Comparison of stakeholder coalitions across pricing policies designed to improve public health, as represented in UK newspapers: a discourse network analysis study

Abstract Background New methodological innovations are needed to better understand the role of complex stakeholder networks in gaining support for contested health policies. The news media provides a valuable setting for stakeholders to present arguments for and against such policies. We used two examples of UK policies, minimum unit pricing (MUP) for alcohol and the soft drinks industry levy (SDIL), to assess the value of discourse network analysis in comparing the competing discourse coalitions evident in the UK news media. We aimed to assess the similarities and differences in the composition and structure of policy discourse networks, and to find out whether public health advocates could use network insights to counter industry opposition to health policies. Methods Existing discourse network analyses for MUP and SDIL were harmonised in Visone to allow direct comparison. We applied a common tie-weight threshold to reduce ties to robust argumentative similarities and used the Girvan-Newman edge-betweenness community detection algorithm to identify clusters of stakeholder subgroups with argumentative similarities within the discourse networks. We calculated network measures of size, density, and E-I index and used them to compare the principal coalitions in each network. Findings Both networks involved a similar range of stakeholder types and formed two discourse coalitions representing proponents and opponents of the policies. The SDIL network was larger (175 stakeholders) than the MUP network (87), particularly the proponents’ coalition (109 stakeholders for SDIL compared with 33 for MUP). The same concept (“policy is supported by the evidence”) was the most polarising concept in both networks. We were able to identify tight discourse coalitions of manufactures and commercial analysts acting in opposition to policy supporters in both debates. Public health actors appeared to be siloed in their areas of interest. Interpretation Discourse network analysis enabled a novel direct comparison of the discourse coalitions across two highly contested pricing policy debates, allowing visualisation of the complex network of actors and relationships operating to potentially influence policy-making through the media. This method shows promise for better understanding of the common tactics used by different unhealthy commodity industries to disrupt public health policies and how public health actors could better work across policy and commodity arenas. Funding SH, CB, GF are funded by the UK Medical Research Council/Chief Scientist Office of the Scottish Government Health Directorates, and TH and PL by the University of Glasgow.