Collaboration between Agencies Promoting Physical Activity in Brazil: A Network Analysis

Purpose: In recent decades, there has been a dramatic increase in lifestyle-related illnesses such as those caused by physical inactivity and unhealthy diets in Brazil (Rivera et al., 2004; Malta, 2009). Although various networks exist for promoting physical activity, there is sparse knowledge on how agencies are working together to promote evidence-based interventions. In this study, we focus on a network of 35 agencies that are participating in physical activity promotion in Brazil including a subnetwork formed by the institutions currently participating in project GUIA (Guide for Useful Interventions for Activity in Brazil and Latin America). The goal of GUIA is to enhance the capacity of agencies to identify and implement evidence-based interventions that promote physical activity. In this study, we investigate 1) how organizational characteristics predict collaboration between agencies, 2) the role of project GUIA-related agencies in facilitating evidence-based physical activity practice, and 3) gaps and barriers for collaboration in promoting evidence-based interventions.

Method: We collected network data from 35 organizations participating in physical activity promotion in Brazil, 13 of which are part of the GUIA project. These organizations were selected using a modified one-step reputational snowball sampling process (Doreian & Woodard, 1992; Farquharson, 2005). An individualized email and a recruitment statement were sent to the agencies inviting them to complete a web-based survey. Those who did not complete the on-line survey after the second email reminder were invited to complete the survey over the telephone. Using R-Statnet, we developed stochastic network models to understand how organizational characteristics predict collaboration of agencies with one another. Participation in GUIA, area of physical activity (research, education, policy development, promotion or practice), distance from the capital, and years working in the area of physical activity, were included as predictors of collaboration between organizations. We compare the subnetwork of GUIA with the general network. A series of goodness of fit tests were used to evaluate model fit.

Results: Collaborative partnerships were less likely to be in place the further away the agencies were from the capital (b = -0.48, std.error = 0.21, p < .001). Collaborative partnerships were more likely to exists if both agencies were research agencies (b = 0.62, std.error = 0.13, p < .001) or education agencies (b = 0.26, std.error = 0.13, p < .05). The longer the agencies have been in the area of physical activity, the higher the likelihood of collaboration (b = 1.02, std.error = 0.45, p < .05). No significant difference was found between agencies that participate and do not participate in project GUIA. More than half of organizations reported that bureaucracy limited the ability to form partnerships and a third identified having different goals, a lack of time, and inter-institutional policies as the main barriers for collaboration.

Implications: Findings on the main predictors of collaboration between agencies will guide future public health efforts on disseminating information and implementing evidence-based programs. The understanding of barriers for collaboration will allow for strategic planning efforts to address gaps and identify opportunities to expand evidence-based interventions.