In 2009, Liesbet Hooghe and Gary Marks declared it time for a ‘post-functional’ theory of European integration (Hooghe and Marks, 2009). In the face of deep European integration and deep discontent among many voters, they argued, the old elite consensus on European integration was starting to fracture. Rather than a ‘permissive consensus’ on the desirability of further European integration, Europe had entered an era of ‘constraining dissensus’ in which European integration was a subject of party political contestation. In such an environment, the basically elitist (neo)functionalist model of European integration would lose validity. Voters, awakened, could put an end to previously automatic mechanisms of European integration.

In the aftermath of a series of referenda, the 2014 European elections and the Cameron government’s commitment to a renegotiation and referendum on the United Kingdom leaving the European Union (EU), the argument of Hooghe and Marks’s influential work might seem to have just gained power. In the extensive literature on voting behavior that engages with their thesis, the evidence seems to be good: voters are intruding on the old elite consensus, and the main question is whether identity or economic concerns shape new pro- and anti- integration cleavages. The result for the EU should be that big and visible steps toward further integration are not taken because public and elite support for integration has broken down.

Eppur si muove

Most discussions of Hooghe and Marks’ post-functionalist approach, including the original article, focus on its electoral politics, showing the changing patterns of elite and public opinion with regard to integration. But the integration side of the equation – the actual change in the powers and development of the EU – throws up a number of puzzles. European integration seems to have taken an ambitious step forward since 2009 with the development of EU fiscal governance, including the Troika, the ‘six-pack’ and ‘two-pack’ of legislation, the European Semester and the Treaty on Coordination, Stability and Governance all purporting to insert the EU deeply into the budgeting and policy processes of all member states. Rushing through an important new treaty and several large packages of legislation while putting four Eurozone states into expensive receivership, overriding both elected governments and a Greek referendum, certainly caused some dissensus, but it is not obvious evidence of constraint (Schimmelfennig, 2014).

We suggest that the politicization of Europe in party politics as studied by Hooghe and Marks, and the resulting dissensus, need not logically be tied to the end of neofunctionalist integration, though the combination might speak ill for the long-term legitimacy and stability of the EU. Instead, the fact of public dissent with regard to European integration is compatible with neofunctionalist integration mechanisms as outlined by Haas ([1958] 2004) and Scheingold (1965). In particular, the next section of the article makes the case for neofunctionalism as a theory of integration when paired with a multiple streams analysis of politics that specifies its practical mechanisms. In our reading, neofunctionalist mechanisms work through the three streams of policy, politics and problem that set agendas and specify policy alternatives. Even if member states in the Council, and perhaps MEPs, try to restrict integration at the actual decision stage, the agenda is about European integration because the three streams are biased, by neofunctionalist mechanisms, toward putting further integration on the agenda.

Recent European integration has perhaps been most dramatic in fiscal and economic governance, but the politically sensitive and expensive health sector has also been increasingly incorporated into EU politics and policy despite member state resistance. Our cases are from two areas of health policy: health technology assessment (HTA) and communicable disease control. Health policy is an important test case for Hooghe and Marks’ constraining dissensus hypothesis. Virtually the whole history of European integration in health policy and public health took place in their era of a constraining dissensus, when elites were supposedly disinclined to risk voters’ wrath with further European integration. Treaty bases for health policy are weak because member state governments have long resisted giving the EU a role in health policy. The Lisbon treaty specifically excluded the EU from determining the ‘organization and finance’ of health care (Art. 168) – we could even call that a constraining consensus. But as the articles in this special issue show, the constraining dissensus did not seem to prevent the emergence of an EU health-care policy and a variety of public health policies since 1992 along the three tracks of health-care integration: public health, internal market and austerity (Vollaard and Martinsen, 2016). The engines pushing them forward have been quite traditional in neofunctionalist analysis of the EU, above all the activities and entrepreneurship of the European institutions and the decisions of the CJEU (Greer, 2006). They did not prevent the development of an EU role in one of the oldest social policies and at the same time the longest-standing track of integration in health policy – communicable disease control. Nor did they prevent integration in an area with no obvious need for European integration – HTA. In our cases, it seems more like European integration in a permissive dissensus (Greer, 2013).

Integration Through Agenda-Setting and Alternative Specification

Neofunctionalism focuses on spillover, the ways in which a little bit of European integration can spill over to other areas of European integration (Haas, [1958] 2004). There are two analytically separable routes this can take. One is spillover somewhere in society, as when the development of a pan-European trade in live animals led to the spread of variant Creutzfeldt-Jakob (‘mad cow’, vCJD) disease and pressure for a European policy on trade in blood and blood products as well as much tighter EU regulation of farm animals. In these cases, integration creates a problem in society, and pressure for a European solution to the problem. Such spillover includes the effects of deepening trade and other economic and social relations, which bind countries together more closely and create shared interests and problems in their more integrated sectors that spill over (Stone Sweet and Sandholtz, 1998). The other route of spillover, less studied, is when the existence and activities of the supranational institution (in Haas’ words) create a disturbance in a new policy area – as when the CJEU (then ECJ)’s decisions on patient mobility created incentive for even those who opposed the idea of an EU health policy to start lobbying in Brussels for restrictions on it up to and including EU legislation (Stone Sweet, 2004; Vauchez, 2015). This dynamic has been noted by EU health policy researchers in particular, in this special issue and elsewhere (Martinsen, 2005; Lamping and Steffen, 2009; Martinsen and Blomqvist, 2009; Obermaier, 2009; Brooks, 2012; Lamping, 2013).

The objections to neofunctionalist explanation focus on its indeterminacy (Vollaard et al, 2016): it is unclear about the nature of the policy, the timing of integration or the endpoint. Original formulations do not always specify convincingly why more integration, rather than a retrenchment, come about as a result of European problems (Niemann and Schmitter, 2009; Lefkofridi and Schmitter, 2014), and lean heavily on elite socialization to explain the integrationist effect of spillover. The response of Schmitter and collaborators is to move into a policy typologies approach, trying to differentiate policies and situations that are more or less prone to produce further integration (Niemann, 2006; Niemann and Schmitter, 2009; Lefkofridi and Schmitter, 2014).

The risk of such policy typology approaches is that the mechanisms and micro-foundations producing the policy are left unclear, with a resulting risk of overdeterminism and missing variables stemming from lack of attention to the political process. Simply put, policy typology arguments, for example that post-functionalist EU voters accept regulatory policies but oppose redistributive ones, depend on the assumption that voters agree with scholars about the salience and public construction of policies as more or less redistributive, just, beneficial and so on. Schimmelfennig (2014), responding to this inevitable social construction, focuses on the mechanisms European political elites have used to preserve an essentially neofunctionalist integration process in fiscal governance despite public opposition.

We pursue a different approach, in which we start with a much-validated perspective on policymaking, the multiple streams approach developed by John Kingdon and later extended to EU studies (Kingdon, 2003; Zahariadis, 2008; Princen, 2009; Ackrill et al, 2013; Baeten and Vanhercke, 2016, for a discussion of the extensive validation, see Greer, 2015). Starting with a theory of politics and policymaking rather than of European integration allows us to avoid producing a theory of European integration that fails to match actual political practice. Starting with a theory of politics that emphasizes ambiguity and chance, as Kingdon does, allows us to avoid the problem of excessive determinism with which critics tax neofunctionalism and policy typologies.

Kingdon’s analysis focuses on agenda-setting and alternative specification, rather than the actual decision phase of politics. It posits that there are three ‘streams’ made up of policy ideas, political action and problems. When they come together, with politicians willing to support a policy idea as a response to a problem, the policy alternative is on the agenda. Appearance on an agenda does not of course guarantee passage. Passage occurs at a decision stage with distinct politics. In the EU the decision stage starts with the Commission proposal and ends with the passage or failure of legislation.

The multiple streams approach directs our attention to the nature of the three streams. The ‘coupling’ of policy, problem and politics is indeterminate, as Kingdon stresses, but the three streams are each patterned. Some ideas and policy entrepreneurs are more effective than others for structural reasons (funding, legitimacy). Some kinds of problems are more visible because of problem-setting mechanisms (media, government statistics). Some kinds of political incentives are more powerful than others. If the three streams at the EU level are each susceptible to influences that bias them toward integrative policy, the result will be an agenda-setting process likely to produce integration.

Whether the three streams are biased toward integration is an empirical question but there are mechanisms well known to neofunctionalist analysis and EU politics that suggest they are. In many policy areas, the policy stream is substantively constituted through EU action such as financial support to networks and research (Sanchez-Salgado, 2007), including soft law mechanisms that formulate recommendations, harmonize European practice and legitimate European actions. This can produce ‘cognitive Europeanization’ (Guigner, 2007) in the relevant policy elites, who form into networks with shared objectives and worldviews that then promote European-oriented policy alternatives. The development of EU-focused networks in specific policy areas, with a tendency to promote EU-level policies and solutions, and often governing capacity of their own, has been documented in many fields by authors of quite different theoretical orientations (Shore, 2007; Sabel and Zeitlin, 2010; Puetter, 2012; Fierlbeck, 2014).

In the politics stream, the EU has a large number of politicians whose incentives are to make a visible impact by taking political action (as with most democratic politicians), and who therefore might attempt putatively popular EU policies. These credit-seeking politicians can include Commissioners, rotating Presidencies, MEPs and often health ministers in the Council who seek credit for time-consuming trips to Brussels. The problem stream is comparatively un-Europeanized because the continent’s media is not integrated, but the combination of political incentives and integrating policy entrepreneurs helps to identify or create problems in the EU agenda, while the elites are often quite ‘Europeanized’, if nothing else as a result of EU network-building policies.

Methods and Approach

In short, we propose to start with a theory of politics, the multiple streams approach, which has been repeatedly validated in polities including the EU. The multiple streams framework directs our attention to the sources of integration and policies in the composition of the problem, politics and policy streams at the EU level. If neofunctionalist mechanisms explain integration, by biasing one or more of the streams (and therefore the alternatives and agenda) toward integration, then the need for an overarching post-functionalist theory of integration may be questioned.

The research undergirding the next two sections is based on document analysis, surveys of HTA and communicable disease control and semi-structured face-to-face or telephone interviews in 2013–2015 (N=15 for HTA, N=11 for communicable disease control) with mainly Brussels- or Luxembourg- based representatives of EU institutions and member states, national HTA or public health bodies, one MEP, patient groups and the pharmaceutical and medical device industry. The interviews were used to guide our identification and interpretation of other kinds of data cited in the text, such as official documents and publications by key figures.

In each case, we identify the three streams of policy, problem and politics at the EU level, focusing on the actors and ideas in the policy stream, the manifestation of a problem and the politics that would lead EU-level actors, especially the Commission, to want to put the policy alternative on the EU agenda. In each case, we find a long history of EU-financed, often Commission-led policy development leading up to the alternative (networks of increasingly similar member state actors, and at the EU level an agency for communicable disease control and a network with a permanent secretariat for HTA). In each case, it comes against the background of important diversity between member states, diversity increased by EU expansion and international debates about the importance and proper organization of the field.

This development of the policy stream, against a background of very diverse and fragmented institutional landscapes in the member states, means that ‘European’ solutions such as a communicable disease agency were available when there was a focusing event that created interest in the policy area – communicable disease outbreaks in the case of communicable disease control and a larger EU legislative package on cross-border patient mobility in the case of HTA. When coupled with a problem and Commissioners with an incentive to propose, the result as the alternative appeared on the agenda. In both cases, the result was a decision to adopt that alternative. The outcome was that a history of institutional support for networks developed an EU-level policy alternative and identification of problems as ‘European’ ones, which gave the opening for EU politicians to seek credit for taking the action.

Communicable Disease Control

Communicable disease control is an area with a long history as public policy, and a relatively long history of European cooperation (Liverani and Coker, 2012). It varies enormously across the EU, with member states differing on almost any axis of concern, be it their formal organization, resources, staff, priorities, scientific capacity and militarization (Elliott et al, 2012; Reintjes, 2012; Rechel and McKee, 2014). In the European context, there are a number of obvious collective European actions that could be taken to respond to communicable diseases, including capacity building, improved surveillance and communications, agreed lists of diseases for monitoring (surveillance), shared scientific resources, creating schemes for accelerating vaccine production, developing a framework for joint vaccine and medicine purchasing, and creation of an EU-level coordinator.Footnote 1 EU member states have, since 2001, taken all of these actions. The most symbolically and practically important act, the focus of the discussion here, was the creation of the European Centre for Disease Prevention and Control (ECDC), a freestanding EU agency for risk assessment, preparedness and capacity-building.

Policy

The ECDC is both a typical EU agency (Vos, 2005; Rittberger and Wonka, 2011) and an example of a kind of internationally prestigious kind of public health agency that a number of large EU member states have created (Elliott et al, 2012). Such agencies, in both EU policy and global public health debate, are responsible for preparedness and risk assessment but not the more ostensibly political task of risk management. In other words, ECDC draws on ideas and practice that have been present in EU debates and public health literature since at least the 1980s.

In particular, policy streams in the EU have a long record of discussing agencies as a policy idea for managing relatively technical areas of policy such as food safety, civil aviation safety, medicines licensing or communicable disease (Rittberger and Wonka, 2011; Hervey, 2012). Establishing an EU agency, as a policy idea, has some apparent virtues. It avoids the drawbacks of program financing (for example, different programs for different diseases) such as fragmentation, duplication, shifting and scattered secretariats, and vulnerability to changes in EU funding or personnel. Once it is established that a given area of activity is an EU responsibility, it makes sense to have resources that are not dedicated and specific to a given disease but rather can be redirected as required. Then there is a case for an agency rather than Commission action. Because an agency has a strict legal framework, a specific purpose and specific control mechanisms (for example, a board), the Council and Parliament often regard it as more trustworthy than the putatively empire-building Commission (Pollack, 2003).

In the case of communicable disease control, EU action dated back to the 1990s when a number of mostly northwest European member states were considering creating public health agencies, drawing in part on their experience working on issues such as AIDS (Steffen, 2012). The heads of a number of member states’ relevant agencies and government departments started to meet in an informal ‘Charter Group’ to discuss strategy and coordination (MacLehose et al, 2002) amidst a larger debate about potential EU-level collaboration on communicable diseases (Greer, 2012). The step from a highly fragmented landscape of diverse organizations to an EU agency seemed too large, and instead the EU’s step into the field took the form of a network of responsible member state bodies that would coordinate and share real-time information such as surveillance data on outbreaks with cross-border transmission potential (2119/98/EC). This Decision formalized a Network for Epidemiological Surveillance and Control of Communicable Diseases in Europe, an intergovernmental body composed of member state representatives (mostly from the unit of the government tasked with communicable diseases and international public health cooperation).

Perhaps more significantly, it created a mechanism for Europeanizing the practice of epidemiology, complementing a procedure for developing a shared European set of case definitions, lists of diseases that must be by member states to the EU, and methods for epidemiological and microbiological inquiry (elaborated in Decision 2000/96/EC). The Decision, supported by funds from the slowly increasing EU health research budget and general EU research, was starting to create a European practice of communicable disease control, veering away from alternatives such as reliance on the WHO’s rather patchy system, the US Centers for Disease Control and Prevention or older member state practices. Harmonizing European case definitions and methods made European practice and scientific knowledge distinct from that of others. It also brought implications for member state policies, since once there was a practice, it might be thought inappropriate not to invest in the resources and expertise necessary to carry it out (for example the laboratory capacity to identify an agreed upon important disease).

While the very slow approximation of epidemiological and microbiological practice started, a specialist debate about institutional change continued as well. Unsurprisingly, the European Commission financed research on the quality of EU communicable disease control coordination, and, equally unsurprisingly, researchers found vast gaps and divergences between member states that the EU might address (Brand et al, 2000; Mossialos and Permanand, 2000). The debate hovered around the options, ranging from strengthening the existing networks formalized by the 1998 decision, through to the most ambitious goal of creating an EU agency. By the time of the September 11th terrorist attacks in the United States, then, there was an EU-supported discussion of options for European integration in communicable disease control, against the background of increasing harmonization in the basic underlying work of epidemiology, microbiology, and management and basic agreement on the contours of both European networks and agencies as organizations, and broad public health concepts.

Problem

In communicable disease control, problems are among the most obvious kinds of problems that can appear in politics. Not all political issues are regularly the subject of blockbuster movies or popular games. Problems include anthrax attacks (bioterrorism), pandemic influenza, new epidemic diseases and the spread of serious diseases such as malaria, dengue fever and polio. Since the end of the Cold War, public health advocates around the world had been exploring new ways to link their ideas with security. In particular they latched onto the concept of globalization, and the idea, heavily promoted by American public health agencies, of ‘emerging infectious diseases’. Globalization meant, effectively, that any disease in the world was no more than 12 or 18 h from Frankfurt, Paris, Amsterdam or London airports. That ostensibly new proximity created the challenge of ‘emerging infectious diseases,’ diseases that at least had not emerged in Europe, which might circulate into Europe and threaten its citizens (Weir and Mykhalovskiy, 2009).

In the case of the EU, the discussion in the policy stream was given a very different cast by its confluence with a suddenly very active problem stream: first the September 11th terrorist attacks in the United States, followed by anthrax attacks in that country, followed by the global outbreak of Severe Acute Respiratory Syndrome (SARS) in 2003 and all against a background of increasing nervousness about new strains of influenza that seemed to presage a new pandemic strain. These issues took on added European interest in light of the ongoing construction of a post-vCJD apparatus for managing veterinary health and blood in Europe (Clergeau, 2005; Farrell, 2005; Grant, 2012). They triggered increased attention to communicable disease control and changes within member states. The United Kingdom, for example, created a powerful new Health Protection Agency (Department of Health, 2002). They also emerged as shared problems in Europe, involving multiple EU countries, leading to calls for action in virtually every member state, and calls for some inter-state solidarity for states with weaker public health resources (Greer, 2012).

The tipping point for ECDC was the SARS outbreak of 2003, which pushed the developed alternative of an EU agency – and the lack of a reliable administration for EU-level cooperation – to the forefront. In other words, not only was SARS (or the ongoing threat of pandemic influenza) a problem, the lack of a coordinated European response was made a problem by extensive discussions of the failings of EU coordination in a borderless Europe where diseases could be expected to spread quickly. Now-forgotten events, such as the identification of dead birds in Germany and Croatia that tested positive for the new H5N1 strain of influenza, dominated the attention of EU public health policy makers one of us interviewed in autumn 2005 and made them interested in resolving the multifarious problems of emerging and circulating infectious diseases.

Politics

In the classic rendering of EU politics, the Commission sets the agenda by virtue of its right of initiative. In more complex renderings of EU politics, the Commission is empowered or influenced by other actors, some as clear as statements of the Parliament and Council and some more opaque (Page, 2012; Kleine, 2013). In the case of the ECDC, the politics were obviously favorable to a proposal in the aftermath of SARS, and the Commissioner for health took the opportunity to put it forth.

The Directorate-General (DG) for health, known as SANCO at the time and later renamed SANTE, is one of the smallest and weakest parts of the Commission, reflecting the recency and legislative weakness of the EU in health relative to other policy areas, so Commissioner David Byrne seized the opportunity to claim credit for addressing an issue as pressing as a global emerging disease and threats such as pandemic influenza. Member states were positive for the same reason – the need to be seen to be doing something, simply put – as were MEPs, which is also why the decision phase was over quickly with passage in the first reading (Greer, 2012). His crucial decision, which makes sense as a bid for credit by a minor EU politician, was to opt for an agency as against a stronger network. The idea of an agency and the idea of a need for EU coordination had already been fleshed out in the policy stream, and the frustration of ministers and EU politicians with a stream of bad public health news, from influenza to bioterrorism threats publicized by the US government, to SARS made them receptive to action. The structure of the agency, meanwhile, meant that they could keep the ostensible technical work of risk assessment and preparedness in an autonomous body while keeping member state control of actual response, with the Commission servicing a fledgling and largely intergovernmental Health Security Committee (de Ruijter, 2015).

The result was that it was a relatively safe proposal for the Commission, one supported by the Parliament’s key MEP on the subject (John Bowis, a former UK health minister) and one unlikely to tread on member state responsibilities. Its passage on the first reading, a rarity under what was then known as co-decision, showed how in the EU context three streams could be biased to produce integrating alternatives and a shared European agenda even at time when EU member states were otherwise at odds about issues such as the Iraq War.

Health Technology Assessment

HTA is a multidisciplinary analysis evaluating the medical, economic, ethical, legal, social and other consequences of health-care treatments – drugs, medical devices, surgical interventions and so on. More often than not, its focus tends to be on a health economic summary of ‘value for money’ of new pharmaceuticals, which is used in Europe today to give guidance to public health-care payers if a new drug should be funded or not. HTA is a relatively new scientific field, emerging in the 1970s in the United States and gaining interest among academics and practitioners in Europe in the 1980s.

The history of HTA in Europe has been described by those making it (Drummond, 1987; Banta et al, 1997; Jonsson et al, 2002; Banta, 2003; Jonsson, 2009; Luce et al, 2010) but very rarely by outside observers, such as political scientists (Altenstetter, 2005; Böhm and Landwehr, 2014). This has created a considerable body of accounts of HTA at the intersection of project reports, research agendas, policy papers and professional memoirs, which documents in detail the history of HTA in Europe as well as the European Commission’s long-standing interest in the field. Indeed, the Commission has since the early days of HTA intervened at two interlinked levels. On the one hand it funded projects that focused on the content (scientific and methodological) side of HTA. On the other hand it, by the means of these very projects as well as by additional demonstrations of political will, encouraged the creation of networks of individuals and institutions active on the topic both at the EU and member state level.

Policy

The active involvement of the EU in shaping European understanding of HTA is one of the key characteristics of the policy. The beginnings of the Commission’s activity in HTA came early: in 1986 with the initiative on Methodology of Economic Appraisal of Health Technology (Drummond, 1987), while HTA was still an emerging field. Since the mid 1990s, EC-funded projects have run almost continuously and the Commission also kept HTA on the political agenda, at level beyond technical expertise. The EU, and in particular the Commission, has been interested and actively supporting a certain (European-wide) vision of what HTA as a scientific discipline and a policymaking tool could look like. In fact, HTA in the member states and at the European level developed simultaneously and interdependently (Böhm and Landwehr, 2014). Many European countries set up national HTA bodies at about the time of the 2004 Communication, which first mentions HTA as a political priority (European Commission, 2004, p. 11), and of the beginnings of the EUnetHTA Project.

The various EU-funded projects created a specific understanding of HTA as a scientific and policy-making field. The EUnetHTA Project produced a definition of HTA, which is today being widely used by EU policymakers as well as European HTA practitioners and observers. It also generated the ‘HTA Core Model’, a trademarked cookbook-style framework providing guidelines on what an HTA evaluation should practically look like. The Core Model and other EUnetHTA recommendations serve today as a readily available template for countries interested in setting up HTA institutions.

Beyond content, a clearly stated objective of most EU HTA projects from 1994 on was to create networks of experts (Banta and Jonsson, 2009; Banta et al, 2009). Undoubtedly, the number of member states participating in EU initiatives has risen steadily, from 5 countries in 1994 to 16 in 2006 and all 28 member states plus a number of non-EU countries since 2012. Manifestly, no member states wish to be excluded from cooperation on HTA. Even countries with almost no HTA capacity nominate at least one formal contact person, often from the Ministry of Health. The empowerment, or effective creation, of networks of experts with the legitimacy to formulate desirable course of action is a key characteristic of the policy stream in the case of European HTA. Likewise, the diffusion of what the expert networks define as best practice promotes convergence on both what member state actors do in the field, and the extent to which member states do it at all.

Problem

The name ‘communicable disease control’ makes it clear what problem ECDC and similar policies purport to address. HTA, on the other hand, can be constructed as a solution to several problems: public finance, inequality of access, sustainability of health systems and so on. Since its beginnings, HTA presented itself as a solution to the problem of uncontrolled diffusion of costly new medical treatments putting a strain on health-care budgets (Drummond, 1987; Banta, 2003). The promise of HTA for policymakers is to limit inefficiencies, in other words to deny or limit funding to treatments with an unfavorable value-for-money ratio, even though some HTA practitioners and theorists are quick to point out that HTA should not be a rationing tool (Jonsson et al, 2002). There is little doubt that rising health-care costs are a pressing issue for most EU member states, and new drugs and technologies are regularly singled out as the largest contributor to expanding health-care budgets. At the EU level this seems to be the frame of choice of the HTA policy community: there is talk of balancing pressures on public spending, efficient use of resources or promoting cost-effective technology (Jonsson et al, 2002; Kristensen, 2012). HTA agencies can also be useful shields for politicians who would otherwise have to account for coverage decisions (Williams, 2015).

To confirm the basic premise of the multiple streams framework that policies can be presented as solutions to various problems, HTA has also been portrayed (although somewhat cautiously) as a response to the problem of unequal access to healthcare innovations for patients across the EU. On the eve of the Eastern Enlargement, the Commission juxtaposed HTA and inequalities of access in its 2004 Communication by mentioning its agenda to ‘ensure universal access to high-quality services’ and by using EU-level HTA initiatives to ‘help to ensure that patients throughout Europe benefit from care reflecting the latest advances in medical technology’ (European Commission, 2004, pp. 3, 11). Similarly, HTA has recently been presented in the context of improving ‘resilience’ of health systems (European Commission, 2014). The added value of European activity lies in information sharing and avoiding duplication of efforts, as member states’ HTA bodies often assess a technology at about the same time. In other words, the same arguments that have been used to develop a case for an EU network or agency in other fields, such as communicable diseases (or medicines, licensing, food safety and drug addiction) are at work; advocates of a stable EU organization point to the staffing and financial instabilities, duplication, rigidity and fragmentation of an approach based on project finance.

Politics

The landscape of HTA in the EU was even more varied and fragmented than that of communicable disease control. All our interviewees identified a clear division between member states with longer-standing tradition in HTA such as the United Kingdom, Germany or France, and countries with little or no experience with the topic, mainly Central and Eastern European and some Southern European ones (also mentioned in Ecorys, 2013). This division was often portrayed as big versus small member states, suggesting that HTA capacity is limited in smaller countries and that bigger agencies were reluctant to have a perceived problem of duplication solved by letting other countries free ride on their analyses. A leap to an agency seemed implausible, but the step to a network with a coordinating body was popular among the stakeholders identified in the European HTA world by an insider’s analysis (Banta, 2003).

The Commission, on the other hand, has shown initiative on similar issues in the past, just like with the ECDC. In the case of HTA, however, it decided not to test the limits of the member states’ willingness to let go of their competences in health policy in any way. In the mid-2000s, HTA became integrated with a number of other health-related issues in the first major piece of legislation of DG SANCO’s doing on EU health care after the Kohll/Decker judgments with which the (then-) European Court of Justice started to shape European health policy: the Cross-Border Healthcare Directive. The main topic of the directive – reimbursement of cross-border care – was highly controversial and the directive’s existence was marked by member states’ resistance from its very beginnings (Hervey, 2016; Martinsen and Mayoral Diaz-Asensio, 2016; Vasev et al, 2016; Vollaard, 2016).

The list of relatively benign issues loosely related to cross-border treatment grouped under the pillar on cooperation in health care (eventually Chapter IV of the directive), including provisions on HTA, therefore achieved a double purpose for the Commission.Footnote 2 First, it gave a legal basis to a number of DG SANCO’s often ad hoc projects and initiatives that could later be developed by secondary legislation, and to institutionalize them in some, albeit minimalistic, form. Second, it served to shift the focus of the directive from its most contentious points on reimbursement (as well as quality and safety standards) to a more comprehensive list of health-related proposals. The pillar on cooperation was supposed to make the directive more palatable to member states, who were almost unanimous in their opposition to its general direction; to paraphrase one interviewee from the EU institutions, this was the only way the Commission could see the directive pass. This also meant that member states directed their opposition to the points that mattered to them the most – reimbursement – and did not waste their energy on what they saw as minor issues of cooperation.

The Commission, on the other hand, also preferred to fight on a single front and decided to preempt any member state opposition to issues of lesser priority by not venturing too far in their proposal. It was aware of member states’ concern that increased EU integration in HTA could lead to ‘Brussels dictating to [national governments] what to reimburse’ and repeatedly voiced reassurance that it would not interfere with national competences – to the point of adding a guarantee in this sense verbatim in the final text of the directive. It decided not to push the question of what institutional shape to give to the new HTA body and opted for a less menacing ‘voluntary network connecting national authorities or bodies responsible for HTA designated by the Member States’ rather than an independent agency of any kind, although there have been reports of various actors from EU institutions admitting that ideally the HTA network would be an agency, as well as vague hopes that, to paraphrase several interviewees, ‘all these things start as networking exercises’. This suggests that the idea of creating a European HTA agency was being considered by the Commission and discarded as an unviable option. A similar logic was likely also at play when it came to implementation of the directive’s provisions: the instructions of DG SANCO (formally of the Executive Agency for Health and Consumers) to Ecorys regarding their cost-benefit assessment of various implementation options of the new HTA Network included a possibility for the network to be hosted in an existing agency subordinate to DG SANCO (Ecorys, 2013). The report did not identify a new or existing agency as the preferred option of stakeholders involved in HTA; however, the fact that the option of merging the network with EMA or creating a new agency was considered is in itself significant, and serves as an illustration of how EU-level solutions are routinely part of the agenda in the politics stream.

As a result, the alternative on the agenda when the presentation of the Cross-Border Health Care Directive presented a focusing event was a network, and that was what was both proposed and incorporated. The further evolution of the three streams suggests that an agency might appear on the agenda again soon, especially given the prominence of HTA in the European Semester and ECOFIN Council’s interpretation of health-care system efficiency, which echoes the Commission’s main justifications for HTA.

Conclusion

We began with a puzzle: despite the development of a constraining dissensus in EU politics, an EU health policy, including HTA and communicable disease control, has developed. This development of an EU health-care union, discussed in the Introduction to the special issue, is particularly unexpected since even in the days of the permissive consensus, member states often opposed the idea of EU health policy. The Lisbon Treaty’s health article specified clearly that the organization and finance of health-care services was outside the EU’s competencies while its work in public health was to be supplementary to member state work. Nonetheless, EU health policies alongside different integration tracks developed over time (Vollaard and Martinsen, 2016), justified by a variety of treaty bases, especially those of internal market and fiscal governance as well as public health. In some cases, integration occurs even in unlikely areas such as HTA. In other words, the constraining dissensus Hooghe and Marks identify did not constrain the development of an EU health policy, as discussed throughout this issue.

The juxtaposition of EU health policies discussed here and the electoral research focused on Hooghe and Marks’ theory suggests that neofunctional mechanisms continue to work – that even if there is no broadly based agreement on the desirability of new EU initiatives, the mechanisms that connect policy and political spillover to agendas and alternatives continue to work in at least some cases.

The differences that exist between the two cases show some of the importance of these different streams. In particular, the intellectual, often Commission-financed, support for both HTA and European communicable disease policy dates to at least the 1990s. But the political and problem streams both differ. The problem stream of communicable disease is much more dramatic, with events such as anthrax and SARS to focus attention. By contrast, weak HTA is difficult to identify or turn into a crisis. That is why HTA only received legislative attention when the broader development of the Directive on patients’ rights in cross border health care created a focusing event. Likewise, the politics stream is somewhat different. Not only was the creation of ECDC an opportunity for a weak Commissioner to assert himself, it also was an opportunity for ministers, MEPs and EU institutions to gain credit for acting on behalf of Europe’s shared security. Both the problem and political stream, then, would predict fewer focusing events and less activity no matter how Europeanized and vibrant the policy stream. Similarly, though, it makes it rather impressive how successful HTA has actually been, given its technical obscurity and unpopularity with important industries.

Others, facing the same problem, have tried to differentiate policy types, arguing that some kinds of policies have effects and public faces that make them more or less tolerable to voters in a European dissensus. The problem is that the construction of policies is always in itself political, as are alternatives and agendas. Starting with the multiple streams theory of politics rather than the characteristics of the policy allowed us to avoid this policy essentialism and track the ways in which pro-integration biases worked themselves into the three streams. The EU substantially financed and supported a relatively pro-integration, European policy stream, which contributed to perceptions of problems as being shared European ones, and the politics stream remains populated by politicians, such as Commissioners, MEPs, rotating Presidency holders and many ministers who want to make an impression by taking action at the EU level.

On the level where policies are made – and in this case, European integration furthered – we find that the policy, politics and problem streams are all canalized if not created by EU action of the sort Haas discussed. So long as problems are framed and set on the agenda in this way, offering credit for all concerned, they will continue to be addressed with EU policies. There might be arguments about what exactly the EU and its agencies should do, but such arguments about desirable European policy are premised on the European integration having happened. That suggests that ordinary EU policymaking still has powerful integrative mechanisms built in that can defy broader constraining trends. For all the constraint dissensus might cause at the level of intergovernmental negotiations and electoral politics, it is a permissive one at the level of ministers discussing a specific proposal in Council. This disconnection between evidence of spillover and evidence of resistance to integration is not new in discussions of neofunctionalism and European integration. Moravcsik, for example, was concerned to prove that the Commission was unable to set the agenda or control the alternatives for major intergovernmental conferences and treaties (Moravcsik, 1998). What this case might suggest is that neofunctionalism is working well to explain integration through agenda-setting and alternative specification, in processes as low salience as HTA and as high salience as rights to health care and fiscal governance. A constraining dissensus need not necessarily materialize, as in the case of economic surveillance described by Baeten and Vanhercke (2016) in this issue. If it is at work, is at work in the decision stage.

If we are indeed in the era of a permissive dissensus – combining contentious electoral politics about European integration with ongoing neofunctionalist European integration – then the challenges of the EU can be seen in a new light. Persisting neofunctional integration amidst increasing dissensus and partisanship about European integration might be a combination that stores up trouble for the EU later.