Local media monitoring in process evaluation. Experiences from the Stockholm Diabetes Prevention Programme

We present a rationale and approach for longitudinal analyses of media coverage and content, and illustrate how media monitoring can be used in process evaluations. Within a community-based diabetes prevention project, the Stockholm Diabetes Prevention Program, we analyzed the frequency, prominence, and framing of physical activity in local newspapers of three intervention and two control municipalities. In total, 2,128 stories and advertisements related to physical activity were identified between the years 1997 and 2002. Although stories about physical activity were relatively few (n = 224), they were prominently located in all five local newspapers. Physical activity was framed rather similarly in the municipalities. Health aspects, however, were expressed to a greater extent in stories in two of the intervention municipalities. A limited portion (14%) of the articles could be linked directly to the program. It is not possible to assess to what extent the program has had a disseminating effect on the newspapers' health-related content in general, due to weaknesses of the process tracking system and limitations of the study design. Implications for the design is that an evaluative framework should be preplanned and include data collection about media relationships, media's interest in public health, media coverage prior to the program and coverage in other media for comparisons of general trends in the reporting. The material and the current database, however, provide a good basis for quantitative content analysis and qualitative discourse analysis to yield information on the type, frequency, and content of health reporting in local newspapers.