Using media to impact health policy-making: an integrative systematic review

A protocol for this review is registered in PROSPERO International prospective register of systematic reviews under registration number PROSPERO 2015:CRD42015020243 [21]. While we had planned in the protocol to start the search with the year 2000, following our discussion with a media expert, we changed the date to the year 2005—the year of the rise of “web 2.0”. Web 2.0 allowed users of the web to interact and generate content especially through social media [22]. Around that time, traditional mainstream media also started to more seriously integrate social media within their operations.

We used the integrative review methodology that aims to include a range of studies from different methodological approaches (both experimental and non-experimental). We followed the five stages of an integrative review by Whittmore et al.: problem formulation, literature search, data evaluation, data analysis, and data presentation [23]. This methodology is meant to protect against bias, enhance rigor, and improve accuracy of conclusions [23].

Herein the definitions of terms used in the research question:

We stratified the studies based on the stages of the policy-making process. We conducted thematic analysis of all papers reflecting the role of media as reported in each study and presented the results in a narrative summary by stage. Two reviewers (LBK, NH) identified the themes. A third reviewer (GH) settled any discrepancy. We did not identify a mix of quantitative and qualitative data for any of the stages we examined; therefore, we could not conclude whether the included studies confirmed, refuted, or complemented each other [29].

As for case studies, we summarized them in a table (see Additional file 2) listing the study’s name and country, health topic, type of media used, and characteristics of the media intervention and policy outcome.

Figure 1 shows the PRISMA flow chart summarizing the study selection process. Out of 13,674 citations, we identified 21 eligible studies: 10 evaluation studies and 11 case studies. We excluded 530 articles at the full-text screening phase for the following reasons: not an intervention of interest (n = 148), not about public health policy (n = 155), not an outcome of interest (n = 164), and not the design of interest (n = 63). Additional file 2 provides detailed descriptions of the 11 case studies [30‐40]. We provide below the characteristics and results of the 10 evaluation studies [41‐50].

Table 1 describes the characteristics of the 10 eligible evaluation studies [41‐50], including study design, setting, population, media intervention (type and nature), and outcome assessed.

Additional file 3 summarizes the detailed assessment of the methodological appraisal of the 10 evaluation studies. Using the CASP tool, we judged the qualitative study to be of high quality as it reported on 8 of 10 criteria of the tool [41]. Using the tool adapted from Lotfi et al. [28], we judged the three quantitative survey studies to be at high risk of bias since they suffered from a number of methodological limitations [42, 43, 46]. The same applied to the controlled before and after study [47], using the EPOC risk of bias criteria. We were not able to find tools to assess the methodological quality of the other five studies. So we assessed their limitations narratively (see Additional file 3). Overall, they all suffered from at least one major limitation (mainly related to confounding), so we judged them to be at high risk of bias.

Additional file 4 provides themes and results of individual studies. We present below the findings organized according to the stages heuristic framework of the public policy process and present the overarching effect of media on each stage highlighted in italic font [24].

This integrative systematic review identified 10 eligible studies that evaluated the effects of planned media interventions on the different stages of the health policy-making process, except for the policy evaluation stage. None of the evaluation studies assessed social media interventions. We judged all 10 studies to be at high risk of bias. The findings of the evaluation studies suggest that media interventions may have a positive impact when used as accountability tools leading to prioritizing and initiating policy discussions, as tools to increase policymakers’ awareness, as tools to influence policy formulation, as awareness tools leading to policy adoption (and to gain public support leading to policy adoption), and as awareness tools to improve compliance with laws and regulations. In one study, media-created attention had a negative effect on policy advocacy as it mobilized opponents who defeated the passage of the bills that the media intervention advocated for [44]. One could speculate that the reason for such defeat may be due to more effective counter-media interventions funded and supported by powerful special interests and stakeholders.

In addition, the available evidence suggested a number of factors as possible predictors of the success of the interventions. These include the timing of media publicity, building relationships with the media, the mobilization of opponents, and the concomitant use of other strategies such as enforcement campaigns and community mobilization and engagement [42‐44, 46, 49]. This suggests that media interventions do not occur in isolation and a number of factors should be taken in consideration while designing a media campaign for policy change.

A methodological review of published evidence maps found that the most popular domains used to classify evidence gaps were study design, interventions, setting, population, and outcomes [51]. We present in Table 2 the research gaps relevant to the “use of media to impact health policy-making” according to those domains. Similarly, we used a framework developed by Robinson et al. to discuss the reasons behind the research gaps [52]. Using that framework, we identified three reasons related to the topic under review: (1) the bias in conducting research, (2) the indirectness of the evidence, and (3) the insufficiency of information. First, the included studies suffered from at least one major methodological limitation, mainly confounding, that makes their findings potentially biased. Second, in the majority of the studies, the media intervention was part of a multi-component intervention or part of larger programs, making the evidence indirect. Consequently, it was not possible to isolate the effects of media as stand-alone interventions. This approach to intervention design may be due to the fact that using media as stand-alone interventions might not be sufficient when aiming to impact policy [42], given the multiple and complex factors that affect the policy-making process [13]. One additional limitation is that in the survey and interview studies, the outcomes were assessed based on perceptions of respondents on effectiveness of media interventions and not actual evaluation of effectiveness.

The aforementioned challenges and limitations of the included studies made it difficult to make any inferences on effectiveness as the authors of these papers do not explicitly provide an evidence for the link between the intervention and the outcome. Thus, we reported the results cautiously and relied on what the authors of the included papers reported. Third, our review showed limited quantity of research on the effects of planned media interventions and social media in particular on health policy-making.

Another major challenge in the reporting of most of the included studies is the very limited description of the media interventions, knowing that these qualify as complex interventions. This makes it difficult to understand the specific media intervention that was tested, and what component of that complex intervention might have been effective. Indeed, two other systematic reviews examining the effectiveness of mass media on health services utilization [10] and on reducing alcohol-impaired driving and alcohol-related crashes [53] identified similar limitations in terms of the description of the intervention.

There is also a dearth of studies reporting on the effectiveness of media interventions on health policies in low middle-income countries (LMIC). This may be linked to the fact that the production of Health Policy and Systems Research (HPSR) is still in its infancy in these countries [54‐56].

We are aware of a number of systematic reviews assessing the use of media/social media in clinical medicine and public health. For instance, media interventions were assessed for their impact on smoking cessation [8], reducing the risk of alcohol-related injuries or crashes [53], increasing child survival in LMICs [57], and changing health behavior in fields such as HIV prevention [20, 58] and physical activity [19, 59]. Two reviews identified evidence that mass media campaigns can in fact have positive impact [8, 57]. Patel et al. showed that using social media, especially Facebook and blogs, likely improves care for patients with chronic disease [60]. Stellefson et al. reported that Web 2.0 can benefit older adults in managing their diseases [61]. Social media were also shown to be used for a variety of conditions and purposes in child health [62].

This systematic review can inform researchers and funders interested in understanding the interaction between the media and the policy worlds. Researchers are encouraged to conduct more and better-designed primary research studies on social media interventions given the rise of their use in the recent decade and potential impact on policy-making. Researchers should develop and follow guidelines for designing and evaluating complex media interventions. Moreover, there is a need for better reporting of studies in this field, taking into consideration guidelines for the reporting of complex interventions when describing the media intervention used [63‐65].

This systematic review highlighted the challenges of evaluating the impact of media advocacy on health policy given the difficulties in applying experimental methods, the complex nature of these interventions, and the multiple factors influencing the policy-making process [13, 26]. Future primary studies and systematic reviews should include process evaluation and qualitative components to explore factors behind successes and failures, the impact of the context, and elements of best practices in health policy-making media campaigns [66]. Furthermore, future research should focus on the influence of media on health policy-making in diverse settings. Examining the role of media in these LMICs is important to design context-specific strategies and understand how the impact of media campaigns and interventions can differ across various health systems and socioeconomic realities. Funding agencies are also called to support future studies particularly in LMICs, where research production is still at its early stages, and support capacity building for research in the field of HPSR.

Our findings can also inform, to some extent, civil society organizations, advocacy groups, and researchers working to influence policy-making when considering media as tools for policy change. One particular important finding that emerged is the unintended effects of media campaigns to energize opponents. This implies that any media intervention has to be carefully designed and thought through to take into account such situations.

This review has two main strengths. First, and to our knowledge, this is the first systematic review examining the effects of planned media interventions, including social media, on health policy-making. Second, we have conducted the review using standard, explicit, and rigorous methods [67] and we followed standard methods for reporting systematic reviews [68]. This includes a very comprehensive search, including of the Communication and Mass Media Complete database, the main database for media-related articles. One limitation of this review is the use of stages heuristic framework while some academics consider it to assume linearity of the public policy process that does not exist in real world. However, it did help us synthesize the identified evidence by providing a simplified and useful way in viewing the entire public policy process [24].