An Actor-Network Theory Analysis of Policy Innovation for Smoke-Free Places: Understanding Change in Complex Systems

Abstract

Complex, transnational issues like the tobacco epidemic are major challenges that defy analysis and management by conventional methods, as are other public health issues, such as those associated with global food distribution and climate change. We examined the evolution of indoor smoke-free regulations, a tobacco control policy innovation, and identified the key attributes of those jurisdictions that successfully pursued this innovation and those that to date have not. In doing so, we employed the actor-network theory, a comprehensive framework for the analysis of fundamental system change. Through our analysis, we identified approaches to help overcome some systemic barriers to the solution of the tobacco problem and comment on other complex transnational problems.

Understanding complex policy issues requires a socioecological approach,¹ which we used to describe a tobacco-use management system² that not only determined tobacco use but also served other purposes (e.g., profit for companies involved). The model proved useful for defining key problems facing tobacco use management and for making recommendations for the future evolution of the management system. However, systems thinking says little about the processes involved in producing system change.

We analyzed the evolution of smoke-free places, using the actor-network theory (ANT) to explicate the processes involved in producing significant system change. ANT^3–5 offers a comprehensive model of the innovation process, be it for producing new tobacco control policies or other sociotechnical changes. However, before we describe ANT in detail, it is important to understand the context within which ANT is applied.

Systems change their structures and processes (how they do what they do) within limits defined by their response repertoire. This repertoire becomes more limited as the system optimizes its capacity around responses that have previously been successful.^{6(p205–207)} Consequently, the potential for structural or procedural change becomes constrained over time. In addition, systems are typically unable to change their functions (what they do) from within, especially when their activities take place within a context characterized by polarization and stalemate,^7,8 such as the tobacco control domain.^2,9 Fundamental, innovative change requires action outside existing system structures. External actors include internal people (or groups) acting outside their system roles and external people or groups who, while varying in the intensity of their relationship with the system, believe it is not delivering. Management increasingly recognizes advantages in incorporating the views and actions of external actors when analyzing problems or redesigning systems.^10–13 These actors are not structured into any system, but act as a functionally coordinated network with respect to a potential change.^14–16

Policy innovation involves at least 4 stages to create a reformed system: establishing the issue or problem (autonomization in ANT terms), developing possible solutions, contesting possible solutions, and institutionalizing, implementing, and refining the preferred solution. Most extant theories of change or innovation focus solely on 1, or at most 2, of these stages. For example, systems thinking^17–20 focuses on problem identification (but not how it becomes autonomized), and on potential solutions, whereas (social) diffusion theories,^21,22 knowledge management theories,^23,24 and approaches derived from complexity theory²⁵ focus on the last 2 stages.²⁶ ANT is a broader umbrella theory that can explicate all 4 stages; it encompasses Westfall et al.'s²⁷ extended US National Institutes of Health (NIH) categories—translation 1 (basic science to trials), 2 (trials to practice) and 3 (practice-based refinement)—and uses concepts and strategies associated with the more focused theories in doing so.

At each stage of an innovation process, the resolution is not completely determined by evidence and internal coherence, but is codetermined by the capacity of the parties to keep their arguments alive, which involves persistence and power. This capacity is influenced by the different framings^28,29 competing networks use (e.g., protecting health vs protecting choice), leading to the development of different narratives or stories. The evidence, the theories that make sense of the evidence, plus the implications and possible outcomes, are woven together by the protagonists into a discourse intended to persuade others to their point of view. As explained in the following sections, ANT provides a theoretical framework for understanding, and a guide to the prerequisites for, successful change in complex systems; in this case, through policy innovation.

ACTOR-NETWORK THEORY

ANT focuses on science-based innovation processes, and, in analyzing indoor smoke-free regulations, the process necessitates engagement with all the component subsystems (tobacco control, tobacco industry, regulatory, and tobacco use) of the tobacco use management system² as well as relevant features of their context that together determine tobacco use. ANT argues that fundamental system change requires that actors and resources are “translated”^{5(p106–109),30} away from their formal subsystem roles and that a network grows as they join with others, outside the system, to pursue the innovation. Such changes also involve partial engagement (“weak links”³¹) with an even wider range of external actors. Unlike other approaches, ANT maps the full innovation process from genesis to implementation, focusing on the relationship between the scientific evidence and the elaboration of networks employing the evidence.

Actor networks are heterogeneous.⁴ Unlike social networks^16,22 associated with diffusion theory, they are networks of people, research evidence, technologies, financial resources, institutions, and regulation acting together to produce the innovation. The development of an actor network necessarily signals the pursuit of a goal (e.g., banning smoking in indoor public spaces), which generally stimulates the development of counter networks, or “anti programs,”^32,33 with different goals (e.g., dedicated smoking rooms with hi-tech ventilation) or that seek to block the goal being sought. They compete for allies, institutional endorsement, public support, and tangible and intangible resources.

ANT describes a process in which heterogeneous elements are woven together and assembled into reality. This move beyond “social networks” and “social diffusion” represents a fundamental ontological contribution by ANT. The ANT framework is schematized in Figure 1 (after Latour⁴). The central core, or “links and knots,” represents the actors (knots) and the set of relationships between them (links) that first initiate the innovation and then act to organize, and hold in place, all the activities along the 4 “loops” that reach out into the broader context. In evidence-based innovation, the loops represent activities designed to link the core with 4 interdependent sources of power: scientific evidence and the technologies or techniques that depend on it, professional or political authority, allies, and public opinion.

FIGURE 1.

The key processes in an actor-network theory analysis.

Source. Latour's model^4(p99–111) was used as a starting point.

In Latour's original framework, the upper loop in Figure 1 is “mobilization of the world”; the embryonic actor network must link with, and mobilize, a range of nonhuman resources (e.g., intangible resources like scientific evidence and supporting theories; tangible resources like technologies, money, and venues) to produce an innovation.⁴ Recognizing the centrality of language, Latour asserts that mobilization represents “all the means by which non-humans are progressively loaded into discourse.”^{4 (p90)}

In the case of smoke-free policies and similar cases (such as climate change), the key resources are knowledge about the problem (particularly scientific evidence), the theories or ideas that make sense of the evidence, and the technologies or processes used to assess or solve the problem. The evidence and the ideas that are mobilized ultimately reflect the interests of those engaged by an issue, and are mobilized to serve their values and priorities. Mobilization becomes a dynamic interplay between evidence and arguments, constraining which arguments come to dominate the discourse and, consequently, how actions or solutions are framed.

The bottom loop, “alliance building,” is the process by which a putative innovation garners allies who can tie it into extended networks (and their resources) while rejecting those likely to detract from its credibility and power, or who support competing outcomes. All innovation involves contests over alliance partners and, as some are added and others avoided, they contribute to the developing discourse about the innovation.

The “balance of opinion” loop is about gaining the acceptance of individuals potentially affected; first, that a problem or issue exists, and second, that the putative innovation represents the best way of managing it. Policy innovations generally require some level of public acceptance before governments will adopt them, so the contestants often expend considerable resources trying to market their positions to the public. The public also provides a source of recruits for existing alliances, and may seed new “grass roots” activist groups that organize to service interests that are currently not represented.

The final loop, “autonomization” or “institutionalization,” refers to the process by which authoritative acceptance of, and institutional support for, an innovation emerges from the appropriate structures (e.g., professional structures for validating specific concepts and government structures for institutionalizing policies). Setting up inquiries, calling for reports, or, later, establishing issue-specific units (e.g., offices of tobacco control), are typical examples of the “autonomization of an issue” as important and distinct. Such actions establish autonomy by distinguishing the issue from others competing for institutional support. Implementing regulations and programs reflects “institutionalization of solutions.”

INNOVATION IN PRACTICE

The contest to establish the harmful effects of secondhand smoke (SHS; also called passive smoking), and the subsequent campaign for smoke-free places, illustrates how policy innovation works in complex systems. Because it has been successfully achieved in some countries but not in others, it is possible to identify what worked well and what created barriers. The process of institutionalizing smoke-free policies has been broken into 3 overlapping stages: autonomization of the issue, contest to determine the solution, and convergence on solutions (resolution).

Autonomization of the Issue

Establishing SHS as a problem requiring solutions represents the first stage in the process of institutionalizing smoke-free policies (Figure 2). Few saw SHS as a problem until epidemiological studies demonstrated adverse health effects, so mobilizing evidence was critical to establishing the issue (as was the case with climate change), in contrast to issues where the science merely supports or amplifies existing concerns (such as alcohol and road accidents).

FIGURE 2.

Issue autonomization (stage 1) of an actor-network theory analysis of policy innovation for smoke-free places.

Note. NGO = nongovernmental organization; SHS = secondhand smoke; TC = tobacco control.

Stage 1 began with Hirayama's³⁴ demonstration that exposure to SHS increased lung cancer mortality in nonsmoking wives of smokers, although there had been prior evidence of nonfatal effects (largely on children). Several official US,³⁵ Australian,³⁶ and UK³⁷ reports followed, all finding that SHS causes lung cancer in exposed nonsmokers as well as nonfatal respiratory effects, especially in children and babies. More studies followed, extending known effects to heart disease.^38,39

In some countries, tobacco control networks had already achieved influence tackling the direct harms of smoking. They represented, or were supported by, organizations concerned with cancer, heart disease, and respiratory diseases, as well as groups with more general medical and public health interests. The networks saw the issue of SHS, which harms nonsmokers, as a powerful tool for controlling tobacco use, so they pursued it vigorously and grew by feeding off it. Alliances were forged with occupational health and safety groups, infant health interests, and, particularly in the United States, environmental groups and authorities.⁴⁰ Nonsmokers' groups (e.g., Americans for Nonsmokers' Rights) were also established, either with the help of existing tobacco control groups or spontaneously.

People's preparedness to support positions for or against smoke-free regulations was important. The SHS debate involved engagement with 3 overlapping “publics”: nonsmokers who had to breathe smoke, smokers who had to refrain from smoking, and citizens. Smoke-free rules involved regulation, an intervention in the economy, and the social life of the country. How citizens responded—through opinion polls, talk-back radio, the letters pages of newspapers, and voting—was critical.

Initially, all these publics were relatively passive; smokers did not try to change their behavior, and others did not challenge them to do so. However, publicity of the (now more newsworthy) research evidence in the mass media, and stories about affected individuals, led to increased awareness of the problem. Media activity was facilitated by tobacco control groups and complemented by paid advertising. The tobacco industry reacted by challenging both the evidence and the proposed solution of smoke-free places.

Not all studies showed robust effects independently (although meta-analyses were unequivocal), and these apparent failures to confirm effects were exploited to try to deny any effects.⁴¹ However, more concern materialized as evidence increased. The evidence, coupled with the framing of the issue as one of environmental pollution, led to SHS being formally declared an environmental harm by the US⁴² and California⁴³ environmental agencies. The scientific consensus took time to disseminate, especially to non–English-speaking countries. Where evidence was accepted, and opposition was limited, some organizations (including government agencies) applied existing occupational health and safety regulations to make workplaces smoke-free.⁴⁴ Governments also supported public education initiatives to highlight the risk, helping to autonomize SHS, build public understanding, and increase concern.

The Contest and the Negotiation of Solutions

After SHS became an autonomous issue, significant debate emerged about solutions and about whether federal, state, or local governments, venues themselves, or individuals should act. This contest over solutions constitutes the second stage (Figure 3), and responses ranged from doing nothing to completely banning smoking. As neither extreme was acceptable, options and debate focused on a range of indoor smoking restrictions.

FIGURE 3.

Contest (stage 2) of an actor-network theory analysis of policy innovation for smoke-free places.

Note. SFP = smoke-free places; SHS = secondhand smoke.

Indoor smoking bans represented a challenge for venue controllers and threatened the tobacco industry. The industry feared that smoking's ubiquitous social role, especially its association with eating and drinking, was threatened by smoke-free rules. Research confirmed some of their fears: restrictions on workplace indoor smoking did reduce cigarette consumption and helped denormalize smoking.^45,46

The industry strongly contested the evidence, and their challenges cast sufficient doubt that some governments decided remedial action was premature.⁴¹ Industry offered 3 alternative solutions: first, “mutual respect” (for smokers' right to smoke and for nonsmokers' right for clean air); second, improved ventilation; and third—failing ventilation—separate smoking areas. With little evidence of smokers voluntarily refraining from smoking around nonsmokers, regulatory interventions became the only effective options.

Tobacco companies were not directly involved in possible solutions, so they had no place in negotiations. To attract allies who controlled public spaces, they argued, without evidence, that smoke-free rules would cause lost patronage and economic hardship, especially in recreational settings (e.g., restaurants, bars). These claims brought some leisure industries into alliance, increasing government concern over possible economic effects and leading to more caution.

Tobacco control coalitions were always attracted to solutions with the greatest capacity to reduce active smoking; their goal was achieving consensus around bans in all enclosed places, and not having to confront the industry's power directly was a clear strategic advantage. Their proposals were taken up by some governments; negotiations commenced with groups whose primary interests were not diametrically opposed to their own and who would likely accept evidence indicating their interests were not under threat.

Debate was continually fueled by studies of early initiatives. Evaluation studies revealed the following: most organizations complied with smoke-free policies⁴⁴; such policies reduced exposure to cigarette smoke^47,48; support was increasing for such policies, even among smokers⁴⁶; and smoke-free policies reduced cigarette consumption, but their effect on smoking cessation was equivocal.⁴⁶ Ventilation studies showed that at levels acceptable for comfort, removal of smoke was inadequate to meet standards^49,50 and that ventilation was commercially viable only in venues where the costs of banning smoking were high (e.g., casinos).

Even when solutions emerge, implementation mechanisms must exist. The most comprehensive potential mechanism was environmental law. In the United States, particularly in California, an extremely broad strategy emerged, engaging the Clean Air Movement and community activists (concerned about the industry targeting disadvantaged groups)⁴⁰ as well as health and allied professionals. This alliance supported a successful citizen-initiated referendum (Proposition 99) to tax tobacco products, which they hypothesized would help control tobacco consumption and also guarantee a significant budget for smoke-free campaigns and resources to extend their networks.⁵¹ Subsequent activism led to several progressive communities introducing citywide smoke-free policies.⁵² These were accepted with no clear downsides,⁵² which helped California to become the first large jurisdiction where public indoor places were effectively smoke-free.

Australia provides a contrasting example. There, the focus was on occupational health and safety (OHS) rather than environmental laws, and this meant that unions and employer groups needed to be enrolled as allies. Although OHS laws applied only to workplaces, they provided a politically expedient mechanism for introducing smoke-free conditions, since negotiations largely took place within the health suprasystem² and activated fewer competing interests. Many of the arguments used to justify environmental laws also applied to OHS laws, although some aspects lost potency (e.g., protecting babies). Most workplaces went smoke-free without the need for specific rules because generic OHS provisions, coupled with some high-profile compensation payouts, sufficed. However, the recreation sector was slow to move. Australian local governments lacked the powers of their US counterparts, so initiatives needed to come from larger, state-level jurisdictions with responsibility for OHS. States that acted early implemented selective restrictions and gradually escalated to comprehensive smoke-free policies, whereas states that acted later took larger and fewer steps.

Many governments were drawn into debates about possible exemptions (e.g., small pubs, casinos) and about smoking and nonsmoking rooms. Where it was mandated that one room be smoke-free (effectively banning smoking in single-room venues), stalemate sometimes ensued through constitutional challenges over fairness (e.g., in Germany).^53,54 Paradoxically, some interim policies that failed the policy intent, such as mandating smoking and nonsmoking sections within one space (e.g., in some Australian states), may have facilitated total bans because smoke spread throughout, rendering this approach socially unacceptable.

During this stage, tobacco control networks continued engaging the public to denormalize public smoking and normalize smoke-free conditions. Public views about smoking gradually translated into new norms for smoking, and some people, including smokers, began to implement smoke-free homes.⁵⁵ Improved knowledge reinforced the need for action. The public's experience that all the preliminary steps had been positive (or less negative than opponents had predicted) led to a positive feedback cycle of increasing support for smoke-free environments.⁵⁶ This process took 2 decades, a relatively short time for a contested reform process, especially compared with progress in resolving the broader tobacco problem.

Resolution

During the third stage, resolution (Figure 4), supporters of smoke-free places came to dominate the discourse, convincing governments to enact policies ending the contest and embedding reform.

FIGURE 4.

Resolution (stage 3) of an actor-network theory analysis of policy innovation for smoke-free places.

The move to resolution strengthened as people increasingly saw public smoking as contrary to cultural norms. Studies of the economic effects of smoke-free restaurants in early adopter jurisdictions showed no evidence of net adverse effects; some suggested net benefits.^57,58 Consequently, restaurateurs severed ties with the industry, and some introduced voluntary restrictions prior to regulation. Older-style, “booze only” pubs (already declining) and gambling venues did show some adverse effects,⁵⁹ but it was argued that bans served an additional good (reducing gambling losses). In short, the evidence strengthened the case for smoke-free solutions and weakened, or marginalized, alternatives. Action was also made easier by the widespread belief that SHS harms “the innocent” (children, bar and restaurant staff, and so on) and, in some countries, by the declining proportion of the population who smoked. Governments, aware of the evidence and public opinion, saw little reason for inaction and moved toward implementation. In ANT parlance, actor networks had achieved “convergence”³¹ on smoke-free enclosed places as the solution.

Ireland, the first country to go completely smoke-free, illustrates several key resolution issues. Relying on global scientific literature to make the case,⁶⁰ Ireland moved from almost no restrictions to becoming totally smoke-free in one step, exemplifying how science facilitates generalization of facts and arguments and enables “action at a distance.” Not surprisingly, there was resistance to mobilizing science conducted in other places, in some cases because of cultural or structural barriers (e.g., in Germany⁶¹) and in others simply as a strategy for delaying action.

Predicting which jurisdictions will be first to act is difficult. Ireland acted largely through a convergence of interests between a senior government minister (Michael Martin) who recognized the problem, a bureaucracy (Office of Tobacco Control) primed to confront it, and a coalition of health groups (Tobacco Control Alliance) lobbying for action and prepared to work with government. The easiest path was building on OHS infrastructure. The Office of Tobacco Control worked with the Tobacco Control Alliance to win public support. Judicious use of the published literature and access to the powerful international tobacco control communications tool Globalink⁶² enabled the Irish campaign to mobilize evidence at both strategic (e.g., effects of SHS on bar workers) and tactical levels. Tactically, they could counter industry's arguments within hours. For example, when economic discourse was invoked to decry smoke-free rules, “we lodged a post on Globalink at 10 am, and by 3 pm we had a reply” (Fenton Howell, president of the Royal Academy of Medicine in Ireland, oral communication, March 2007). This ability to load evidence into discourse, in real time, was crucial to success.

In Ireland, an atmosphere of inevitability and movement was consciously promoted, which led to people wanting to know more and whether they should “get on board.” This effort involved strategic use of mass media, with paid advertising priming the issue, and news coverage guaranteed by the advertising. This in turn attracted new allies, whereas potential opponents were neutralized.⁶⁰ Others now use evidence of Ireland's success to argue for smoke-free places in their jurisdictions and to dampen fears of adverse effects.⁵⁶

In the United States and Canada, county, city, and community level governments were the chief protagonists of smoke-free regulations.⁶³ The central government was seen as too inflexible and possibly overly sensitive to industry lobbying. Whereas there was public suspicion of “big government,” community-based initiatives had a long and respected history, and the tobacco industry was seen as an “outsider” at the local level. However, this approach can leave rural, low-education “pockets” behind (those without a critical mass of activists, or who lack the tax base to introduce effective compliance regimes),⁶⁴ so more broadly based rules (e.g., at the state or national level) are eventually required to set minimum standards.

In Australia, Ireland, and some other countries, smoking bans have been largely self-enforcing once implemented. Compliance has been high, partly because smokers understand and respect the rationale and partly because the public and key stakeholders (e.g., restaurateurs) have been exposed to sophisticated education campaigns.⁴⁶ In countries like France, however, which needed to enact legislation (as France did in 1992) before it could introduce government education campaigns, a second set of laws (in 2007) was required before there was high compliance.⁶⁵

Evidence from countries that have achieved convergence and successfully resolved the SHS issue reveals that alliance building requires a sophisticated understanding of local cultural norms. In Australia and Ireland, where trade unions and employer groups are credible actors, they became important allies and helped frame the issue. In California, the focus was more on environmental allies because they had the requisite credibility. Consistent with the law of increasing returns,^{66 (p1–12)} successful alliance building creates strong positive feedback for the innovation, and positive feedback “locks in” a single dominant product, technology, or policy; the more connected the network, the more dominant it becomes. In California, solid, well-communicated science reinforced attempts to autonomize SHS and institutionalize the smoke-free solution, making the formation of cross-sectoral alliances more likely (while weakening industry alliances). This enabled powerful public campaigns to gradually denormalize public smoking and provided support for initial (successful) steps by innovative communities, laying the foundations for further action, which was then confirmed by more research, escalating the introduction of smoke-free regulations (a “virtuous cycle”).

Successful institutionalization is instructive, but at least as much can be learned from failures. In Germany, attempts to introduce relatively modest federal smoke-free rules were undermined when several states exercised their constitutional powers and refused to comply. The federal government had to compromise and leave it to individual states to determine both the level and timing of smoke-free conditions. Implementation was less than comprehensive.⁵³ In addition, German culture stresses individual responsibility for health, primarily because Germans are wary of public health interventions, and has a corporatist political system that directly involves powerful interests (including the tobacco industry) in formal decision-making processes.⁶¹Germany also attempted to implement rules without due consideration of equity, such as trying to legislate that all single-room venues be smoke-free while allowing multiroom venues to have designated smoking rooms, a strategy that attracted successful litigation.⁵⁴ Further, the German scientific and political establishment perceives itself as unique and expects local evidence before acting.⁶¹ In ANT terms, the tobacco control community in Germany was “regional,” not “networked,” and thus unable to generalize from experiences elsewhere.

Adoption of the Framework Convention on Tobacco Control Article 8,⁶⁷ which promotes smoke-free rules at a global level, produced hope that the policies would be gradually taken up worldwide. However, local cultural and institutional issues can intervene. In jurisdictions where the arguments have not been properly examined or the detail of the strategies analyzed, there is still potential for oppositional forces to maintain influence. For example, Japan may allow exemptions to smoke-free rules similar to those of Germany, thus reducing the effectiveness of their strategy.⁶⁸

INSIGHTS INTO EACH LOOP

The following sections summarize the contribution of each loop to the innovation process.

Mobilization

Competing priorities for mobilization emerge out of the different values and goals of the competing networks (e.g., tobacco control vs the industry), and this leads to the development of different narratives. These competing stories are tightly woven webs of evidence, ideas, theories, and values. Their relative credibility is a function of their internal coherence, the perspective from which they view the problem (which determines which issues are seen as critical), the quality of the evidence embedded within them (indicating truth value), and their capacity to engage potential audiences. Stories offering resolution (e.g., smoke-free conditions do significantly improve air quality⁴⁸) are especially potent.

There are 2 classes of scientific research mobilized in support of the stories: evolving research relevant to the issue that occurs independent of action (e.g., epidemiology demonstrating that SHS is harmful) and research directed at understanding the problem (once identified) and developing and evaluating (potential) solutions. The net contribution of science to the management of complex problems like tobacco control is a function of the following:

1. Certainty (“immutability” in ANT terms). How definitive is the evidence? Are there plausible alternative explanations of the facts? Are there plausible alternative solutions that have not been adequately evaluated? Answering these questions is a top priority for contested innovations.

2. Generalizability (“mobility” in ANT terms). Does the evidence apply here, or are “we” different? Context significantly influences both the amount and the focus of the evidence that is mobilized. Where local interpretations of the issue dominate, general solutions may not be pursued, or pursued only after local validation. Similarly, jurisdictional powers may constrain which evidence is mobilized.

Alliance Building

There is no “best practice” in alliance formation over and above understanding both the intrinsic power of a given set of stakeholders and how trends in the broader environment, within which all systems are embedded, alter the balance of power within that set.

The groups that become centrally involved in an issue reflect how the issue has been framed and the nature of the specific solution sought. Interest groups are motivated by their perceptions of how the solution will affect their concerns. They can choose to support the goals of the innovation network, to negotiate modifications to the goals to make them more congruent with their own, or to oppose them. Alliance members jointly determine alliance goals, and as alliances grow, the diversity of individual goals grows, so there is greater potential for compromise over solutions. There is a dynamic tension between having an alliance sufficiently strong to achieve its members' goals and increasing its power through more members; the latter could lead to compromises that the original members might find less desirable.

Groups whose interests are secondary to the focus of the changes being negotiated are vulnerable, because they cannot easily become lead negotiators, and their interests may be compromised by focal groups whose key interests are different, even if marginally so. If the perceived impacts on the interests of specific parties (e.g., restaurateurs) can be changed by arguments or evidence, then the allegiance of those parties is likely to shift, changing the balance of power.

Smoke-free rules actually represent one “battle” in a longer-term “war” to minimize tobacco use. Strategically, it represented an “indirect attack”; attacking on a weak flank, where one is cohesive or powerful, and on the ground of one's own choosing. In this case, controllers of indoor public spaces were the weak flank, and the movement was cohesive and powerful because they achieved consensus that smoke-free was the solution (it lacked the potential to open up tensions between harm reduction and zero-tolerance advocates, a tension as intense within tobacco control as in the broader drug control movement⁶⁹), and “innocent” nonsmokers provided a “ground” of choice. This strategy enabled the movement to extend smoking's denormalization without ever engaging the industry head-on.

Engagement With the Public

If the putative solution is not obviously and immediately a plus for all, it has to be justified on a balance of interests (e.g., “the right to clean air” vs “the right to smoke”). The views of affected parties are critical, as are the concerns of the general public about the situations of affected parties. The public is concerned about effects on themselves and, in so far as it affects others, the perceived equity of the solution. Where public support translates into active support for coalitions (including joining them) or taking other actions (e.g., threatening to vote against a government), it will have a much stronger impact on decisionmakers than would passive compliance.

Autonomization and Institutionalization

Having a solution that is defensible in the face of opposition is critical. Success requires targeting the appropriate autonomizing or institutionalizing agents. Achieving this depends, in turn, on understanding local cultural, professional, and political norms, and the wrong institutionalization strategy can produce stalemate. Simple innovations are easier to sell and institutionalize than are more complex ones. Institutions seek the smallest possible change to minimize disruption, so they tend to do the minimum believed necessary to achieve agreed goals.

Accepting an innovation that contains compromises that can corrupt its coherence is a risk: it may not work or it may lead to unintended complications. Alternatively, it may be necessary to gain sufficient support for an innovation to be accepted and to facilitate incremental progress toward more substantial change.

Decisionmakers, especially in government, are more likely to autonomize problems, and institutionalize solutions, when they can see a critical mass of stakeholders being engaged by a specific solution and when alternatives are marginalized.

The use of ANT to understand the process of health policy innovation builds on more focused work by others. In particular, Newton³¹ has applied ANT to the climate change issue. In the tobacco control field, we have used Grüning et al.'s⁶¹ detailed analysis of the cultural and political dynamics that characterized the smoke-free contest in Germany, and Breton et al.⁶³ have articulated the alliance-building process associated with the Quebec Tobacco Act. The latter expertly mapped interactions between longer-term trends and political changes that underpinned the successful alliance-building process.

BROADER IMPLICATIONS

This application of ANT has implications for other complex public health problems, including obesity and climate change. Like SHS, they are difficult to solve, in part because optimal solutions for one set of stakeholders threaten, or may appear to threaten, the interests of others. SHS has a comparatively limited range of stakeholders and thus is easier to understand and solve than are the other problems.

Problems like tobacco control and managing climate change share the need to engage external actors in problem formulation and solution development. No single subsystem, whether government or any other, can solve these problems from within. In both cases, there are stakeholders with sufficient power to stalemate solutions that can be framed as threatening their interests. Governments will increasingly be faced with situations in which the achievement of public goals depends on their ability to constructively engage with external actors. Governments can produce engagement in 3 ways: mandatory requirements, financial or other incentives, or collaboration. In situations in which it is easy to achieve stalemate because of complex interdependencies between finely balanced, powerful coalitions of interest, inducing or forcing actors to engage will not suffice. A collaborative, deliberative approach,^10–12 especially where the mobilization of science is critical to decisionmaking,¹³ produces better outcomes.

In such approaches, widely adopted in Scandinavia and recommended by the US National Research Council,¹³ external actors (e.g., firms, nongovernmental organizations) collaborate because it advances their interests as well as the public interest. We argue that the most effective solutions to complex problems emerge where collaboration facilitates strong actor-network development. More effective regulatory frameworks can be developed by institutionalizing this collaborative imperative, especially where scientific discourse is critical to decisionmaking.¹³

Employing collaborative, deliberative processes requires one caveat. All actors who can envisage a non–zero-sum solution need to be brought to the table, but actors whose goals are intrinsically antithetical to a problem's definition (let alone resolution) should be excluded. In tobacco control, regulation of tobacco marketing should exclude the industry. They necessarily resist marketing constraints but cannot do so openly as participants, so they may dissimulate, and this could corrupt the process.

However, in the case of product regulation to reduce harmfulness, a partial convergence of interests could provide the basis for progress.⁷⁰ The importance of key players having options allowing them to participate is illustrated by the success of the Montreal Protocol negotiations.⁷¹ A key determinant was that patent holders of chlorofluorocarbons (CFCs) causing ozone depletion also held patents on hydrofluorocarbons (HFCs) that were to be employed as a solution. Thus, they could compromise without major loss.

ANT is not a predictive tool; it is a descriptive tool that takes into account the many varied elements contributing to innovation and facilitates understanding of the reasons for success or failure in different contexts. It is also an extremely powerful planning tool, identifying the links necessary for policy innovation to occur. Applying the “links and knots” model, following a systems analysis, enables planners to move seamlessly from “what” to “how.”

There have been some attempts to employ quantitative techniques to model actor networks. These include early work on “socio-technical graphs,”⁷² and more recently the development of the analytic technique known as Reseau-lu,⁷³ that is specifically designed to model heterogeneous networks. It has been applied in allied fields, including climate change⁷⁴ and biomedicine,⁷⁵ as well as the tobacco control domain.⁷⁶ Agent-based modeling (neural networks, cellular automata, genetic algorithms, an so on)⁷⁷ has also been proposed as an approach to modeling the evolution of actor networks.⁷⁸

The strength of the approach presented here is that, informed by a socioecological model,¹ the more focused actor-network model articulates the full process of innovation. Without this, furrows imprinted on the social landscape by earlier successes and failures fundamentally constrain creative innovation. ANT helps avoid the limitations of “best practice” summaries emerging from single jurisdictions, or the difficulty of trying to reconcile differences between them. Rational change does not emerge spontaneously. It needs to be framed from a strategic perspective, encompassing all identifiable influences (social–ecology framework) and enacted via sociopolitical processes (facilitated by ANT analysis).