Turning to the Source: Assessing the Evidence Sources Used to Describe the Potential Human Health Impacts of Wind Turbines by Public Health Organization Websites and Community Group Websites Using a Social Network Analysis Approach

Abstract

Introduction: Wind turbines are a source of renewable energy that has become more common in Canada in the past decades. Concerns have been raised over potential adverse health effects from exposure to wind turbines, particularly wind turbine noise. A disagreement exists over the potential harm from exposure to wind turbines to human health, where many public health organizations state that there are no direct human health impacts from wind turbine exposure, while many community groups state that wind turbines are harmful to human health.

Objectives: 1. Determine the types of evidence cited by community group websites, and by public health organization websites, to support their respective positions on the potential health effects of wind turbines; and 2. Assess the pattern of citations or links to the evidence used by community groups and public health organizations to characterize and interpret these patterns of evidence citation and to see whether and how these patterns differ between the two groups.

Methods: Websites of Canadian community groups, public health organizations, environmental non-governmental organizations (eNGOs) and academic organizations were identified using an Internet search strategy. The identified websites with content on wind turbines and human health that met the inclusion criteria were characterised with a data collection tool to gather information about the webpage structure and its links to evidence sources and other organizations’ websites. Descriptive statistical analysis was performed on the website characteristics and evidence and organization citation data. Testing for significant differences between community groups and public health organizations was done using t-tests and chi-squared tests. Adjacency matrices were created to represent the presence of ties between organization websites and between organization websites and evidence sources. Graphs (sociograms) were created based on the adjacency matrices to visualise the relationship between the different types of organizations as well as between organizations and evidence sources. Additional centrality measures were calculated for the visualised networks and representations of structural equivalence were created to determine whether nodes in the network were similar.

Results: 67 identified websites met the inclusion criteria: 2 academic organizations (3%), 6 eNGOs (9%), 18 public health organizations (27%) and 41 community groups (61%). Significant differences were found between community group websites and public health organization websites in their position on wind turbines and human health, and the presence of website components (social media or a news section). Community group websites were significantly more likely to cite blogs, news, video evidence, and personal accounts/testimony than public health organization websites, but no significant difference was found in the citation of peer-reviewed literature or grey literature. Significant differences for mean citation counts between community group websites and public health organization websites were found for experimental studies with controls, grey literature, and observational study without controls. Community group websites predominantly linked to other community group websites and public health organization websites predominantly linked to government and other public health organizations websites.

Social network analysis of the 67 Canadian organization websites determined that websites tended to link to other organization websites of the same organization type. The network structure lacked a central node and was divided according to the websites’ position on whether wind turbines were potentially harmful to human health—where websites within the network clustered by position. There was structural equivalence between organization websites by organization type, where certain national and provincial websites had similar roles within the network.

The results from examining the network between the 67 Canadian organization websites and the 584 evidence sources identified differences in the specific evidence sources and types of evidence that were cited. When the network analysis was limited to evidence sources with more than two citations, the evidence citations were found to be similar in type (reviews, grey literature and cross-sectional surveys) but varied by the specific evidence source cited. The type of grey literature cited varied by organization type, where community group websites tended to cite grey literature that originated from community groups and public health organization websites tended to cite grey literature that originated from public health organizations, government or industry. Higher quality evidence sources were shared between websites across the organization types, but the lower quality evidence sources citations were predominantly shared between organization websites of the same type.

Conclusions: The network of Canadian organization websites with content on wind turbines and human health was structured according to organization type and position on potential health effects. Grey literature, reviews and cross-sectional surveys were the most frequently cited evidence sources and evidence citation patterns differed by organization type. These results provide a basis for understanding which types of evidence sources are used to substantiate positions on wind turbines and human health and how public health practitioners and researchers can approach the uncertainties in the evidence base on the topic.