Examining and predicting network relationships in a longitudinal study of a community health partnership

Work in public health increasingly requires coordination of multiple organizations within a community to achieve positive change.  Successful collaboration requires the connection of diverse organizations to tackle complex and pressing problems at multiple levels of the social ecological model (McLeroy et al., 1988).  Diversity though can create challenges for the coordination of public health collaborations seeking to make an impact.  Our presentation will use the experience from the development of a community health partnership in central Texas to assess study hypotheses related to the barriers to coordination created by diversity such as differences in sector (government, nonprofit, private), policy domain (public health, medicine, etc.), organization size, and tenure of membership within the collective. The data involve a comprehensive mixed-mode survey of all members to collect both organization-level attributes, connectivity within the network, and network position. Using the most recent of four waves of network surveys, we develop and test an exponential random graph model (ERGM) to test these hypotheses with the focusing being the existence (and persistence) of a tie between organizations.  The ERGM model provides the opportunity to simultaneously model the effects of organizational attributes (e.g. organization size), relations between organizations (e.g. shared sector), and structural aspects of the network (e.g. overall network density or a tendency towards closed triads of actors).  This research will provide insight into the points of vulnerability within public health networks.  Such knowledge will empower collaborative leaders to identify and address points of tension within their network, prevent potential attrition, and maximize collective impact.