How social networks influence the local implementation of initiatives developed in quality improvement collaboratives in health care: A qualitative process study.

Background: Quality improvement collaboratives (QICs) have facilitated cross-organizational knowledge exchange in health care. However, the local implementation of many quality improvement (QI) initiatives continues to fail, signaling a need to better understand the contributing factors. Organizational context, particularly the role of social networks in facilitating or hindering implementation within organizations, remains a potentially critical yet underexplored area to addressing this gap.

Purpose: We took a dynamic process perspective to understand how QI project managers' social networks influence the local implementation of QI initiatives developed through QICs.

Methodology: We explored the case of a QIC by triangulating data from an online survey, semistructured interviews, and archival documents from 10 organizations. We divided implementation into four stages and employed qualitative text analysis to examine the relationship between three characteristics of network structure (degree centrality, network density, and betweenness centrality) and the progress of each QI initiative.

Results: The progress of QI initiatives varied considerably among organizations. The transition between stages was influenced by all three network characteristics to varying degrees, depending on the stage. Project managers whose QI initiatives progressed to advanced stages of implementation had formed ad hoc clusters of colleagues passionate about the initiatives.

Conclusion: Implementing QI initiatives appears to be facilitated by the formation of clusters of supportive individuals within organizations; this formation requires high betweenness centrality and high network density.

Practice implications: Flexibly modifying specific network characteristics depending on the stage of implementation may help project managers advance their QI initiatives, achieving more uniform results from QICs.