1
Speaking in Tongues?
Comparing Stakeholder and Citizen Input in the EU’s Consultation Regime
in a Risk Governance Case
Master Thesis Public Administration International and European Governance Track
Leiden University Anna-Lena Bartz (s2082128) Supervisor: Dr. Caelesta Braun 11.01.2019
2
Table of Contents Page
I. Introduction... 3
II. Literature Review... 5
III. Theoretical Framework... 11
IV. Methodology... 16 V. Empirics... 24 VI. Analysis... 41 VII. Conclusion... 48 VIII. Bibliography... 51 IX. Annex... 59
3 I. Introduction:
''Vaccination is one of the most powerful and cost-effective public health measures developed in the 20th century (...) I find it disheartening to witness children dying
because of low uptake, vaccine hesitancy, or vaccine shortages. Infectious diseases
are not confined within national borders. One Member State's immunisation weakness puts the health and security of citizens at risk across the EU. Cooperating in this area is in all of our interests. Protect our children, vaccinate!" (European Union, 2018).
The following statement by EU Commissioner for Health and Safety Vytenis Andriukaitis accompanied the proposal for a Recommendation for strengthened cooperation against vaccine
preventable diseases on 4 May 2018. The proposal followed Commission President
Jean-Claude Juncker’s state of the union in 2017, in which he called for increased vaccination coverages to prevent further death from vaccine-preventable diseases. One of the most common diseases currently spread across Europe are the measles. This is an increasingly serious problem: according to a 2018 report from the European Centre for Disease Control (hereinafter: ECDC), 57 measles-related deaths EU-wide have been reported since 2016. Currently, measles outbreaks have been reported to increase again this year (BBC, 2018).
The ECDC attributes the steady increase of these diseases to a decrease in vaccine uptake as a result of general vaccine hesitancy. This resurgence of the measles however is not without some surprise: in 2017, the WHO published its report from the 6th European Regional Verification Commission for Measles and Rubella Elimination (RVC). In this report, the experts concluded that the countries under inspection were on the right path to eliminate measles, provided that vaccination coverage rather would not further decrease. Given that the WHO and the EU closely cooperate in the field of public health, this evaluation is highly relevant. However, a year later several outbreaks of measles across Europe have been reported again. The initial implementation of risk governance namely that if coverage rates were kept up the disease would be eliminated, thus failed. In addition, vaccine coverage rates have been further decreasing.
Vaccine hesitancy thus constitutes a considerable risk in European Union public health. Due to the interconnectedness of the European Single Market, diseases are much more likely to spread and cooperation between Member States is crucial to prevent and/or contain this. In the sphere of public health, risks affecting the health of citizens need to be alleviated using insights of the discipline of risk governance. To conduct this fruitfully, risk governance relies on consultations with different sets of stakeholders, in particular with citizens who need to be
4
engaged. However, the raison d’être for a consultation differs depending on the set of stakeholders (i.e. professional stakeholders representing bigger interest groups versus citizen). The European Commission therefore, prior to the proposal of the aforementioned recommendation, consulted various stakeholders and citizens about the current state of EU public health, vaccination, and how to counteract vaccine hesitancy. This consultation constitutes the focus of this thesis’ research. The research question thus is:
How is the input of different sets of stakeholders reflected in the EU’s consultation regime in the area of vaccination?
In order to answer this question, this thesis conducts a quantitative text analysis using R software of aforementioned consultation held prior to the proposal. The consultation was held with stakeholders representing both professional and civic organisations as well as with interested citizens. The goal of this research is to investigate the impact of the input of both stakeholders and citizens on EU legislation.
This is highly relevant as it touches upon several ‘hot button’ issues of the EU today, such as public health, democratic involvement of citizens, and the (perceived) high level of influence wielded by professional stakeholders. Investigating the role of different sets of stakeholders in the area of public health suitably addresses all these issues. The area of vaccination is furthermore a controversial and at times divisive topic in today’s discourse in Europe. While some people accept the importance of getting vaccinated against several diseases such as the measles, an increasing number of people is voicing concerns about the risks of vaccinating themselves and their (young) children. One of these concerns is inter alia that vaccines are being pushed by ‘Big Pharma’ in an attempt to make money. Looking at the actual role of stakeholders’ input from professional and civic backgrounds thus is particularly interesting.
This thesis consists of seven sections and is structured as follows: after this introduction, the second section provides an overview of literature in the field of Risk Governance, linked with insights from consultation and citizen engagement literature. Next, the third section presents the theoretical framework for this thesis and the hypotheses which have been devised to conduct research. The fourth section outlines the methodology and explains how the hypothesis have been operationalised. In the fifth section, the context of the topic of the consultation and the findings are introduced. Following this, section six provides an analysis. Finally, the seventh section provides a conclusion which discusses the limitations of this thesis and offers a recommendation for future consultation regimes.
5
II. Literature Review: Risk Governance and Citizen Engagement
In this literature review, the academic field of Risk Governance is reviewed and placed into the context of the public health sector. In doing so, it uncovers important linkages between Risk Governance and literature on consultation and citizen engagement. Furthermore, the added value of combing these two literatures is outlined. Risk governance is increasingly reliant on communication and cooperation across a diverse set of societal actors. The goal of this thesis, namely to investigate the role of various stakeholders in the EU’s consultation regime, thus relies on insights of both risk governance and consultation literature.
Risks and How to Govern Them
The concept of “Risk” has been of high interest in the social sciences for decades. Scholars have focused on different aspect of what constitutes a risk and how to deal with a risk and its aftermath (cf. Beck, 1992). According to Aven & Renn (2009), risks should be defined as “uncertainty about and severity of the consequences (or outcomes) of an activity with respect to something that humans value” (p. 2). The OECD (2003) further defines this as a “systemic risk”: systemic risks are embedded into the broader context and interdependent on different factors such as social or economic factors. Scholars have also emphasised the pertinent link to the broader concept of governance, which depicts the interaction of various actors of society towards a given goal (Van Asselt & Renn 2011, Van der Vegt 2017). The inclusion of these various actors presents an essential element for this thesis.
Governance in this respect is largely understood as “a horizontally organised structure of functional self-regulation encompassing state and non-state actors bringing about collectively binding decisions without superior authority” (IRGC White Paper 2005; c.f. Rosenau 1992; Wolf 2002). Risk Governance then references the numerous ways different sets of stakeholders are dealing with these uncertainties of a particular risk. The concept of “Risk Governance” is a relatively young discipline, developed to improve policy-makers’ abilities to adequately assess, manage, and evaluate risks. It was first introduced by TRUSTNET (a European think tank funded by the EU) but has been lacking in clear conceptualisation until recently (Van Asselt & Renn, 2011). It particularly lacks a focus on how to involve stakeholders from all spheres of society. The International Risk Governance Council first conceptualised Risk Governance in 2005, introducing the Risk Governance Cycle with four phases: pre-assessment, appraisal, characterisation and evaluation, and risk management (IRGC 2005).
6
This model has been expanded in 2017 to include a fifth phase, namely “cross-cutting aspects” such as stakeholder involvement (IRGC, 2017). The new model can be seen as a reaction to the critique of the 2005 model, such as by Renn, Klinke, and Van Asselt (2011), who define Risk Governance as “deal[ing] with risks surrounded by uncertainty, complexity and/or ambiguity” through various constellations of “actors, rules, conventions, processes, and mechanisms”, with an emphasis on stakeholders (p. 233). Building on this, Renn et al (2011) argue that an effective model on Risk Governance should be characterised by the ability to adapt (p. 235). This is only possible if those dealing with risk are communicating and working with stakeholders and society at large throughout the whole process. Communication here is referred to as “meaningful interactions in which knowledge, experiences, interpretations, concerns, and perspectives are exchanged” (Renn et al, 2011, p. 242). The ‘key challenge’ then is to coordinate this communication across all levels to ensure that risk can be dealt with “in an efficient, effective, and fair manner” (Renn et al, 2011).
This is important: in the beginning, Risk Governance had been characterised by a predominantly technical approach to deal with risks in general, mainly concerned with containing risks (Boin et al, 2008). However, more and more recent works emphasise a normative take (Linke 2011, Renn & Klinke 2014, Van Asselt & Renn 2011). For instance, Renn & Schweitzer (2008) advance the concept of ‘inclusive risk governance’ which calls for stakeholder involvement to ensure “effective, efficient, fair and morally acceptable decisions about risk” (p. 174).
Brender (2014) echoes this idea: According to her, risk governance needs to be evaluated at all levels (inter alia economic, legal) and crucially encompasses the involvement of stakeholders (p. 3, p. 161). Similarly, Löftstedt and Bouder (2011) have remarked upon the recurring situation in which stakeholders and a scared population (amplified by the media) pressure public officials to come up with quick-fix solutions which are often ineffective (p. 411). They argue that public officials need “to become more competent communicators of risk and encourage the development of media guidelines and constructive deliberation” (p.424). A recent contribution of the IRGC (White Paper, Year) has further called for an inclusive risk governance approach. One of the ways to answer this call for more inclusivity is to consult the insights of literature on stakeholder consultation.
Affected by Risk, Dealing with Risk? – Consulting with Stakeholders
According to Renn & Schweitzer (2008), inclusive risk governance is dependent on the processes of inclusion and closure. The sphere of inclusion requires that “the four central actors
7
in modern plural societies: governments, economic players, scientists and civil society organizations” are consulted (p. 175). This is necessary to create “closure”, which refers to an agreement or a compromise of all relevant stakeholders on the matter at hand. In light of this normative or inclusive approach to governing risk, which presupposes the involvement of stakeholders across all societal levels, it seems fruitful to consider literature on consultation (of stakeholders) in general.
Scholars have increasingly elaborated on the importance of consulting stakeholders and representatives of society in recent years (Renn & Walker 2008, Arras & Braun 2017, Renn & Klinke 2014). They have studied the involvement of stakeholders – why, how, to what end – extensively (for example: Jordan & Mahoney 1997, Arras & Braun 2017, Christensen, van den Bekerom & van der Voet, 2017, Bunea & Thomson 2014). Literature can be divided into three foci, namely into the reasoning behind consultation in the first place, the mechanics behind it, and the constellation and mobilisation of stakeholders who are consulted. Firstly, Stakeholders are representatives of certain interests or beliefs, such as business interests or NGOs advocating for environmental protection. Consulting with stakeholders on (proposals for) policies to tackle risk-related phenomena inter alia ensures that those affected by these policies agree with their content. Involving stakeholders serves several purposes such as asking for expertise, addressing legitimacy issues or ensuring (political) responsiveness (Source, Source, Arras & Braun 2017). It can also foster trust of the population in the policy- and decision-makers responsible for dealing with a risk: For instance, turning to the subject of this paper’s case study for a moment, the refusal to vaccinate poses a risk in itself, but the reasoning behind it is due to individual perceptions of the risk posed by a particular disease, and this perception is influenced by historical experiences, religious or political affiliations, and socio-economic status (Larson et al 2011, Horton & Das 2011, Moxon et al 2011, Rees & Madi 2011).
Secondly, the constellation of stakeholders consulted has been subject to research as well (Kurzer & Cooper, Source). The dominance of one set of stakeholders over others is principally dependent on the policy issue that is discussed. Taking the example of committees in three European parliaments, Pedersen, Halpin & Rasmussen (2015) investigated both the composition of actors and the concentration of evidence in informational exchanges with stakeholders. They found that committees can increase the variety in constellation by inviting more diverse stakeholder groups to their talks, but that this might reduce the evidence gathered. However, scholars have identified business interests as dominant in a number of areas: According to Rasmussen & Carroll (2014), “business dominance is very obvious” at EU level and more dominant than other types of stakeholders (p. 457). The dominance of a certain set of
8
stakeholders, however, inter alia depends on how they are mobilised: Simply put, who is most likely affected of a certain policy and feels pressured to share their opinion? The authors argue that one of the more important deciding factors here is the one of costs, i.e. who would have to pay and who would benefit from payment. Expanding on this at EU level, Rasmussen, Carroll & Lowery (2014) clarify that besides financial incentives, political salience of an issue is decisive, too. Crucially, stakeholders representing public interests (such as broader interest i.e. environmental issues or specialised interest i.e. consumer protection) are seen as ‘transmission belts’, telling policy-makers what the public desires. Consequently, policy issues that are more salient, that is, have the potential to affect more stakeholders including the public, are more likely to mobilise at a higher rate (Rasmussen et al, 2014, p. 448). Thus, mobilisation can “refer to the original formation of a given interest organisation as well as to the choices made by existing organisations to become politically active on specific issues” (Rasmussen, Carroll & Lowery, 2014, p. 252). In addition, Yackee (2015) has investigated the impact of societal participation opposed to the impact of businesses. She concludes that the perception that business interests are more likely to be heard by policy-makers than public opinion is detrimental to consultation regimes.
Thirdly, consultation literature has highlighted the issue of ‘insiderness’ as opposed to being an outsider to the exchanges held between stakeholders and policy-makers (Moe 2005, Moe 2006). Who qualifies as an insider? Individuals or groups who enjoy privileged access to policy-makers regularly due to their ability to substantially contribute to the policy-making process (Bunea 2016, Grant 2000, Eising 2008, Fraussen et al 2014). Often, these insiders are part of groups representing business interests. By contrast, outsiders to the policy- and decision-making processes more often represent societal issues. Remarkably, the abilities to contribute may differ depending on the situation and/or policy: for example, Van Ballert (2015) outlines the European Commission’s use of expert groups in times of heightened uncertainty, i.e. when policy-makers are lacking relevant background knowledge. Similarly, Bunea (2018) identifies certain situations in which the Commission tends to consult more expert groups as opposed to the public. However, scholars have also acknowledged that there is no clear-cut distinction between insiders and outsiders, as there are inter alia varying degrees of insiderness (Fraussen, Beyers, & Donas 2014, Bunea 2016).
Once stakeholders have gained access to policy-makers through i.e. the medium of open consultations or a permanent position, it is crucial to represent their interests. This is done through ‘framing’. Following Baumgartner & Mahoney’s explanation (2008), “policy advocates hope to see debates defined in a manner favourable to their position, and with their
9
allies they work to achieve this” (p. 436). To frame a particular issue in a certain way then is an important step to ensure that one’s interest is adequately represented in potential policies. Here, Börang & Naurin (2016) propose the concept of ‘frame congruence’, namely the “degree to which lobbyists preferred frames correspond to the perceptions of the policy proposals among policy makers (p. 501).Eising, Rash & Rozbicka (2015) further argue that these frames can be understood in two categories, namely generic (institutional) policy frames and specific policy frames. Now that stakeholder involvement has been covered, it is important to look into the insights of literature on citizen engagement. While it is important to consult ‘professional’ stakeholders, citizen consultation can provide policy-makers with useful input.
Involving Stakeholders at the European Level – Do Citizens Have A Say?
The issue of stakeholder involvement in risk governance is particularly pertinent to the European Union level. Due to the scope of Union policies, affecting 508 million European citizens, consulting with representatives of certain interests is crucial to fulfil the EU’s democratic obligations. Hence, Greenwood (2011) specifies that “EU institutions use advocacy groups as the best available proxy for an otherwise absent civil society” (p. 2). Arras & Braun (2017), for example, illustrate that EU agencies involve stakeholders in their policy-making processes to fulfil agency needs such as expertise or organisational capacities as well as reputation-building. Fung (2015) even argues that “participation can be an effective means to accomplish the values of good governance” and thus suitably address legitimacy issues (p. 515). Stakeholder involvement is presented as an “instrument of legislative control”, however in different forms (Bunea, 2016). For example, differences in stakeholder constellations such as permanent bodies as opposed to public consultations, show that involvement can fulfil different agency needs.
The issue of public consultations is also addressed by Quitkatt (2011): in its early stages, online consultations were presented as a tool to improve governance and tackle the democratic deficit by involving society in decision-making processes. However, at the time, Quitkatt criticised the lack of transparency as well as their exclusionary nature. Liu (2016) similarly explores the consultation of society: technological advances, here represented by the possibility of surveying citizens online and thus reaching potentially more people, are considered to be advantageous in reducing the power of a selected elite of civil servants and thus reducing the apparent democratic deficit. Another obstacle faced to counteracting the democratic deficit is found by Roed & Hansen (2018): while (online) consultations do allow the participation of various groups and individuals, the authors
10
identify a certain ‘participation bias’. This bias is showcased when groups have to outweigh costs against benefits of participating in a consultation. Hence, groups with for example bigger financial resources tend to participate more than groups who are not always able or willing to afford the cost of meaningful participation.
Interestingly, the focus on citizens’ involvement in general and their input in particular is currently lacking in research. While some works have investigated the question of “why” citizens should be involved in these consultations, the question of “how and to what end” has largely been neglected (Yackee 2015, Marxsen 2018, Kohler-Koch & Finke 2007). This is particularly lacking in the context of multilateral organisations’ risk management (Brender, 2014, p. 172).This is odd because the 2015 Better Regulation guidelines adopted by the Juncker Commission expand the use of (online) consultations and aim to improve transparency.
However, Kohler-Koch & Finke (2007) have argued that the European Commission has been advocating for the inclusion of societal actors in their consultation regime since the 1990s. They outline the involvement of societal actors in three phases: during the first phase, in the beginning of the European project, mainly targeted consultations were conducted, focusing on expert groups. Yet, in the in the second phase during the mid-1980s, the Commission widened the net of participants by establishing inter alia the European Social Dialogue with a variety of actors. The third (and current) phase, starting in the mid-1990s, has seen an increase in participatory governance. Here, the 2001 White Paper on Governance can be cited as one of the stepping stones towards inclusive consultation regimes. It would thus be fruitful to look into the status of the most current consultation regime.
Marxsen (2015) has examined the Commission’s consultation regime at the time, criticising the lack of participatory chances for citizens (p. 258). While open consultations are designed to include individuals, there is a clear understanding that citizens need to “join forces to [meaningfully] intervene at the European level” (p. 276). It should be noted, however, that many of the more detailed critique of Marxsen has partly been addressed by the Better Regulation guidelines (for example: Marxsen criticised that survey results were not accessible, currently they are published at the Commission’s website).
Risk Governance in European Public Health: Is the Public Engaged?
According to Van der Vegt (2017), the demand for more public engagement in the European Union in regard to risk management erupted with the outbreak of the BSE crisis in 2004. Since then, public engagement has been considered crucial particularly in risks pertaining to
11
public health and natural disasters (Renn & Walker, 2008). Interestingly, the sphere of public health in risk governance is predominantly characterised by security arguments instead of an emphasis on actual health risks. According to Brown and Harmann (2011), “part of global health governance involves the utilisation and framing of fear and susceptibility to risk as an instrument to help shape global health policy and to elicit political and technical mobilisation and cooperation from a vast array of stakeholders” (p.773). This framing interlinks security with health and heavily impacts the legislative process. Understanding these mechanisms is therefore crucial to understand how and why public officials cope with risks in the sphere of public health (Battams Year, Bretherton & Vogler 2006, Roller & Chang 2016). However, managing public health risks by involving the public is increasingly considered as essential: In his critique of a lack of regulatory studies concerned with EU public health, Flear (2015) recognises the potential threats of ‘public health problems’ and their capacity to negatively affect a society (p. 13, p. 2). In his view, to alleviate this potential threat and to improve risk regulatory governance, citizen participation is vital (p. 6). Moreover, he argues that public health is considered at the European level because it has the potential to spill-over onto other regulatory fields: “the regulation of public health is about ensuring the security of the internal market in order to ensure its optimisation” (Flear, 2015, p. 170). The issue of public health is thus highly relevant to the European Union and its Member States.
This literature review has given an overview of the field of risk governance, particularly risk management. It has linked the risk management stage of the widely accepted risk governance cycle with literature on consultation, particularly consultation regimes of the European Union. An emphasis here lies on the reasons for and ways of engaging with societal actors. Lacking hitherto is more research into the actual value/effect of the contribution of societal actors. More specifically, what is the role of different sets of stakeholders in consultation regimes, particularly in the EU’s public health policies.
Furthermore, the review has showcased linkages between one of the risks in need of public engagement, namely public health. In a nutshell, when managing public health related risks in the EU, consultation with societal actors seems essential. Existing literature has dealt with each topic on its own but has not yet linked the issues to investigate this field. Therefore, the following section presents a theoretical framework aimed at addressing EU consultation regimes in the field of public health related risk management.
12 III. Theoretical Framework
This section presents the theoretical framework constructed to guide this research. It relies on the propositions of the International Risk Governance Council (IRGC) as well as on insights from literature on consultation. First, it presents general conceptualisations of the respective theories. Second, it presents the hypotheses derived from both the literature review and this framework. It thus combines insights of the fields of risk governance, consultation (and citizen engagement) to present a comprehensive account of different sets of stakeholders’ engagements in a risk governance case in the sphere of public health.
Risk & Governance – Conceptualisation
Before this framework can introduce the concept of risk governance, its two components need to receive more attention. Firstly, for the purposes of this research, ‘risk’ is understood as “an uncertain consequence of an event or an activity with respect to something that humans value” (IRGC White Paper, 2001, p. 19).Risks are furthermore understood as systemic as they are interlinked in a set of naturogenic and anthropogenic factors with consequences for multiple contexts (OECD, 2003). Simply put, a risk is something caused by events from nature and/or humans and can affect both in various constellations. In preparing a plan to manage a risk, it is helpful to identify the type of risk that needs to be dealt with. The IRGC presents a taxonomy of risks, categorising risks into six different types: three agents (physical, chemical, biological), natural forces as well as hazards (social-communicative, and complex) (IRGC White Paper, 20xx, p 20). Due to the scope of this research, which is concerned with (European Union) Public Health, risks pertaining to this area can be categorised as a biological agent as they are mainly related to the spread of inter alia viruses. It thus renders necessary the question of what exactly constitutes a (public) health risk. From a medical perspective, the state of health refers to the “absence of disease” (Flear, 2015, p. 23). However, health related to public health is usually understood from a more normative perspective. One of the most widely accepted definitions, the WHO’s, refers to health as a
“state of complete physical and mental and social well-being, and not merely the absence of infirmity” (WHO, 1948).
The emphasis here lies on so-called ‘pre-conditions of health’, such as for instance social well-being. This presumes inter alia necessary infrastructure such as housing, access to medical help, but also societal factors to support mental health. It is further “subject to wide individual, social, cultural interpretation, is produced by the interplay of individual perceptions and social
13
influences” (Jones, 1994 cited in Flear, 2015). This is what creates ‘public health’ (Flear 2015, Jones 1994). Moreover, since ‘health’ is intertwined in the context of society it is clearly interlinked with the second concept, namely the concept of governance.
Following this, the concept of ‘governance’ requires further elaboration. One of the main obstacles to adequately capture ‘governance’ lies in the ambiguity of it – many scholars struggle to agree on a definition suitable for all ranges of governance. Peters (2012) boils the concept down to the following:
“Societies require collective choices about a range of issues that cannot be addressed adequately by individual action, and some means must be found to make and to implement those decisions” (Peters, 2012, p. 20).
The focus on collective action is increasingly necessary in today’s globalised context, as well as in the sphere of the European Union, encompassing several state and non-state actors.
Risk Governance Defined
Having briefly summarised the two components of this research’s theory, we turn to ‘risk governance’ itself. This research utilises the framework of the IRGC for the following reasons: firstly, the IRGC is a widely accepted organisation in the (academic) field of risk governance research. In addition, as outlined in the literature review above, the IRGC provided one of the first comprehensive models of risk governance and has been devoted to the topic since 2001. Secondly, the framework fits within the scope of this research, as it acknowledges the importance of participatory mechanisms. Thus, according to the IRGC (2001)
“risk governance includes the totality of actors, rules, conventions, processes, and mechanisms concerned with how relevant risk information is collected, analysed and communicated and management decisions are taken“ (IRGC White Paper, 2001, p. 22). Expanding on this, risk governance increasingly requires that decisions are taken in cooperation with a variety of different societal and non-governmental actors. More specifically, this requires
“the inclusion of stakeholders and public groups within the risk handling process and, consequently, on the establishment of adequate public-private partnerships and participatory processes” (IRGC White Paper, 2001, p. 49).
Thus, in the framework for risk governance, the IRGC refers to stakeholder engagement as one of the “cross-cutting aspects”, meaning it is an important feature throughout the process of governing a risk.
14
Figure 1: Framework for Risk Governance by the IRGC (2018)
How do we involve stakeholders? – Consultation and Citizen Engagement
Having established the importance of stakeholders in the framework of risk governance, it is necessary to elaborate on the nature of stakeholders in general, and the way through which they are engaged, namely the process of consultation. Once again referring to the IRGC, they define
“stakeholders as socially organised groups that are or will be affected by the outcome of the event or the activity from which the risk originates and/or by the risk management options taken to counter the risk” (IRGC White Paper, 2001, p. 49). From the perspective of consultation literature, this encompasses non-governmental actors, societal groups, businesses, as well as governmental actors – in brief, everyone whose interest is ‘at stake’. This is because all of these groups would likely be affected by a (public health) risk and would thus be eligible to engage in the risk governance process. In short, everyone should be consulted when it comes to issues relating to the sphere of public health. Consulting with stakeholders is an active process which gives both sides the opportunity to ‘take notice’ of the others’ interests, viewpoints and opinions (Stewart, 2009, p. 8). Engaging with stakeholders thus can be defined as a set of
15
“deliberate strategies for involving those outside government in the policy process [as] ways of making policy decisions and ways of implementing them” (Stewart, 2009, p.3). There are different types of engaging stakeholders of which consultations are one. They are understood as an ‘information exchange’ to identify preferences across a wide set of representatives. One hindrance that can be encountered here is mobilisation/participation bias, resulting from a bias in selecting only a certain group of stakeholders to consult or failing to mobilise a diverse set of stakeholders (mainly society) to be consulted. Thus, the IRGC (2017) highlights the need to conduct consultations “so that the various actors are encouraged to contribute to the process in those areas in which they feel they are competent and can offer something to improve the quality of the final product”.
In a nutshell, the theoretical framework of this thesis consists of two building blocks, namely those of Risk Governance and Stakeholder Engagement. In the sphere of public health, risks affecting the health of citizens need to be alleviated using risk governance. To conduct this fruitfully, risk governance relies on consultations with different sets of stakeholders, in particular citizens who need to be engaged. However, the raison d’être for a consultation differs depending on the set of stakeholders (i.e. professional vs citizen). Based on this and the review of the literature above, the following hypotheses have been derived from both this framework and the afore literature review.
Hypotheses:
H1: The input of (professional) stakeholders is used and their ideas are incorporated into
the proposal more than the input of citizens. Alternatively:
HA: Citizen and other stakeholders’ input is used in the same way - their input is equally
reflected in the proposal.
This section elaborated on the theoretical framework of this research and presented the hypothesis devised for this thesis. The next section discusses how this can be operationalised by presenting the methodology.
16 IV. Methodology
This section presents the methodology of this thesis. Firstly, it provides an overview of the case study and the relevant data. Secondly, it presents software employed to analyse the data. Thirdly, the hypotheses outlined above are operationalised.
IV.I Case Study and Data Collection
The case study of this thesis is the European Commission’s consultation regarding a Proposal for a Council Recommendation for strengthened cooperation against vaccine preventable
diseases from 4 May 2018. The proposed Recommendation is the Commission’s attempt to
deal with the increasing spread of vaccine hesitancy and the risks this entails. The controversy is centred around the necessity of vaccines versus alleged health risks of vaccinating. The Recommendation calls for cooperating at EU-level to inform and educate EU citizens about the need for vaccination and to contain recent outbreaks of diseases such as measles. The consultation was held from December 2017 and March 2018, aimed at various stakeholders and EU citizens.
This consultation regime is part of the Commission’s 2015 Better Regulation initiative and consisted of a roadmap consultation, a face-to-face consultation with stakeholders and an open online consultation as well as a targeted online consultation. The two latter ones are of particular interest for this research and had 8927 participants in total: 239 stakeholders (33 to the stakeholders’ consultation and 206 to the open public consultation) and 8688 citizens (representing 97.3% of all respondents) (European Union, 2018c). The data has been gathered and published by the Commission. It can be accessed via the website1 of DG SANTE2 in the form of MS Excel files. The open online consultation aimed at both stakeholders and citizens consisted of a mixed format questionnaire with a total of 22 questions (17 multiple choice questions and five open questions). The closed consultation intended for targeted stakeholders consisted of 43 questions in total, out of which 29 were open questions. Both questionnaires included a complimentary question, allowing for additional remarks and feedback outside the parameters of the questions. The open questions are of particular interest for this thesis, as they give stakeholders with the opportunity to provide individual statements outlining their preferences (given the parameters of the question). These statements thus represent the subject of this research because they are indicative of stakeholder preferences.
1 Website: https://ec.europa.eu/health/vaccination/consultations/cooperation_vaccinepreventablediseases_en 2 DG SANTE is the Directorate-General for Public Health, Food Safety and Consumer Rights thereof.
17
Form? Aimed at? Responses? Consists of?
Closed targeted consultation Stakeholders (selected) 33 43 Questions
• 14 Multiple Choice Questions • 29 Open Questions Open consultation Stakeholders Citizens 206 8688 22 Questions
• 17 Multiple Choice Questions • 5 Open Questions
Proposal after
Consultation
Council n.a. Explanatory memorandum
• outlining how consultation was
conducted and how results filtered into proposal
Proposal text
Table 1: Overview Data Structure
IV.II Method - Quantitative Automatic Content Analysis
In order to analyse the data above, this thesis utilises a quantitative content analysis. It can be described as
“a systematic, replicable analysis of text (...) through the application of a structured, systematic coding scheme from which conclusions can be drawn about the message content” (Rose et al, 2015, p. 1).
The analysis, in particular the coding scheme, is aided by software, namely R3. R is a “language
and environment for statistical computing and graphics”, which can be inter alia used to conduct data analysis (CRAN Project, 2018). In particular, the method of text-mining has been conducted in R to generate information about the content of the consultations outlined above. Text-mining essentially describes a way of analysing data in the form of natural language text akin to methods of numeric analysis. Essentially, the method asks what topics are discussed by whom and how? This gives crucial insights into whether the input of stakeholders or citizens is used more in the resulting proposal. The table below outlines the steps taken to analyse the data. The steps have been inspired by Silge & Robinson’s (2018) text-mining approach.
18
Step Description
1. Prepare data:
o Only responses to open questions o Only responses in English
o Automatic translation of non-English-language responses to citizen consultation
o Convert to R-compatible format o Combine all data into one file 2. In R: import prepared files and clean data
o Remove stop-words (such as “and, “this”, “it”) o Remove white spaces, punctuation and white spaces o Convert to lowercase letters
3. In R: generate term frequencies for each file
o Most common terms used by targeted and non-targeted stakeholders as well as by citizens, and in proposal
o Clean up frequencies to remove “noisy” words such as “vaccines” that are used due to the nature of the topic, but do not contribute to understanding
4. In R: Sentiment Analysis o Using tidytext model 5. In R: Topic Modelling
o Using tidytext model
Table 2: Steps taken to analyse data
The goal of this method to analyse the content of the responses to the consultations with the help of the R software. Firstly, the data needs to be prepared so that it can be utilised within the software. The data provided by the Commission was published in the form of three excel files containing responses to all questions of the respective questionnaires in all 24 official languages of the European Union. In order to conduct a feasible automatic word analysis, only English responses were able to be considered. Due to the fact that the closed consultation with targeted stakeholders was held in English, this did not affect this particular file. However, the open consultation with both stakeholders and citizen, allowed for responses in all official EU
19
languages. The open consultation with stakeholders still generated a workable response. However, the open consultation with citizens had to be cut down considerably. The table below showcases this.
Responses All Responses ... in English
Closed Consultation with Targeted Stakeholders 32 32 Open Consultation with Stakeholders 206 97 Open Consultation with Citizens 8688 52 Table 5: Overview of English-language Responses
To increase the representativeness of citizen input, the responses from stakeholders and citizens in the open consultation were automatically translated using a free translation service4 provided by Microsoft Azure5. The service, ‘Translator Text API’, recognises the original language, translates into the intended language (here: English), and indicates whether translation has been successful (whereby O indicates failure and 1 indicates success to translate). The code can be found in the annex.
Example: Excerpt of Translation Generated with Software
Therefore, the number of workable responses returned to 8688 and 206 responses respectively. Similarly, the responses to multiple choice questions were not considered, as they do not provide statements with the possibility of presenting preferences. The files are then converted into a format that is compatible with R (in this case: csv).
4https://docs.microsoft.com/en-us/azure/cognitive-services/translator/ 5https://azure.microsoft.com/
20
Secondly, the prepared files need to be imported into the R workspace and cleaned to remove all words and signs that could be counterproductive to generating the most frequently used terms. For example, so-called stop-words such as “and”, “it”, “is” can be removed as well as punctuation to avoid the most frequent term being i.e. a question mark. Thirdly, term frequencies for each file can be generated. For clarity, terms that are too noisy (i.e. do not say much about the topic) such as “vaccines” have been removed too. Fourthly, the ten most frequent terms in each file can be used to create a term dictionary. Fifthly, the relative term frequency of the most common frames from each set of stakeholders in the proposal text can be created. More specifically, the fifth step involves a dictionary method to extract the occurrence of frames which illustrates the amount of attention that a text pays to a certain theme. Sixthly, a sentiment analysis was conducted: According to Silge and Robinson (2018),
“When human readers approach a text, we use our understanding of the emotional intent of words to infer whether a section of text is positive or negative, or perhaps characterized by some other more nuanced emotion like surprise or disgust“ (Chapter 2.1).
It is possible with R and the tidyverse package to analyse the words used by stakeholders and categorise them into a binary fashion (negative/positive connotation of a word used) or even into emotions associated with certain word usage. Seventhly, topic modelling was conducted, which identifies two main topics in a given number of documents and associates sub-topics with them. This helps to structure the word frequencies created in steps three and four. To sum up, this method first identifies with which words (word frequency) stakeholders talk about concise topics (topic modelling) and how they feel about it (sentiment analysis).
This method is based on the idea of George Zipf, namely that the relative frequency of a term or a frame occurring in a spoken or written word documentation relate to the importance placed on it. Consequently, “Zipf’s law states that the frequency that a word appears is inversely proportional to its rank” (Silge & Robinson, 2018). In other words, if the word ‘education’ appears more often than the word ‘legislation’, we can assume that i.e. educating people is more important than making laws. For the purposes of this thesis, we can also assume that the more often terms are used by a certain set of stakeholders are used, the more they define the collective ideas of this set.
We also assume that different sets of stakeholders focus on different terms more frequently, and thus their overall frequency differs. By then comparing which (and whose) terms are more dominant in the final outcome, namely the proposal, we can determine whose
21
ideas have been considered more and thus been given more weight. The following section explains how code was employed to work with the data.
IV.III Coding Instructions: Excerpt
This section shows an excerpt of the codes used to analyse the data to illustrate the usage of the software. The codes in full can be found in the Annex. The following explains the steps taken and showcases shortened snippets of the code that was used. For the construction of the term frequency tables (Steps two and three), the pre-prepared data had to be imported into R.
➢ library(tm)
➢ reviews = read.csv("FILE.csv", stringsAsFactors=FALSE)
➢ review_text <- apply(reviews, 1, function(x) paste(x, collapse=' '))
Once the data and the necessary R package (text-mining) were loaded, it needed to be transformed into something the software could work with. In this case, it meant creating a ‘corpus’ which could then be transformed. As a reminder, the goal of this task was to identify the most commonly used terms in each consultation by generating their frequency. This meant that words that could distort the outcome (such as so-called ‘stop-words’, like “that”, “and”, “it”) had to be removed, as to avoid the most frequent term be i.e. “and”. The data was further manipulated to exclude punctuation, higher case letters and whitespaces. Given the nature of the consultation, it was expected that the words “vaccine”, “vaccines”, and “vaccination” were very frequently used and were removed to generate more meaningful terms.
➢ review_source <- VectorSource(review_text) ➢ corpus <- Corpus(review_source)
➢ corpus <- tm_map(corpus, content_transformer(tolower)) ➢ corpus <- tm_map(corpus, removeWords, stopwords('english')) ➢ corpus <- tm_map(corpus, toSpace, "vaccines")
Next, the cleaned-up data was transformed into a Document-Term Matrix, which illustrates ...
22
➢ dtm <- DocumentTermMatrix(corpus) ➢ dtm2 <- as.matrix(dtm)
➢ frequency <- colSums(dtm2)
➢ frequency <- sort(frequency, decreasing = TRUE) ➢ words <- names(frequency)
➢ barplot(...)
As for the fifth step, the aim was to compare the relative frequency of terms in the document of the proposal created by the Commission after the consultation. After the required packages (tidyverse, tidytext, pdftools) were loaded, the proposal’s text needed to be imported and put into a data frame, which holds the body of the text.
➢ this_is_the_proposal <- pdf_text("proposal.pdf") %>% readr::read_lines() ➢ proposal_text <- data_frame(txt=this_is_the_proposal)
Once in a data frame, a Document-Term Matrix was created and once again, unnecessary stop-words and noisy stop-words (vaccines) were removed.
➢ proposal_dtm <- proposal_text %>% unnest_tokens(word, "txt") %>% anti_join(stop_words) %>%
anti_join(data_frame(word=c("vaccines" , "vaccination" , "vaccine"))) ➢ proposal_dtm
Next, the most frequent terms were counted. Using the results from steps two and three, outlined above, the ten most frequent terms from each consultation were used to search the proposal. For example, the ten most frequent terms in the citizen consultation included inter alia the terms ‘healthcare’, ‘citizens’, and ‘information’. We thus create a dictionary of frequent terms.
23
➢ count <- proposal_dtm %>% ➢ count(word, sort= T)
➢ count
➢ frame_a <- c("most frequent words”)
With these frames defined, the term ration could be calculated counting the total number of words and the dictionary to identify the relative frequency of each frame.
➢ total <- proposal_dtm %>% ➢ total[[1]]
➢ total_a
Operationalisation
In order to conduct this analysis, the hypotheses derived for the theoretical framework need to be operationalised. Hypothesis H1 expects that stakeholders’ input is generally more used and
their ideas are more likely to be found in the proposal created by the Commission. Thus, in order to prove this, the most frequent terms used by stakeholders should have a high relative frequency. Consequently, citizen input is less used for the actual proposal and thus should have a lower relative frequency in terms used. An alternative hypothesis HA expects the input of both
stakeholders and citizens to be equally frequent. Therefore, the independent variable is the consultation with stakeholders regarding a legislative proposal. The dependent variable then is the degree of overlap with stakeholder and citizen preferences (frames and sentiment) with the frames and actual initiatives proposed in the recommendation.
This section presented in detail the methodology employed to conduct this thesis’ research, namely a quantitative text analysis using R software to compare frequency of terms employed by stakeholders and citizens in the proposal text by the Commission. In the next section, the findings generated in this way will be presented.
24 V. Empirics: Case Study
V.I Context
This section introduces the topic of this research, namely vaccines and vaccine hesitancy, and places it into the context of European public health. Vaccine hesitancy is considered to be a significant risk which needs to be managed at the European Union level.
Risk: People Are Dying
Risk Governance in the sphere of public health are mainly concerned with communicable diseases. A focus here lies on vaccine-preventable diseases. A vaccine is “a biological preparation that improves immunity to a particular disease” (WHO, 2018a)6. Earliest vaccines
such as against smallpox in the 18th century, and against rabies, typhoid or cholera in the 19th century have greatly contributed to the well-being of humanity (Plotkin, 2014). The 20th century produced vaccines against inter alia polio, measles, or Lyme disease. Currently, the 21st century has produced vaccines against for instance the human papillomavirus (Plotkin, 2014). Besides preventing or easing illness for individuals, national vaccination programmes have also eradicated diseases such as smallpox, polio and diphtheria. Due to the coordination of world-wide vaccination programmes such as under the World Health organisation (hereinafter WHO), smallpox has been eradicated worldwide, with polio and measles seen as candidates for eradication in the near future.
In order to eradicate or prevent a disease from spreading, the WHO recommends a threshold of 95 per cent vaccination coverage. This is necessary to ensure so-called herd immunity, which means that “if enough people in a community are immunised against an infectious disease, there is less of the disease in the community, which makes it harder for the disease to spread” (Australian Government, 20177). Simply put, if a certain proportion of the
population is vaccinated, a disease is unlikely to spread even to the most at-risk members of society, such as infants, the elderly or pregnant women (Van Boven & Van Lier, 2016). However, herd immunity rates across the world have been declining at an alarming rate. In the European Union, for example, herd immunity for some vaccine-preventable diseases such as measles is under the 95 per cent threshold. Therefore, in his 2017 State of the Union, Commission President Jean-Claude Juncker said
6 WHO website, accessed October 2018: https://www.who.int/topics/vaccines/en/ 7https://campaigns.health.gov.au/immunisationfacts/community-herd-immunity
25
“It is unacceptable that in 2017 there are still children dying of diseases that should
long have been eradicated in Europe. Children in Romania or Italy must have the
same access to measles vaccines as children in other European countries. No ifs, no buts. This is why we are working with all Member States to support national vaccination efforts. Avoidable deaths must not occur in Europe” (European Commission, 2017, emphasis added).
The issue highlighted by President Juncker is still a concern in 2018. According to the European Centre for Disease Control (hereinafter: ECDC), the year 2018 has so far seen 33 deaths from measles infection alone. Currently, they are ongoing outbreaks Romania (5 222), France (2 727), Italy (2 295) and Greece (2 290) (Report November 2018, ECDC, 2018). It is noteworthy that 87 per cent of those affected have not been vaccinated (ECDC, 2018). Interestingly, vaccination against measles (often given in a combination vaccine against measles, mumps and rubella (MMR)) has not increased in response to the outbreaks, but rather declined. The Figure below (from the ECDC’s State of Health Report of 2018) depicts this.
26
Another communicable disease, seasonal influenza, is equally widespread: it causes “four to 50 million symptomatic cases in EU/EEA each year, and 15 000 to 70 000 European citizens die every year of causes associated with influenza” (ECDC, 2018a8). In 2009, the European Union
adopted the Council Recommendation on seasonal influenza vaccination (2009/1019/EC). However, a 2014 assessment of the state of play of the recommendation found that threshold coverage rates of 75 per cent were not adhered to in most member states. Why is this still the case? According to the European Union, the roots of the problem lie inter alia with “unequal access to vaccines and the waning of public confidence in vaccination” (Europa, 2018). Vaccine hesitancy is defined as:
“[a] delay in acceptance or refusal of vaccines despite availability of vaccine services. Vaccine hesitancy is complex and context specific, varying across time, place and vaccines. It is influenced by factors such as complacency, convenience, and confidence” (WHO, Report of the SAGE Working Group on Vaccine Hesitancy, 2014).
The loss of confidence in the efficiency of vaccines is a phenomenon which can be found across the word, with the “anti-vax” movement particularly prominent in the United States. Larson et al (2018) define vaccine confidence as “the trust in the effectiveness and safety of vaccines and trust in the healthcare system that delivers them” (p. x). Across Europe, hesitancy to vaccinate especially children has been spreading since the 1990s.
Critique against vaccination has been prominent since its conception and ensuing systematic distribution in the 19th century, however it has declined over the years. Meanwhile, in 1958, a first version of a vaccine against measles is tested, in 1971, the triple vaccine shot against measles, mumps, and rubella (MMR) is introduced. The anti-vaccine movement had been given a boost in 1998 when the medical journal The Lancet published Andrew Wakefield’s study, which claimed a link between MMR vaccination, and autism (Wakefield, 1998). Albeit scholars quickly falsified his study and Wakefield admitted to fraud in 2004, the claims have persevered in the public – especially amongst young parents. According to the Vaccine Confidence Project (2016), confidence in the safety of vaccination is decreasing in the European region, with four EU member states showcasing the lowest levels of confidence out of a survey total of 67 countries. The reasons for this are manifold, with fears of vaccination leading to autism, cancer, or development disorders in young children at the
27
forefront of the (misguided) criticism. Dubé et al (2013) propose a conceptual model of vaccine hesitancy, shown in the figure below. Besides the loss of confidence in vaccines, unequal access to vaccination in general across member states is further contributing to the problem and facilitates the outbreak of diseases. Therefore, vaccine hesitancy and its consequences can be seen as a considerable risk in the sphere of public health.
Figure 2: Conceptual model of vaccine hesitancy (Dubé et al, 2013.
Vaccine Hesitancy at the European Union Level – What can be done?
While the above quotation of President Juncker highlights the urgency with which vaccine hesitancy and a lack of access to vaccination need to be dealt with, the European Union’s competencies do not quite reach the sphere of public health wholly. Public health is still a competency of the individual member states. However, the European Commission has justified their input in accordance with the following legal bases. According to Article 168 TFEU,
“A high level of human health protection shall be ensured in the definition and implementation of all Union policies and activities. Union action, which shall
complement national policies, shall be directed towards improving public health,
preventing physical and mental illness and diseases, and obviating sources of danger to physical and mental health. Such action shall cover the fight against the major health
28
scourges, by promoting research into their causes, their transmission and their prevention, as well as health information and education, and monitoring, early warning of and combating serious cross-border threats to health” (Article 168 (1) TFEU, emphasis added).
Thus, Article 168 TFEU allows for complementary action to member states’ activities in the sphere of public health. In addition, Articles 114 (approximation of laws) and 153 (social policies) TFEU are considered to be lending legal leeway. Exemplary instruments issued by the European Union include Decision 1082/2013/EC to promote crisis preparedness and coordinated responses to cross-border threats. Moreover, the 2014 Council conclusions on vaccinations as an effective tool in public health inter alia emphasised the costly implications of a lack of awareness of the benefits of vaccination in general and disease outbreaks in particular. In addition, the Horizon 2021-2027, as part of the Multi-Annual Financial Framework for the next seven years, currently foresees a new emphasis on issues of public health, aimed at inter alia strengthening health systems and supporting EU legislation on public health (Europe, 2018b9).
Furthermore, the European Union cooperates with several international organisations, notably the WHO. In 2001 and 2015, they committed to cooperation to ensure “better health” (Europa, 2018c10). Regarding communicable diseases, the WHO closely works with the European Centre for Disease Control (ECDC), an EU agency designed to monitor diseases. These coordination and harmonisation attempts are essential, because the risk of disease and a lack of vaccination knows no borders and can thus significantly affect the functioning of the European Union’s Single Market.
V. II Consultation with Stakeholders and Societal Actors
As one of the distinct measures to tackle vaccine hesitancy, the Commission proposed a
Council Recommendation for Strengthened Cooperation against vaccine preventable diseases.
In the run up to the proposal, the Commission conducted several consultations with stakeholders and citizens. The consultation included four main activities, namely a roadmap consultation, an open public consultation, a targeted stakeholder consultation, and a face-to-face consultation. Of particular interest to this research are the consultations conducted online with stakeholders
9https://ec.europa.eu/health/funding/future_health_budget_en
29
and citizens (rows three and four in the table below), running from December 2017 to March 2018. The survey results have mostly been made accessible on the Commission’s web pages.
Form of Consultation Time Frame Format Accessibility/ Addressed to Roadmap consultation
4 weeks Free-text format on the Commission’s website All EU languages
Everyone interested in the topic: citizens, academics, NGOs, lobbyists
Open Public Consultation
12 weeks Web-based questionnaire All EU languages
International organisations, vaccine industry, health professionals’ associations, patient organisations, civil society, scientific community
Targeted Stakeholders’ Consultation
4 weeks Web-based questionnaire English
Registered stakeholders at Health Policy Platform
Consultation n.a. Face-to-face meetings Vaccine industry, health professionals, public health NGOs, international
organisations
Table 3: Consultation Regime. Source: European Commission (2018d)
Consultation – Questions Asked
The open consultation for stakeholders and citizens and the closed consultation for targeted stakeholders consisted of two separate questionnaires with differing questions (both open and closed). Both questionnaires were structured around the three pillars presented in the first draft of the proposal, namely (I) Tackling vaccine hesitancy and improving vaccination coverage,
30
(II) sustainable vaccination policies in the EU, and (III) EU coordination and contribution to global health. Thus, the questionnaire outlined that the goal of this proposal would be to develop “a joint action on vaccination, co-funded by the Union´s Health Programme, will start in 2018, focusing on strengthened interaction of immunisation information systems, fostering of vaccine supply management, enhanced prioritisation of vaccine research and development, and tackling vaccine hesitancy” (Questionnaire, 2017).
Questions were usually asked in multiple choice format, sometimes with the possibility to specify or add to their response in a comment. There was also the possibility to provide lengthier statements to open questions (see example below). Both questionnaires can be found in the annex.
Excerpt of the Questionnaire for Targeted Stakeholders: Example Open Questions (Source: European Commission)
Similarly, the questionnaire for the open consultation consisted of multiple choice and open questions structured around the three pillars of the proposal. However, the open questionnaire was much shorter and provided less opportunities to comment.
31
Pillar Closed Consultation Open Consultation
I 2, 3, 6, 7, 8, 9, 12, 13, 14 (total: 9) 2, 3, 11 (total: 3) II 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 14, 15, 16, 19 (total: 15) 5 (total: 1) III 1, 2, 3, 9 (total: 4) N.A. Additional 1 (total: 1) 1 (total: 1)
Table 4: Questions used for Analysis
Respondents
Chart 1: Types of Stakeholders Responding to Open and Closed Consultations (Source: European Commission, 2018d)
The above chart depicts the different types of stakeholder organisations which have responded to both the open and closed consultation. Most responses were registered from Non-Governmental Organisations (23 per cent), health administrations (15 per cent), and
32
professional associations (12 per cent). Interestingly, 12 per cent of responses also came from “other” stakeholders. Overall, stakeholders seem to represent opinions and preferences from a considerable level of expertise in the field of public health in general and vaccination in particular.
Similarly, the declared working experience of individual stakeholders alludes to a strong level of expertise in these fields, as most come from the healthcare sector and pharmaceutical industry. By contrast, not all citizens declared a relevant working experience, but those who did, mainly come from the healthcare and education sector. This can be seen in the Chart below.
Chart 2: Declared Working Experience of Citizens and Stakeholders (Source: European Commission, 2018d)
The distribution of nationalities of citizens responding to the open consultation is shown in Graph 1 below. It indicates the percentage of responses given by the ten most re-occurring countries; and shows the absolute number of responses in the box at the end of the graph. It shows that the most responses were given by French citizens, followed by Belgian, German and Italian citizens.
33
Consequently, most responses were given in French. However, a considerable number of English responses were also registered., as can be seen in Graph 2 below. Graph 2 indicates in what languages most responses were given by showing the top ten most re-occurring languages.
Graph 2: Responder per Language (Source: the author)
Graph 1: Responder per Country (Source: the author)
34
Findings: Term Frequencies in Targeted and Open Consultations
For each consultation, the most frequent terms used by either stakeholders or citizens have been calculated with the aid of the R software as outlined in the fourth section of this thesis (methodology). Graph 3 below shows the ten most frequent terms used by stakeholders and citizens. The top chart on the left shows the ten most frequent terms in the Commission’s Proposal issued after the consultations were held. The top chart on the right shows the results of the term frequency search for the responses given by citizens in the open consultation. Next, the bottom left chart similarly indicates the stakeholders’ responses to the open consultation. Lastly, the bottom right chart shows the ten most frequent terms occurring in the targeted consultation with stakeholders. The aim of this exercise was to see what terms are considered important by the different survey groups indicated by the frequency of their occurrence.
Graph 3: Most Frequent Words in Proposal (red), Open Consultation with Citizens (green), Open Consultation with Stakeholders (blue) and Targeted Consultation with Stakeholders (purple)
