The delivery of population health services is a cross-sector effort that involves many different organizations in the community. While research has shown the collaborative nature of public health, very little has examined differences in system composition based on the size of the jurisdiction served. This study uses Social Network Analysis (SNA) methods to examine how the composition of local public health systems differs based on the size of the population served.
A cohort design was used to examine the network characteristics of local public health systems. The National Longitudinal Survey of Public Health Systems (NLSPHS) contains a stratified random sample of the nation’s 3000 local public health officials (n=397) who were surveyed in 1998, 2006, 2012, and 2014 (70% response) to measure the availability of 20 core public health activities within their jurisdictions, the range of organizations that deliver each activity, and the perceived effectiveness of services. The initial sample included health departments serving populations of 100,000 residents or more, however, in 2014 the sample was expanded to include smaller jurisdictions. We analyzed the networks for four key SNA measures, two at the network level and two at the organization level. Degree centralization measures the connectedness of organizations in the network as a whole and network density measures the proportion of total possible relationships that exist between organizations in the network. Degree centrality captures the number of relationships that an organization maintains with other organizations in the network and betweenness centrality measures the degree an organization lies between other organizations in the network. We then used ttests to determine if significant differences exist between small and large jurisdictions.
All U.S. agencies meeting the national definition of a local health department.
Large systems are 20% more central (p<0.01), meaning that less organizations are providing the bulk of population health activities, while smaller systems have more diffuse delivery of population health activities. LHDs showed the highest levels of betweenness centrality and degree centrality, indicating their key roles in connecting organizations in both large and small systems, although our findings indicated that LHDs are 30% (p<0.01) less likely to be providing the majority of population health activities in small systems.
Significant differences exist in the network structure of large and small systems. Our analysis suggests that systems serving larger population sizes are more likely to capitalize on the wide range of partners available to participate in population health actives, while smaller jurisdictions see more diffusion in who is providing the bulk of activities.
Implications for Public Health Policy or Practice:
Network analytics may provide important insight into how the structure of our local public health systems differs based on community size. Stakeholders may find these measures useful in understanding how the role and contributions of organizations that participate in population health shift in small versus large settings, especially as a building block for additional analysis that explores whether more dense networks are related to better outcomes and understanding if smaller communities have enough potential partners available to grow their networks or are they able to do more with less.