity was not the focus of SCOPE, we found that the overall network connectivity within and across regions increased over time in both provinces, which included connections to new teams as well as strengthening existing relations (from knowing each other to communication and collaboration). The gradual increase of connectivity over time suggests that SCOPE was successful in building relationships between teams across nursing homes. This could be attributed to the participatory nature of the SCOPE intervention that provided various opportunities to collaborate, including Learning Congresses where teams were encouraged to interact and share knowledge and experiences, and regular progress meetings with a Quality Improvement advisor who spanned across teams. SCOPE was also subject to a rigorous process evaluation where teams gathered to discuss their experiences of participation and progress with their QI intervention.(36) Early evidence from the SCOPE intervention showed that its team-based, collaborative structure facilitated the perception of communication and support among the members of QI teams within nursing homes, as reflected in cohesion scores by team members.(37, 38)
Improvement in connectivity across teams could facilitate formation of ‘communities of practice’, in which members collaborate to address a common problem.(39) In other care settings, information sharing and peer support in communities of practice can help healthcare providers break down professional, geographical and organizational barriers, learn from each other about common challenges and solutions to address them, develop collective expertise, and gain a sense of identity and belonging, (16) which results in value creation and efficiency. (40) Few studies have assessed the development of communities of practice in nursing homes, but some limited evidence supports their effectiveness in facilitating the development of a culture of collaboration and support among nursing home staff, building capacity for evidence-based practice (41) and sustainable quality improvement.(42) Studying the dynamics of network formation in the context of inter-organizational interventions can provide insights onto the social processes that may affect the formation and sustainment of communities of practice.
In our analysis, we found that geographic proximity and co-membership in organizational chains were important predictors of connectivity before, and during, SCOPE. In both provinces, teams were more likely to connect to others in the same region and those who belonged to the same organizational chains. This may explain higher network density before SCOPE in AB (compared to BC), since 10 out of 14 AB teams shared chain memberships and most teams in AB- Central & Calgary were geographically close to each other, as opposed to BC, in which 8 out of 17 shared chains, and many BC teams, particularly in the Fraser Health region, were geographically dispersed, independent, family-run nursing homes. Proximity is an important factor in the creation of ties between teams. (43, 44, 45) Geographic proximity facilitates knowledge diffusion and information exchange through serendipitous communications between neighboring actors.(46) Organizational proximity, on the other hand, occurs when network actors share institutional membership, such as belonging to chains. Our findings suggest that nursing homes that were geographically closer were more likely to build new relations. Co-affiliation to chains (which potentially involved shared standards of care, leadership, staffing, and prior recognition of each other) was significantly associated with relationships at each timepoint (but not a significant predictor of building new ties), indicating stable connectivity among nursing homes belonging to same chains throughout the study.
In BC, many teams did not have connections before SCOPE (as reflected in the smaller density) but developed a dense and highly reciprocated network over time, which was reflected in the increased knowledge of other teams and development of communication and collaboration ties. The larger increase in reciprocity in BC compared to AB implies that independent teams probably benefited more from the network-building nature of SCOPE and used this opportunity to build connections with new partners. In other contexts, healthcare professionals, particularly in remote and rural areas, have expressed needs to belong to communities of support and information, to help addressing isolation, and finding others with similar issues and experiences.(47)
A common pattern in both provinces was an increasing trend of out-degree centralization. Based on the SAOM model results, teams in AB that were already active became even more active after the intervention. This implies that the SCOPE intervention provided opportunities that were disproportionately utilized by teams who were already outgoing and active; thus, less-outgoing teams benefited less from the networking aspects of the intervention. This suggests that, if identified and engaged, active network members might be intentionally engaged to transmit new ideas, and help developing trust and sustainable relationships, and may form the next generation of opinion leaders in the network.(48)
This study has several limitations, including the lack of parallel control group, and inability to assess the network building impact of various SCOPE components. In addition, we did not assess the relationship between connectivity and network positions with QI outcomes and satisfaction with the intervention. Larger scale studies can assess and optimize the impact of different strategies included in complex team-based interventions using advanced experimental designs, such as factorial designs, and using strategies for intervention optimization, such as Multiphase optimization strategy (MOST).(49) The findings should be interpreted in light of the nature and dynamics collaboration networks among nursing homes in study provinces, and might have limited generalizability to other contexts.