Evidence and Power in EU Governance of Health Promotion: Discursive Obstacles to a “Health in All Policies” Approach

This article explores the relation between the meaning of what constitutes &#8216;evidence&#8217; in the European Commission (EC) and the Health in All Policies (HiAP) concept. Since the 2006 Finnish EU presidency, HiAP is regularly referred to by the Commission, but has not yet been implemented as an overarching political vision. While there is a growing literature on technical implementation of HiAP, little work has delved into the political obstacles to HiAP. This article explores three ways in which the dominant meaning of &#8216;evidence&#8217; in the EC reinforces neoliberal institutional characteristics in a way that undermines HiAP: The problematization of health reinforces constitutional asymmetry; the definition of &#8216;EU added value&#8217; hampers positive integration; and the politicization of evidence strengthens the Better Regulation meta&#8208;regulatory agenda. The article suggests that the meaning of evidence in the EC reinforces neoliberal rationality present at institutional level, and calls for more dialogue across public health ontologies.


Introducing 'Health in All Policies' (HiAP)
In the European Union (EU), non-communicable diseases (NCDs) are the leading cause of death, disease and disability. An estimated 91.3 per cent of deaths and 86.6 per cent of disability-adjusted life years are caused by NCDs within the EU28 (EU Science Hub, 2019). They are often preventable and determined by modifiable factors such as diet, smoking, alcohol consumption and physical activity. However, there is a growing understanding that targeting public health interventions solely at changing people's behaviour is insufficient and even results in 'victim-blaming' (WHO, 2008;Ayo, 2012;Baum and Fisher, 2014;WHO, 2018a). In turn, public health researchers are increasingly concerned with understanding how social, macroeconomic and political factors drive the NCD burden (Labonté, 1998;Bambra et al., 2005;Navarro et al., 2006;Mackenbach, 2014;Ottersen et al., 2014;Kickbusch, 2015;Schrecker, 2016;Gkiouleka et al., 2018). Their work shows how health is mainly influenced by policies outside the health sector, and consequently highlights the need to consider health across policy sectors.
One way this understanding of complexity was taken up and transposed into policy recommendation, was through the Health in All Policies (HiAP) approach. HiAP can be defined as: '[…] an approach to public policies across sectors that systematically takes into account the health and health systems implications of decisions, seeks synergies and avoids harmful health impacts, in order to improve population health and health equity.' (Leppo et al., 2013, p. 6).
Rather than a pre-defined action plan, HiAP represents a way of working, in which societal health and wellbeing is considered a priority in all areas of policy-making. This approach recognizes that all public policy areas have an impact on health directly or indirectly, especially through influencing socioeconomic equity (Leppo et al., 2013). The essence of HiAP is normative and can be traced back to the Alma-Ata Declaration of 1978 and the WHO 'Health for All' strategy (WHO, 1978(WHO, , 1981. HiAP was raised on the EU agenda most notably during the 2006 Finnish presidency (Ståhl et al., 2006). Indeed, HiAP is relevant not only to local level policies, but also to policies at national, supranational and global level. Given the complexity of health determinants, health promotion at local level is likely to have limited effectiveness if European policies with predominantly non-health interests (such as internal market, competition, trade and fiscal coordination) are not considered through a HiAP lens (Ståhl et al., 2006, p.14). And while costs of health intervention in non-health policy areas tend to be overestimated, the long-term benefits resulting from health and wellbeing promotion are often more difficult to quantify (Sihto et al., 2006;Smith et al., 2010a).
HiAP is both about 'technical' health mainstreaming, and about the adoption of a 'political' overarching vision for a healthier, more just and sustainable society (Ståhl et al., 2006;Leppo et al., 2013). The majority of research on HiAP has focused on technical implementation, 1 and little has been said about political challenges to HiAP, including in the EU (Oneka et al., 2017). However, investigating how power plays out in the uptake (or lack thereof) of HiAP at EU level could help better understand why, despite a particularly rooted neoliberal rationality prevailing in the EU governance, the HiAP concept has nevertheless emerged and become routinely referenced (for example: EU Council, 2006European Commission, 2007, 2009, 2017a. Most importantly it would allow to better understand some of the limits to taking up HiAP as a normative vision. Due to its technical/political duality, HiAP carries an inevitable level of vagueness, and its meaning requires contextual adaptation. With the recognition that all policy areas impact on health, HiAP opens up a terminological 'Pandora's box' of what qualifies as public health promotion, and whether 'health' remains an appropriate term when used in such an all-encompassing way (Synnevåg et al., 2018). Such use of language runs the risk of being perceived as 'health imperialistic' (Banken, 2001;Kemm, 2001). This terminological debate calls for a more explicit discussion of the normative aspects of HiAP, which again justifies studying HiAP from a political science perspective. This article presupposes a fluid understanding of health as physical, mental, social, cultural and spiritual wellbeing, and in turn sees HiAP as a way to foster societal wellbeing broadly understood, rather than a health imperialistic disciplinary agenda. The purpose of this article is to explore how a mutually reinforcing relation between institutions and discourse undermines a HiAP uptake, specifically neoliberal institutional characteristics of the EC, and the dominant meaning of evidence within the EC.

Structure of the Article
The remainder of the article is divided into four parts. The first part introduces the theoretical underpinnings of this article, then outlines its methodological approach. The second part presents the EC institutional context empirically, before situating it theoretically. It contextualizes HiAP within the EC institutional setting, emphasizing some of the neoliberal biases identified in the literature. To do so, it draws on-and explains the concepts of constitutional asymmetry, and the disbalance between positive and negative EU integration (Scharpf, 1998(Scharpf, , 2006Greer, 2014). However, the article argues that, in order to get a more nuanced understanding of the obstacles to HiAP, it is necessary to look beyond institutions alone and explore how knowledge is produced, in particular the meaning of evidence.
The third part introduces the role of the Directorate General for Health and Food Safety (DG SANTE) as a 'knowledge broker' empirically, before situating it theoretically. It explains the importance of evidence and evidence-based policy-making, and unpacks the dominant meaning of what represents 'evidence' in the EC. This is then situated in relation to the literature critical of evidence-based policy-making, and specifically Parkhurst's (2017) notion of evidence-based policy-making biases.
The fourth and final part explores how the dominant meaning of evidence and institutional factors interact in a way that reinforces a dominant neoliberal rationality and undermines HiAP. Three dimensions are developed: the problematization of health as an economic investment and its relation to constitutional asymmetry in the EU; the definition of EU added value and its influence on positive integration; the politicization of evidence and its relation to the Better Regulation agenda.
While dominant discourses need to be understood in their institutional contexts, the article draws attention to the reproduction of the dominant meaning of evidence occurring in spaces outside the EC. This is especially relevant in the context of public health policy, which is rooted in a positivist, hypothetico-deductive research paradigm. This paradigm however, faces shortcomings when used to research political determinants of health (Mykhalovskiy et al., 2019). Ultimately, the article suggests that challenging the limits and constraints of HiAP in a neoliberal setting needs to go alongside promoting more dialogue between (positivist and post-positivist) ontologies in the study of political determinants of health, and calls for the need to foster more explicitly diverse relationships between research and policy.

Theoretical Underpinnings and Methodological Approach
To study ideas in institutional settings, the article considers the role of both institution and discourse (Panizza and Miorelli, 2013;Larsson, 2015;Schmidt, 2017). The term discourse is used to refer to 'ways of thinking and speaking about aspects of reality', and can shed light on the 'inextricable link between power and knowledge' (Cheek, 2012, p. 356). 'Institutions', here, are defined as constraining and enabling constructs co-constitutive of agents acting within them (Schmidt, 2010, p. 48). Defined as such, institutions are not simply existing out there, independently of their active and intentional social construction. They are shaped and constructed by actors themselves. This is an important consideration in order to avoid overly deterministic conceptualizations of institutions as immutable entities.
In turn, this article suggests that limiting the analysis to institutional factors, and neglecting discursive factors, would fail to expose more subtle, often overlooked, obstacles to HiAP. In this case, it would fail to tease out the relation between meaning of evidence, neoliberal rationality, and orthodox public health ontology. The need to take into account institution and discourse is driven by the notion that, while both can be seen as co-constitutive, the latter is more than a path-dependent result of institutional power dynamics.
Discourses, and the meanings they carry, can be seen as manifestations of power present in language (Belsey, 2002). They are not neutral and independent of each other, but contain normative assumption and worldviews, shaping a 'system of meanings' which contributes to creating dominant rationalities (Bacchi, 2012;Panizza and Miorelli, 2013;Flear, 2015). One well-documented such rationality, which clashes with HiAP and the idea of fostering health equity, is neoliberalism. Neoliberalism refers to the widespread governance rationality which promotes market liberalization and economic growth while curtailing public expenditure, and which at the same time shapes the social sphere along the same rationality, by disciplining citizens to become free yet responsible, entrepreneurial subjects (Tickell and Peck, 2002;Rose et al., 2006;Joseph, 2012;Parker, 2013). Exploring the reproduction of neoliberal rationality, consequently, cannot be limited to analyzing institutional factors, but should also occur in discourses, for this can lead to uncovering less obvious spaces of neoliberal reproduction.

Methodological Approach
This article's focus was guided by 28 semi-structured elite interviews undertaken with 29 participants between March 2018 and July 2018 in Brussels and Luxembourg. The research project was initially intended to explore knowledge production in the EU Platform for action on diet, physical activity and health, and its limits to addressing political determinants of NCDs. Interview participants were eight officials of EU health advocacy groups, eight of the EC (seven current and one former official), three of the European Parliament (EP) (one MEP, one MEP assistant and one former MEP assistant), five representatives of health ministries of EU member states, four members of associations representing the interests of the food and retail industry, and one representative of a research and evaluation company. The interviews were conducted by the researcher face-to-face (24), over Skype (2), by telephone (1) and via email (1).
The semi-structured nature of the interviews provided flexibility to develop broader topics, including the evolution of DG SANTE, the EC's involvement in public health promotion, the role of the EP, and latest policy developments in the area. This created a different picture in which unexpected topics appeared as more important than the original focus. These topics were investigated in increasing depth (in two cases, interviewees were contacted a second time to elaborate on the new foci): first, the challenges to implementing HiAP. Secondly, the role of evidence in public health promotion: how to process and review it, how to devise effective assessment methods and identify best practices. Interviewees were split between a majority who firmly adhered to the dogma of evidence-based health policy-making, and a few public health policy-makers and advocates who took a critical stance vis-à-vis evidence-based policy-making.
This article's research question developed as a result of grounded conceptualization (Belfrage and Hauf, 2017). That means it departed from an initial provisional conceptualization of an issue, which was later modified and re-theorized as a result of fieldwork. This process brought to the forefront the ontological dilemma in public health between aspiring to produce 'normatively neutral' knowledge, and recognizing that the political and social dimensions of public health elude objectivity. This dilemma appeared to be salient in the attempts to implement HiAP, because HiAP refers to both a normative vision and a technical process. The EC context illustrates these tensions particularly well because of both its 'technocratic' nature and its institutional neoliberal bias. The research question resulting from this 'retroductive' reflection process (Belfrage and Hauf, 2017) can be articulated as follows: In the EC, how is the dominant meaning of evidence related to neoliberal institutional dimensions, and how does that undermine a normatively meaningful HiAP?
Based on this question, the interview data as well as publicly available EC documents were analyzed using a set of theoretical concepts as a lens. These theoretical concepts are: constitutional asymmetry, negative versus positive integration, and Parkhurst's (2017) evidence biases. They will each be explained in more detail in the relevant sections below. These lenses allowed to make sense of pathways between neoliberal institutional dimensions in the EC and meanings of evidence. The documents collected and analyzed were EC publications concerning HiAP, public health promotion, evidence-based policy-making (including Better Regulation guidelines) and overarching EC strategic documents (including on the EC governance and Juncker priorities).

Contextualizing HiAP within the EU Institutional Setting
Article 168 of the TFEU states that '[a] high level of human health protection shall be ensured in the definition and implementation of all Union policies and activities'. Despite little to no legislative power on healthcare delivery narrowly understood, growing understanding of the inter-sectoral nature of health has led to a growing awareness of how the EU influences population's health nonetheless, for example through shaping determinants of health Koivusalo, 2010;Karanikolos et al., 2013;Jarman and Koivusalo, 2017;Goldner Lang, 2017). The idea that health is shaped outside the health policy area, and that it is necessary to foster an inter-sectoral approach to health promotion was well understood by seven of the eight DG SANTE officials interviewed, 2 for example: When you're looking at the issues like health determinants or disease prevention and health promotion, […] you have to look at the entire picture I think. And that's where of course changing lifestyles is very difficult because of different factors intervening and different cultural aspects, different economic aspects, and it's only by taking a holistic approach that you can hope to make any progress over time. (Interviewee 1).
Existing research around HiAP and its relevance in the context of health promotion has made its way to the EC, at least to DG SANTE. When introduced under the Finnish presidency in 2006, HiAP was welcomed by the EU Council at the Employment, Social Policy, Health and Consumer Affairs Council meeting, during which all of the core messages of HiAP were stressed: a recognition that health is valuable per se, that it is not merely a matter of personal choice, and that health determinants are shaped largely in policy areas outside the health sector. It invited the EC to set out a plan to implement it and include it in its Health Strategy, as well as to improve the knowledge base and identification of the EU policies that have an impact on health (EU Council, 2006). A second set of Council conclusions reiterating the commitments to HiAP was published in 2010 (EU Council, 2010). The 2017 Companion Report published by DG SANTE puts forward as its first conclusion the need to address social determinants of health, the vicious circle between ill-health and poverty, and that this action requires multi-sectoral collaboration (European Commission, 2017a).
However, despite the awareness of health policy-makers, a HiAP uptake as part of a political vision for the EC and a rationality existing beyond public health policy circles, has not been meaningful so far (Koivusalo, 2010;Ollila, 2011;Leppo et al., 2013;Greer, 2014). Indeed, HiAP as a political vision for the EC and the EU, in practice, would entail a radical shift towards a social Europe. Arguably, and amidst the current growth of far-right Euroscepticism, believers in the European project seek to put a (renewed) emphasis on the social pillar of the EU (Ferrera, 2017). While the EU project was primarily one of economic integration, expectations to strengthen its 'social logics' started intensifying as economic integration became deep enough for legal constraints on domestic welfare states to be felt (Scharpf, 2002).

Positive versus Negative Integration
Scharpf explains how 'negative integration', which acts to remove barriers to trade and promote the freedom of the single market, is systematically stronger and faster than 'positive' (that is, market correcting) forms of integration (Scharpf, 1998, p. 164). Because of the executive power of the EC and the authority of the Court of Justice, negative integration was depoliticized, whereas positive integration needs political legitimation and a high level of consensus. This is often hard to achieve due to strongly varying interests and modes of welfare governance of member states (Scharpf, 2006). Relatedly, Scharpf explains that, while 'product related regulation' guaranteeing quality and safety standards have been fairly easy to put in place alongside negative integration, 'process related regulation' at EU level, such as those pertaining to labour rights, are not easily justifiable and agreed upon because they do not directly determine the safety and quality of the end product in the market (Scharpf, 2006, p. 854).
In the case of health promotion, this rationale is illustrated in the importance given to food safety and sanitary and phytosanitary regulations, as opposed to HiAP-inspired health promotiondespite NCDs being by far the biggest burden of disease in the EU. Greer (2014) argues that there are three different 'faces' through which the EU influences health: the first one pertains directly to health policies. The second face is the influence of internal market harmonization on health, which according to Greer is illustrative of the EU constitutional asymmetry. 'Constitutional asymmetry' refers to the priority given to the economy over the social as intrinsically built into the EU project from the beginning (see Section Constitutional Asymmetry and the Problematization of 'Health'). And finally, the third face, which developed most recently, concerns the EU fiscal governance and its impact on health, most worryingly the post-eurozone crisis commitments to austerity (Greer, 2014). Promoting health across sector, in turn, would entail considerable priority shifts and changes not in DG SANTE, but across DGs. Greer goes as far as to argue that health concerns are not the main driver of EU health policy (Greer, 2009).
Arguably, thus, the EU institutional setting is geared in favour of an economic, neoliberal logic of EU integration, which renders a HiAP uptake in the EC unlikely. The EC however should not be seen as an entirely coherent neoliberal monolith. A space for HiAP in the EC did emerge, by virtue of growing understanding of the complexity of NCD determinants, but it has not led to a fundamental transformation. For a more nuanced investigation of the obstacles to HiAP in the EC, it is necessary to look beyond institutions alone. The next sections analyze how the dominant meaning of what represents legitimate, usable evidence in the EC contributes to the neoliberal bias and to undermining HiAP as a normative vision.

Unpacking the Meaning of Evidence in the EC Governance of Health Promotion
DG SANTE is a strong supporter of the evidence-based approach to health policy-making. This emphasis, when put in relation to the HiAP awareness, points towards a contradiction: the simultaneous acknowledgement that public health is political and normative (when talking about HiAP and inequalities), while aiming to approach this issue apolitically, following an evidence-based rationale. This suggests that both political and technical aspects can/need to be disentangled. The EC governance structure has a clear separation between its political and its technical components (European Commission, 2017b, p. 3). In the area of health promotion and NCD prevention, this is reflected by DG SANTE endorsing the role of a knowledge broker. This is aligned with Article 168, which stresses the responsibility of the EU to facilitate collaboration between member states, complement their activities and facilitate knowledge exchange (Article 168, TFEU). The 3rd Health Programme gives DG SANTE a mandate to collect and process policy evidence, as well as support and facilitate best practice exchange and learning between member states (European Union, 2014; European Commission, 2017c).

Issue Bias and Technical Bias in Evidence-Based Policy-Making
The concept of evidence-based policy-making was preceded by, and aims to emulate, the principles of evidence-based medicine (EBM) (Parkhurst and Abeysinghe, 2016). EBM is embedded in the positivist research paradigm characteristic of applied natural sciences. It traditionally regards randomized controlled trials as the gold standard and most reliable type of evidence (Nutley et al., 2013). The adequacy of transposing this approach to social policy, however, is highly contentious and is being criticized in a growing body of literature (Petticrew and Roberts, 2003;Sanderson, 2006;Neylan, 2008;Greenhalgh and Russell, 2009;Bacchi, 2012;Parkhurst, 2017). It has been argued that this understanding of evidence is not useful when dealing with complex, 'wicked' problems (Dryzek, 1990;Sanderson, 2006;Parkhurst, 2017). Parkhurst (2017) categorizes two main types of problems that follow from the transposition of an EBM approach to public policy-making. The first one, 'issue bias', relates to the risk of depoliticizing politics, that is the technocratization of inherently political and value-laden policies. This obscures underlying norms and values that are inevitably present in policies, and shuts down possibilities for dialogue around these norms. The second type of bias relates to the risk of politicization of science, the instrumental (mis)use of scientific evidence for political ends -'technical bias' (Parkhurst, 2017).
In line with the evidence-based policy-making approach as a replication of the principles of EBM, legitimate, usable evidence in the EC and DG SANTE is predominantly conceptualized as quantifiable, measurable and deriving from a positivist approach that isolates causal links. This however neglects the key difference that NCD determinants are largely social rather than medical (Smith, 2013;Parkhurst, 2017). It has already been suggested that biomedical paradigms of health, with their appeal to normative neutrality, are often neatly compatible with neoliberalism (Rushton and Williams, 2012;Schrecker and Bambra, 2015).

The Interaction between Meaning of Evidence and Institutional Context
This final section of the article seeks to tease out the mutually reinforcing relation between the two previous sectionsthe institutional neoliberal biases of the EC and the meaning of evidenceand reflect on how it undermines HiAP. It draws on the theoretical concepts explained in both previous sections (negative/positive integration and constitutional asymmetry, and Parkhurst's notion of biases in evidence-based policy-making) to outline three ways in which the meaning of evidence reinforces neoliberal institutional characteristics of the EC, and in turn undermines a normatively meaningful HiAP uptake: first, the dominant meaning of evidence shapes the problematization of health in a way that strengthens constitutional asymmetry. Secondly, the dominant meaning of evidence informs how EU added value is defined in a way that challenges possibilities for positive integration. These two points relate to Parkhurst's issue bias. Thirdly, the dominant meaning of evidence lends itself well to politicization in the context of the EU's agenda setting guidelines for how to regulate, the 'Better Regulation agenda', which represents an instance of technical bias.

Constitutional Asymmetry and the Problematization of 'Health'
The rationale for the 2014-20 3rd EU Health Programme illustrates constitutional asymmetry in the way it is tailored to feed into the current EC's big political priorities (socalled 'Juncker priorities') largely geared towards economic growth and competitiveness, and towards creating entrepreneurial citizens as the 'social' goal (Juncker, 2014). It claims that health is a prerequisite for a 'smart, sustainable and inclusive economy promoting growth for all ' (European Parliament, 2018, p. 5). The call for more investment in health promotion is still dominantly justified in terms of economic gains and increased productivity rather than referring to social justice and human rights: 'Deaths from major non-communicable diseases translate into EUR 115 billion in potential economic loss each year' (European Commission, 2017a, p. 10). This economic justification for health was criticized by six of the 13 health policy-makers interviewed 3 (former health policy-makers and MEP assistants excluded), for example: Prevention for example, I proposed to appoint a Commissioner for prevention. Because the money we allocate for health disappears in the hospitals more or less. But prevention means to work with healthy people and prevention is still not considered as an economic category but when you land in the hospital then you become an economic category. (Interviewee 2).
This justification seemed to be perceived by those interviewees as nevertheless the only realistic strategy to raise health higher onto the agenda. As this suggests, the economic justification for investing in health is difficult to challenge, and ill-health becomes seen in terms of risk, as an economic burden and a productivity loss to avoid (Flear, 2015;Smith, 2013;). One of the obstacles to challenging this discourse, this article argues, is the dominant understanding of what constitutes legitimate research evidence aligned with a biomedical paradigm of health. In mutual reinforcement with the EU's constitutional asymmetry, this meaning of evidence maintains the conceptualization of health as a means to an end rather than an end in itself. Mobilizing evidence aligned with a view of health as a medical issue and a financial burden to avoid, precludes a focus on the social dimensions of health, such as drivers of diet-related NCDs (Bambra et al., 2005;Stahl et al., 2006;Smith, 2013). Instead, such an understanding of evidence lends itself more to measuring health impact in financial terms. Yet this task may be bound to fail, for the complexity and multi-faceted nature of health does not allow a break down into a small list of measurable indicators. The puzzle of collecting data across member states, defining indicator and measuring impact was a point highlighted by five interviewees 4 as an important challenge: […] when you think of the EU you think of standardized, very strict standardized data collection that applies for a limited number of indicators. A good example is if you look at employment you have four or five key indicators that are used at EU level […] In health we were never able to reach that degree of lightness so we started with something like a couple of thousands of indicators, and now we have a short list that is under 100. But even under 100 … it's not 4 […] (Interviewee 3).
While measuring health impact is certainly an important part of HiAP, the core message of the HiAP approach is not reducible to health impact assessments alone, especially if the way impact is defined buys into the conceptualization of health promotion as primarily an economic investment. In this way, the dominant meaning of evidence risks reducing HiAP to a list of measures with financially quantifiable benefits, and fails to fundamentally challenge the economic, neoliberal narrative of health promotion as an investment with high returns. Despite the likelihood that no health policy-maker interviewed would think of health as merely an economic investment, the assumption that this is a necessary 3 Interviewees 2, 6, 12, 13, 14, 15 out of 1, 2, 3,5,6,7,9,10,11,12,13,14,15. 4 Interviewees 1,3,6,7,16. frame, and that a radically different frame would be inadequate, is maintained through the overarching neoliberal rationality that is manifested in the meaning attached to evidence.

Positive Integration and EU Added Value
'EU added value' is an important criteria guiding resource allocation (European Commission, 2017d. Indeed, the upcoming EU Health Budget prioritizes what is straightforwardly perceived as having greater EU added value, such as collaboration on rare disease diagnostic and treatment, digitalization of health and care, infectious disease threats, and supporting EU health system reforms (European Commission, 2018a). Arguably, the accent put on added value might reflect a need to clearly define the role of DG SANTE at a time where the mantra '[The EU should be] big on big things and small on small things' (Juncker, 2014;European Commission, 2017e, 2018b prevails. Admittedly, DG SANTE in particular works with a small budget, 5 which significantly constraints what it can do and justifies this type of prioritization. However implementing HiAP is not the responsibility of DG SANTE alone, and requires an EC-wide, even EU-wide commitment. When it comes to health and health promotion, EU added value has been narrowed down along very specifically defined lines. Indeed, the Consumer, Health, Agriculture and Food Executive Agency (CHAFEA) has elaborated a formalized 'EU added value score' from 1 to 10 to evaluate potential funding under the 3rd Health Programme (European Commission, 2017f). This suggests that EU added value should be objectively quantifiable. CHAFEA streamlines seven EU added value criteria: best practice and knowledge exchange; benchmarking; (multi-stakeholder) network building; addressing cross-border threats; health issues where internal market is strongly concerned (for example, patient mobility); innovation in healthcare; and optimizing the use financial resources (avoiding duplications). The executive summary of the mid-term evaluation of the 3rd Health Programme suggests that it may be sensible to shift the focus away from issues as complex and broad as NCDsexcept for the knowledge sharing angledue to the difficulty to clearly define the scope of action and avoid spreading resources too thinly (European Commission, 2017d, p. 4).
Commonly agreed in interviews 6 and present in DG SANTE documents was the notion that one of the most important EU added value in public health promotion is sharing evidence and best practices (European Commission, 2017a. This implies that EU competencies themselves are of little relevance to health promotion. However, going back to the HiAP logic and recognizing the complexity and interrelatedness of political determinants of health, EU policies appear indeed much more relevant to health promotion and NCD prevention (see introduction). To meaningfully take such a broad vision of health promotion into account would be a complex task related to the more fundamental debate around promoting a social logic of EU integration.
The meaning of added value as mainly evidence sharing suggests an instrumental relationship between evidence and policy. However, research utilization is far from limited to an instrumentalist view of research 'filling a knowledge gap' and providing ready-touse solutions to problems (Weiss, 1979). The research/policy relationship can also be a more fluid one of awareness raising, where theoretical perspectives slowly inform different ways to make sense of issues, potentially redefining policy agenda on the long run (Weiss, 1977). On the topic of social determinants of health, Smith (2013) explains that an approach to thinking about evidence and policy efficiency in narrow problem-solving terms leads to developing health policies that target low socio-economic groups, following an idea of 'closing a quantifiable gap'. But it precludes policies that modify the system which brings about socio-economic inequalities (Smith, 2013).
Applied to the EU, a narrow understanding of added value related to an instrumentalist view of the evidence/policy relationship, confines the responsibility to promote health and tackle NCDs to member states or local authorities, which might implement health promotion policies targeted at vulnerable groups without calling into question the link between macroeconomic governance and ill-health. This separation between EU affairs and (national) public health also relates to the difficulty and/or reluctance (justified or not) of member states to envisage and work towards more 'positive' integration. Meanwhile, it shuts down possibilities to think of structural changes, such as those advocated by the HiAP approach. In that sense, the meaning of evidence and in particular how the relationship between evidence and policy is thought about, reinforces the neoliberal rationality prevailing at EU level, as it deters from thinking about potential EU added value that would lie in fostering structural changes, in moving towards more social EU integration. Instead, it implicitly confirms the idea that EU added value in health promotion is limited to evidence exchange between member states.

Better Regulation and Lack of Evidence
There exists a wealth of medical, statistically significant evidence regarding the pathological effects of poor nutrition or physical inactivity. But as suggested throughout the article, isolating the effect of a particular policy, measuring and proving its effectiveness or harm is often very challenging or even impossible, given the complex, multi-factorial nature of the issue, and the time needed for outcomes to become visible. This does not mean that research evidence on that matter does not exist, but that it is not necessarily the kind of evidence that is perceived by the EC as usable, namely narrowly measurable and statistically significant. In turn, 'lack of evidence' can become a justification for inaction. In the EU, calls from industry for increasingly intensive impact assessments have been criticized for representing a way to oppose and delay proposals for public health regulations (Smith et al., 2015;Corporate Europe Observatory, 2016).
This issue is particularly relevant in the EC, which mainstreams the Better Regulation approach across DGs, as a cross-cutting priority. Among other things, Better Regulation makes mandatory stringent impact assessments of all new major EU policy, and embeds a culture of close stakeholder consultation (European Commission, 2017g). Some researchers have shown how corporate interest, notably the tobacco industry, lobbied to shape the impact assessment in a way that prioritizes business' economic interests over other kinds of impact, including health (Smith et al., 2010b). The same study shows how Better Regulation empowers corporate actors by involving them at an early stage, giving them space to challenge potential and existing legislation. The instrumentalization of evidence as a means to oppose public health legislation, and using complexity as an excuse for inaction were the main points raised by the three interviewees 7 who expressed critical views towards the role of evidence in public health policy-making: Impact assessment now has become much more complex.
[…] It's also become a battle ground; so again [regarding] the most recent tobacco control directive. An impact assessment was done […]. And then the industry lobbied against it, called it into question, complained to the impact assessment board in the EC. So the EC had to, they decided to commission not just one new impact assessment on the draft directive, but they commissioned 5 impact assessments. Which built in another 2, 3, 4 years delay. (Interviewee 4).
This article stresses that because the meaning of evidence is narrow and carries much authority, especially in the technocratic context of the EC, evidence is easily instrumentalized to oppose public health regulation. This is even more relevant in light of the Better Regulation governance culture porous to interest representation and lobbyism. This 'technical bias' is a particularly pervasive lobbying tool, because it consolidates current power dynamics and a governance culture in which public health and societal wellbeing is not given priority over other interests. The erosion of the precautionary principle (that is, the possibility to regulate based on suspicion of risk in the absence of clear scientific evidence, see Smith et al., 2015) through increased requirements for a particular type of evidence undermines possibilities for policies geared towards structural, political change, which again limits the scope for a meaningful HiAP uptake.

Conclusion
This article has looked at how the dominant meaning of evidence represents an obstacle to implement HiAP in the EC. It explored three ways in which the meaning attributed to the notion of evidence interacts with neoliberal institutional factors and reproduces a neoliberal rationality which hinders a normatively meaningful shift towards HiAP. The dominant meaning of evidence shapes the problematization of health in a way that reinforces constitutional asymmetry. It informs the definition of 'EU added value' in a way that precludes the possibilities to consider positive integration. Finally, it is susceptible to politicization by private interests, especially in the Better Regulation context.
The dominant meaning of evidence does not appear a priori, and should be understood within the institutional context: indeed, the use of evidence to quantify health investment in monetary terms, the technocratization of EU added value and the instrumentalization of evidence by private interests can all be seen as political strategy to maintain the status quo. In turn, it is by no means argued that the meaning of evidence is the sole, independent factor that undermines HiAP in the EC. It is nevertheless important not to reduce discourses to the inevitable result of institutions and political strategy. This article suggests that the dominant meaning of evidence is also an often overlooked site of interaction between a biomedical paradigm of health and neoliberalism. Pointing this out can open up interesting avenues for engagement between critical EU studies and public health.
Instead of calling for more evidence on HiAP to support its implementation, this article concurs with Hawkins and Parkhurst (2016)