Unpacking complexity in public health interventions with the Actor–Network Theory

Unpacking complexity in public health interventions with the Actor–Network Theory Abstract This article proposes a sociologically informed theoretical and methodological framework to address the complexity of public health interventions (PHI). It first proposes three arguments in favour of using the Actor–Network Theory (ANT) for the framework. ANT: (1) deals with systems made of human and non-human entities and proposes a relational view of action; (2) provides an understanding of the intervention–context interactions and (3) is a tool for opening the intervention’s black box. Three principles derived from ANT addressing theoretical problems with conceptualisation of PHI as complex systems are proposed: (1) to focus on the process of connecting the network entities instead of their stabilised form; (2) both human and non-human entities composing networks have performative capacities and (3) network and intervention shape one another. Three methodological guidelines are further derived: (1) the researcher’s task consists in documenting the events that transform the network and intervention; (2) events must be ordered chronologically to represent the intervention’s evolution and (3) a broad range of data is needed to capture complex interventions’ evolution. Using ANT as a guide, this paper helps reconcile technicist and social views of PHI and provides a mean to integrate process and effect studies of interventions. Actor–Network theory, complex system perspective, public health intervention research Introduction Actor–network theory (ANT) has been developed by Bruno Latour, Michel Callon and others since the 1980s. Originally, ANT was developed as a sociology of science focussing on understanding the production of scientific facts and technology. Because this theory developed a renewed vision of the social world, it evolved into a grand sociological theory known by the name of sociology of translation or ANT. Latour (2005) proposed an in-depth presentation of this theory in his classic book Reassembling the social. ANT is now recognised as a useful conceptual tool for appraising complex situations and analysing the production of change. The theory is increasingly used to understand change in complex systems such as health care, where it has been employed to study the introduction of technology or of new strategies for service improvement (Ramiller, 2007; Cresswell et al., 2010, 2011; Bleakley, 2012; Stoopendaal and Bal, 2013). It offers a powerful tool for opening up the black box of public health interventions as it provides a means for mapping the genesis of interventions in the form of a networking process (Bisset and Potvin, 2007; Timpka et al., 2009; Young et al., 2010; Duff, 2011; Roke et al., 2012; Rushton, 2014), while showing how effects are produced (Correa and Domènech, 2013). More than an analytical tool, ANT is a methodology for understanding the world (Latour, 2005; Broer et al., 2010a; Bleakley, 2012). Given that it achieves its full heuristic potential when applied to concrete situations, we thought that if we could successfully apply it to the analysis of population health intervention it would provide useful (and much needed) theoretical and methodological frameworks for population health intervention research (PHIR). In this article, we set out a conceptual framework and methodological guidelines we have derived from ANT for investigating PHI based on both our own field studies on public health interventions and other researchers’ studies that have employed ANT to unpack the black box of complex interventions. We first justify the use of ANT for the analysis of PHI; we then present our ANT-based framework for (complex) population health interventions, followed by a series of methodological guidelines derived from the framework. Finally, we discuss the usefulness of this work to public health research and practice. The appropriateness of applying ANT to the analysis of PHI PHI addressing social determinants of health at a population level is still an innovative field in public health practice. It challenges traditional public health knowledge, especially in the formulation of intervention theories that needs to incorporate social determinants and actor mobilization for social change (Potvin et al., 2005). Such interventions are complex systems (Hawe, 2015) that mobilise actors from various sectors (e.g. education, transportation, work, urban planning, welfare) and spheres (public, community, philanthropic, private), at the community and government levels, together with a variety of non-human entities such as specialised knowledge, resources, communication technology and so forth. Such disparate assemblages of entities are called hybrid actor–networks (Callon et al., 2001). Interventions of this type must not be exclusively conceived of as techno-scientific objects, meaning knowledge-based activities developed and implemented in response to problems (Hawe, 2015). Instead, they must be conceived of as evolving objects that take form and transform in response to the contexts in which they are implemented (Potvin et al., 2001; Poland et al., 2008; Hawe, 2015). This view is in line with the conceptualization of interventions as systems, which focuses on the appraisal of the whole rather than its parts, emphasizing relationships, interactions and feedback loops between the various components and between the system and its context (Hawe et al., 2009). Although it is always possible to focus on specific areas or subsystems, such as the different levels of action, strategies and resources involved, conceiving PHIs as systems provides a framework for problematizing relationships between interventions and their contexts, and how such relationships transform them both (Poland et al., 2008; Potvin and McQueen, 2008a; Shiell et al., 2008; Hawe et al., 2009; Potvin and Clavier, 2012). Focusing on the dynamic and recursive interactions between an intervention and its context leads to examining the connections between the various elements of the intervention and context, the network they form and the evolution of that network. This causes the observer to pay attention to interactions between the various entities—human and non-human—comprising such networks, their converging and diverging interests, their mutual influence, and the various actions undertaken by intervention providers or beneficiaries. From this perspective, intervention constitutes a strategy for enabling actors to change roles or create new ones, establish or strengthen connections within and between existing networks, create new networks and mobilise new resources, with the aim of producing changes in context. PHIs addressing social determinants of health at a population level are still in need of theoretical general frameworks to provide adequate support for innovation and transform practices. Meanwhile, leaders and researchers in the field have called for theories elucidating the roles of social actors, their agency and associations, and the opportunities and constraints that exist in the contexts in which they interact (Potvin, 2005; Willis et al., 2007; Hawe et al., 2009). Contemporary social theories offer relevant frameworks for addressing the shortcomings in public health interventions aimed at producing social change, especially the under-theorization responsible for slowing down the production of general knowledge about the processes at work. Three arguments support the pertinence of using ANT to analyse PHIs as processes of social change. Firstly, ANT suggests a relational view of action; secondly, it conceives of the context as being defined by the actors and their actions; and thirdly, it allows for the investigation of how effects are produced. ANT is a powerful theoretical tool for studying PHIs in that it proposes a relational view of action. It locates humans together with material and technical objects inside a complex web of associations. It conceives of actors as performing actions in association with other actors. Focusing on actors, ANT encompasses and scrutinises the large network of connections supporting or restraining action. This perspective is well suited to complex population health interventions that often involve partnership and intersectional action, while assembling a wide range of entities, whether material—such as technical devices and resources—or social, such as organizations (Bilodeau et al., 2011; Potvin and Clavier, 2013). It is not the actors per se that comprise the focal point of ANT but rather the connections between them through which they act. Interventions are thus seen as systems of actions in the form of networks (Latour, 2005; Cresswell et al., 2010, 2011; Van Herzele et al., 2011). This is in conformity with Hawe et al. (2009)’s view of interventions as networks of events. The second argument in favour of ANT is that it promotes the conceptualization of context in relation to networks of actors and their actions. Social actors are constantly positioning themselves and their actions in relation to their context, identifying the universe in which they act, its contours, scale, and make-up, all of which evolve through their action. The researcher’s task therefore consists of following and documenting the evolution and constant reassembling of network components (Latour, 2005). Instead of context being conceived of as a backdrop for networks, intervention contextualization is understood as being the by-product of networks (Andrews et al., 2013). Thus, a researcher employing ANT would view contextualisation (i.e. the intervention in context) as a jumping-off point for understanding a situation, conceiving of it in terms of connectedness and the web of connections linking the entities involved in the situation. In network analysis, entities only acquire meaning and importance with regard to their connections (Latour, 2005; Callon and Feray, 2006; Andrews et al., 2013). Contextualisation thus consists of identifying the various relevant components in a situation (social actors and their roles, physical objects in the form of material and financial resources, and symbolic objects in the form of scientific or experiential knowledge, values, norms and rules), analysing their attributes, strategic positions and power relationships, points of convergence and divergence with regard to the situation in question, and what is at stake for them, and distinguishing areas of stronger and denser connections. The third argument for the relevance of ANT to the study of PHIs is that it is a powerful theoretical system for opening up the black box of interventions (Bleakley, 2012). It allows for investigation of the production of effects, i.e. the complex process by which PHIs act and transform the context in which they are implemented. Documenting how intervention effects are produced remains a challenge for public health intervention research (Potvin and McQueen, 2008b). ANT offers a solution to this problem in that it provides an appropriate analytical framework for capturing actions operated through interventions, and makes it possible to reconstruct the various connections through which interventions change contexts and impact population health. The ANT-based analytical approach helps illustrate how effects are produced by connecting various entities (social actors, ideas, facts, resources) that act collectively in a network, and then document the chain of events that produces new connections leading to observed effects (Callon and Ferrary, 2006; Cresswell et al., 2010, 2011; Correa and Domènech, 2013). Paying close attention to the processes involved in producing effects contributes to a more comprehensive view of the complexity of PHI. ANT assumes that relationships among and between actors (both human and non-human) in an emerging network are mostly unpredictable and affect social environment in ways that are often unexpected (Cresswell et al., 2011). ANT however suggests that strengthening connections within a network leads to a better organised network and that the predictability of a network’s behaviour (effects) increases with its level of organisation (Latour, 2005). ANT-based framework for public health intervention research Among the numerous principles founding ANT we have identified three as the most important to the study of PHIs. First, ANT focuses on the connections between heterogeneous entities that make up a situation rather than on independently existing entities. It is concerned with the process of assembling social connections rather than their stabilised form. Thus, it considers ‘an actor [to be] what is made to act by many others’ (Latour, 2005: 46). Actors can only act in association with others and in constellations that give them the possibility of acting (Mol, 1999; Latour, 2005). Therefore, in the context of PHI, ANT concentrates on the processes responsible for producing the effects of interventions, based on the premise that observable changes are, in fact, the end result of completed processes. For example, the development of a tobacco control policy requires actors from different domains to connect with one another and mobilise scientific evidence in advocacy actions in order to influence the public discourse in such a manner as to promote the restriction of smokers freedom so as to convince policy makers to act accordingly (Breton et al., 2008). Second, one of the foundational proposition in ANT is the principle of symmetry which essentially rejects the dualism between human and non-human entities. As a consequence, ANT affords human and non-human entities in networks the same treatment. This means that both human and non-human entities are considered to be capable of agency, depending on how they are associated within a given network. Non-human entities’ agency is granted to them by human entities that activate them within a network (Potvin and Clavier, 2012). Consequently, ANT examines what objects make humans do as well as what humans make objects do, using the same conceptual repertory (Callon, 1986; Latour, 1987; Latour, 2005; Callon and Ferrary, 2006). Thus, in an area such as PHI, in which social actors (humans, organizations, institutions, etc.) are at the forefront, ANT would highlight both the performative capacities of non-humans entities, such as scientific evidence, expert knowledge, financial resources and regulations, as well as human agencies in shaping PHI. For example, recommendations concerning the plain packaging of tobacco products are based on evidence that cigarette packages themselves influence choices among certain groups (British Heart Foundation, 2011). Third, another consequence of the principle of symmetry is that humans and technologies they produce are intertwined, and they cannot be understood independently from one another. Applying this principle to public health intervention as technology (Lehoux, 2006), this means the sociotechnical network and the intervention shape one another. The intervention is made possible only to the extent that it is supported by a sociotechnical network. In turn, the network is gradually constructed around the intervention. The intervention develops through the construction, expansion and strengthening of the network. (Callon and Latour, 2006). Thus, ANT offers a new way of understanding innovation in PHI, as well as their sustainability and dissemination. For example, the routinization of a programme within an organization requires linkages between elements from the programme and the environment to be strengthened at the expense of other competing events (Pluye et al., 2004). Central to ANT is the concept of translation, which refers to the linkage operation that connects disparate entities—humans, things, new ideas, interests, values, specialised and lay knowledge, financial resources, institutions, regulations, etc.—in a situation. These hybrid collections of connected entities are called sociotechnical networks. Translation is the process by which networks are created, expand, and act. They act through the creation, reconfiguration, strengthening of connections and expansion of a sociotechnical network that connects new actors, resources and knowledge, resulting in the development of innovative solutions. The translation process consists of four non-linear operations (Callon, 1986). Problematization involves the identification and connection of entities that are relevant to the situation at hand. It consists of assigning the entities roles and interests related to the situation. It also includes identifying problems, as well as potential solutions and controversies. The latter means the confrontation of differing views that actors in a given situation may hold and that are tied to their role and interest. Problematization evolves over time through interactions and controversies within the network (Broer et al., 2010a). Interessement is the development of convergence in a network through the negotiation of roles and interests (Akrich et al., 2002). It involves actors deploying strategies and mobilizing new entities in order to rally others to their own problematization. Enrolment is the product of successful interessement. It is the alignment of actors’ interests based on their acceptance of roles as defined through negotiations. Enrolment implies certain shifts in actors’ positions (actor displacements) and the closure of controversies. Mobilization consists of achieving a critical mass of connected actors able to coordinate their actions, thereby testing the solidity of those connections and the capacity of spokespersons to act and speak on behalf of the broader network. A certain level of solidity is required for a network to acquire sustainability and the capacity to act (Potvin and Clavier, 2012). Once actors have been enrolled, the network may stabilise for a certain amount of time and produce results (Latour, 2005; Andrews et al., 2013; Stoopendaal and Bal, 2013). In order for networks to operate and produce adequate solutions, new roles must be proposed to relevant entities and accepted by them. Commitment to new roles involves actors assuming new identities, which necessarily produces controversy (Potvin and Clavier, 2013). The notion of controversy is central to the translation process. Controversies link together and enmesh the techno-scientific and political contents that make up the issues facing actors (Lascoume, 2002). Solving controversies is essential in order to develop collaboration within a network. It is the closure of controversies among actors that creates agreement among them and strengthens the network, stabilizing the intervention. Controversies are solved through translation by the addition of knowledge, other viewpoints and argumentative elements, as well as by the strengthening of existing connections and the enrolment of relevant new actors bringing new knowledge and resources necessary for action (Sénécal, 2012; Potvin and Clavier, 2013). Solving controversies consists of concerned actors striving to elaborate new, more robust solutions that would allow them to improve their cooperation while satisfying their own interests, at least in part (Callon, 1986; Callon, 1988; Bilodeau et al., 2002; Latour, 2005; Akrich et al., 2006). It should be noted that different networks pursuing disparate or opposite goals may compete for allies, institutional endorsement, public support and resources, at a particular time and place (Young et al., 2010). ANT also includes the concepts of mediator and of displacement and inscription into social or material forms that act as intermediaries in sociotechnical networks. Translation is operated by mediators, who are actors having the capacity to displace other actors thereby creating new connections that prompt changes in the network’s entities. Displacements are the result of negotiations and persuasive actions through which actors manage to gain adhesion from others, reorganise connections, and shape the network (Latour, 1987; Correa and Domènech, 2013). As opposed to mediators, whose actions transform networks, intermediaries are entities that ‘transport meaning or force without transformation’ (Latour, 2005: 39). Such inscriptions—i.e. documents, devices, trained individuals or money (Callon, 1991; Stoopendaal and Bal, 2013)— circulate among the network’s actors, stabilizing agreement and making innovations sustainable and capable of diffusion and dissemination. Dissemination involves connecting the interventions with other networks in which they are to be adopted (Callon and Latour, 1986). ANT-derived methodological guidelines As shown above, ANT provides a powerful theoretical framework that offers solid foundations for interpreting interventions as they are developed, interact with their context, produce effects, and are sustained and disseminated. It may be used as a methodology for opening up the black box of PHI in that it elucidates the way in which a wide range of entities are linked together and displaced to produce social change. Three methodological guidelines have been derived from ANT as relevant to PHIR: developing a system of inscription to follow actor networks, adopting a chronological ordering of events and making practical decisions during the field work, and using a broad range of data sources. Firstly, the researcher’s task consists of following and documenting transformations in sociotechnical networks, their actions, and the context in which this occurs, and analysing how the actors themselves define what is going on (Callon, 1986; Latour, 1987; Broer et al., 2010a, b; Potvin and Clavier, 2013). Drawing on the concept of inscription (Callon, 2006; Latour, 2007; Potvin and Clavier, 2012; Stoopendaal and Bal, 2013), the researcher develops a system by which data and initial interpretations and analysis may be recorded in a material form that aggregates and synthesises information. Such system must include specific tools designed for following the evolution of the actor–networks, their intervention and context. These systems of inscription capture reality and help create meaning and knowledge. They shape the interpretation of situations under analysis by organizing data in the sense of ANT. They make it easier to manage and synthesise vast amounts of information, and permits thoughts, analysis under construction and primary results to be expressed in compact form. This type of system also facilitates the sharing of thoughts with stakeholders and practitioners during collaborative research or for collective reflection. The second methodological guideline is the adoption of a chronological ordering of events, as ANT focuses on processes and processes evolve over time. This makes it possible to record PHIs as they develop (for data collected prospectively) or have developed (for data collected retrospectively). Consequently, it entails adopting the position of ‘the researcher following the thing’ (Cresswell et al., 2011; Van Herzele et al., 2011), i.e. making practical decisions about objects to be investigated and methods during field work rather than only a priori, based on the evolution of the process and the research objectives. The third guideline is the use of a broad range of data at different system levels to capture the complex ways in which PHIs evolve, instead of planned data collection such as planned one-shot interviews with stakeholders—as Latour stated: ‘Everything is data’ (2005: 133). The various recommended ethnographic methods and types of longitudinal data are as follows: (i) observational notes taken during events (meetings, planning days, team-building sessions); (ii) logbooks kept by board members, managers or practitioners throughout the intervention; (iii) interviews conducted at appropriate times; (iv) government documents, internal administrative documents (strategic plans, progress reports) and intervention tools; (v) media (local and national newspapers, websites) and ephemera, such as letters, mails, contracts, meeting agendas and reports. Implications A systemic perspective that conceives of PHIs as transformative processes in complex systems opens up new possibilities for the creation of scientific knowledge relevant to public health planning. Having the necessary conceptual and methodological frameworks for capturing and interpreting observable manifestations of ongoing processes is a prerequisite for achieving this aim (Potvin et al., 2010; Petticrew, 2011). The initial challenge is the development of a theoretical understanding of public health intervention, how it works and produces change (Datta and Petticrew, 2013). ANT offers a comprehensive framework for the analysis of complex systems of change. It provides a framework based on a general social theory for interpreting empirical data and making sense of PHIs. The theoretical and methodological frameworks presented in this article help capture the observable manifestations of social processes that take place during PHI, connecting them to a larger body of knowledge on similar social processes. This perspective is highly relevant in light of repeated calls for more comprehensive (high level) theories of change pertaining to such interventions (Larsen, 2001; Barnes et al., 2003; Auspos and Kubisch, 2004; Potvin, 2005; Potvin et al., 2005; Willis et al. 2007; Hawe et al., 2009). Anticipating and observing changes at the various levels at which they occur is another challenge to understanding and theorizing complex systems of actions. Such complex systems have properties that pertain to the system as a whole, and data must be gathered on multiple levels with tools designed for this purpose. This type of methodology is considered a prerequisite for the study of such interventions (Shiell et al., 2008; Datta and Petticrew, 2013), and equipping PHIR with a system of inscription such as that proposed by ANT would be an appropriate means of achieving this end. Such systems are comprised of diverse sources of data concerning the sociotechnical network, the intervention and the context in which it is implemented. All of these are necessary to capture changes in the system as a whole. Furthermore, complex systems of change require research designs that focus on the intermediate changes that act as preconditions of system-level change, such as movements in the positions of key actors in the network (Shiell et al., 2008). This prerequisite is adequately met by ANT-derived theoretical and methodological frameworks, especially the emphasis on the processes by which changes are produced. Finally, to address adaptivity and feedback processes in complex systems requires an adaptive methodological approach in order to follow actors as they redefine their action as well as any planned or unintended consequences (Shiell et al., 2008; Datta and Petticrew, 2013). The methodological principles of ANT foster adaptive and longitudinal methodologies consisting, on the one part, of following actors and their actions so as to capture the chains of events that occur in complex systems, regardless of whether the events result in changes, planned or otherwise; and on the other, of making decisions regarding methods during field research rather than simply a priori. This sort of research position is well suited to the less predictable nature of change in complex systems, as well as to the different levels on which such change can take place. Adaptivity and feedback processes in complex systems underscore the fact that knowledge about interventions is generated by practitioners as well as by intervention researchers. Complex systems are governed by feedback that may transform interventions. Agents in these systems are thoughtful, and capable of self-monitoring. They correct their courses of action in response to feedback (Hawe, 2015). Thus, research methodologies that are founded on the co-production of knowledge are more appropriate to capturing this reality than traditional designs that suppose a distant researcher. This implies a reciprocal practitioner-researcher relationship consistent with the researcher’s epistemological position. Conclusion Many difficult issues about PHIR are still being debated within both researcher and practitioner communities. This article helps advance these debates by proposing a theoretical and methodological framework derived from ANT that makes it possible to overcome two important traditional cleavages in public health intervention research. The first is between a technicist view (intervention as a programme, a technology or a set of products) and a social view of PHI (interventions as relationships, social actors and resources, power structures, ideas and set of values) (Hawe, 2015). By proposing a sociotechnical view of interventions, ANT illuminates the process by which social actors and technical entities are linked in action networks. In such sociotechnical networks technical entities have agency in the same sense as social actors, in that they can be credited with being the source of actions. The second cleavage is between process and effect studies of interventions, which are traditionally conducted separately using different methodologies, making it difficult to establish empirical and logical connections between the two (Craig et al., 2008). ANT helps reconcile these two types of study by supporting a research approach that employs the same conceptual repertory and methodological principles, whether it be in the identification and observation of transformations in the sociotechnical network and in the intervention it supported, or in the context in which intervention is implemented. This type of approach makes it possible to capture data starting from the first significant changes in processes right through to the end results, expected or otherwise, that transform contexts. Ethical approval Ethical approval is not required because this article is a theoretical and methodological one. Funding information This work was supported by the Canadian Institutes of Health Research, Operating Grant: Programmatic Grants to Tackle Health and Health Equity 2011–2016 [ROH 115211]. Louise Potvin holds the Canada Research Chair in Community Approaches and Health Inequalities (CIHR 950-228295). Acknowledgements This article has benefited from fruitful discussions with colleagues, notably Professor Sylvie Gendron from the University of Montréal, Chantal Lefebvre and Marilène Galarneau, research professionals at the Centre de recherche Léa-Roback sur les inégalités sociales de santé de Montréal. The authors wish to express their thanks to Elizabeth Reeve, Certified Translator, for proofreading this article. References Akrich M., Callon M., Latour B. ( 2002) The key to success in innovation. Part I: The art of interessement. International Journal of Innovation Management , 6, 187– 206. Google Scholar CrossRef Search ADS   Akrich M., Callon M., Latour B. ( 2006) Sociologie de la traduction. Textes fondateurs . École des Mines, Paris. Google Scholar CrossRef Search ADS   Andrews G.J., Evans J., McAlister S. ( 2013) Creating the right therapy vibe: Relational performances in holistic medicine. Social Science & Medicine , 83, 99– 109. Google Scholar CrossRef Search ADS   Auspos P., Kubisch A.C. ( 2004) Building Knowledge About Community Change. Moving Beyond Evaluation . The Aspen Institute, New York. Barnes M., Matka E., Sullivan H. ( 2003) Evidence, understanding and complexity. Evaluation in non-linear systems. Evaluation , 9, 3, 265– 284. Google Scholar CrossRef Search ADS   Bilodeau A., Chamberland C., White D. ( 2002) L’innovation sociale, une condition pour accroitre la qualité de l’action en partenariat dans le champ de la sante publique. Revue canadienne d’évaluation de programme , 17, 2, 59– 88. Bilodeau A., Galarneau M., Fournier M., Potvin L. ( 2011) L’Outil diagnostic de l’action en partenariat: fondements, élaboration et validation. Canadian Journal of Public Health , 102, 4, 298– 302. Google Scholar PubMed  Bisset S.L., Potvin L. ( 2007) Expanding our conceptualisation of program implementation: lessons from the genealogy of a school-based nutrition program. Health Education Research , 22, 737– 746. Google Scholar CrossRef Search ADS PubMed  Bleakley A. ( 2012) The proof is in the pudding: Putting Actor-Network-Theory to work in medical education. Medical Teacher , 34, 462– 467. Google Scholar CrossRef Search ADS PubMed  Breton E., Richard L., Gagnon F., Jacques M., Bergeron P. ( 2008) Health promotion research and practice require sound policy analysis models: the case of Quebec’s Tobacco Act. Social Science & Medicine , 67, 11, 1679– 1689. Google Scholar CrossRef Search ADS   British Heart Foundation. ( 2011) The Plain Truth: How Tobacco Plain Packaging will Save Lives . British Heart Foundation, London, UK. Broer T., Nieboer A.P., Bal R.A. ( 2010a) Opening the black box of quality improvement collaborative: an Actor-network theory approach. BMC Health Services Research , 10, 265– 274. Google Scholar CrossRef Search ADS   Broer T., Nieboer A.P., Bal R.A. ( 2010b) Quest for client autonomy in improving long-term mental health care. International Journal of Mental Health Nursing , 19, 385– 393. Google Scholar CrossRef Search ADS   Callon M., Latour B. ( 1986) Les paradoxes de la modernité. Comment concevoir les innovations? Perspective et Santé , 36, 13– 25. Callon M. ( 1986) Some elements of a sociology of translation: domestication of the scallops and the fishermen of St. Brieux Bay. In Law J. (ed), Power, Action and Belief: A New Sociology of Knowledge? . Routledge & Kegan Paul, London and Boston, pp. 196– 233. (Sociological Review Monograph Series, 32). Callon M. ( 1988) Introduction. In Callon M. (ed) La science et ses réseaux. Genèse et circulation des faits scientifiques . La Découverte, Paris, pp. 7– 32. Callon M. ( 1991) Techno-economic networks and irreversibility. In Law J. (ed.) A Sociology of Monsters: Essays on Power, Technology and Domination . Routledge, London and New York, pp. 132– 161. (Sociological Review Monograph Series, 38). Callon M., Lascoumes P., Barthe Y. ( 2001) Agir dans un monde incertain. Essai sur la démocratie technique . Le Seuil, Paris. Callon M. ( 2006) Sociologie de l’acteur réseau. In Akrich M, Callon M, Latour B. (eds), Sociologie de la traduction. Textes fondateurs . École des mines, Paris, pp. 267– 276. Google Scholar CrossRef Search ADS   Callon M., Latour B. ( 2006) Le grand Leviathan s’apprivoise-t-il?. In Akrich M., Callon M. (eds), Sociologie de la traduction. Textes fondateurs . École des Mines, Paris, pp. 11– 32. Google Scholar CrossRef Search ADS   Callon M., Ferrary M. ( 2006) Les réseaux sociaux à l’aune de la théorie de l’acteur-réseau. Sociologies pratiques , 2, 13, 37– 44. Google Scholar CrossRef Search ADS   Correa G., Domènech M. ( 2013) Care networking: A study of technical mediations in a home telecare service. International Journal of Environmental Research and Public Health , 10, 3072– 3088. Google Scholar CrossRef Search ADS PubMed  Craig P., Dieppe P., Macintyre S., Mitchie S., Nazareth I., Petticrew M. ( 2008) Developing and evaluating complex interventions: the new Medical Research Council Guidance. British Medical Journal , 337, 979– 983. Google Scholar CrossRef Search ADS   Cresswell K.M., Worth A., Sheikh A. ( 2010) Actor-Network Theory and its role in understanding the implementation of information technology developments in healthcare. BMC Medical Informatics and Decision Making , 10, 67– 78. Google Scholar CrossRef Search ADS PubMed  Cresswell K.M., Worth A., Sheikh A. ( 2011) Implementing and adopting electronic health record systems: How actor‐network theory can support evaluation. Clinical Governance: An International Journal , 16, 4, 320– 336. Google Scholar CrossRef Search ADS   Datta J., Petticrew M. ( 2013) Challenges to evaluating complex interventions: a content analysis of published papers. BMC Public Health , 13, 568– 586. Google Scholar CrossRef Search ADS PubMed  Duff C. ( 2011) Networks, resources and agencies: On the character and production of enabling places. Health and Place , 17, 149– 156. Google Scholar CrossRef Search ADS   Hawe P., Shiell A., Riley T. ( 2009) Theorising interventions as events in systems. American Journal of Community Psychology , 43, 267– 276. Google Scholar CrossRef Search ADS PubMed  Hawe P. ( 2015) Lessons from complex interventions to improve health. Annual Review of Public Health , 36, 307– 323. Google Scholar CrossRef Search ADS PubMed  Larsen P.D. ( 2001) From programme theory to constructivism: On tragic, magic and competing programmes. Evaluation , 7, 3, 331– 350. Google Scholar CrossRef Search ADS   Lascoume P. ( 2002) Chapitre 7. De l’utilité des controverses sociotechniques. Journal International de Bioéthique , 3, 68– 79. Google Scholar CrossRef Search ADS   Latour B. ( 1987) Science in Action: How to Follow Scientists and Engineers through Society . Harvard University Press, Cambridge. Latour B. ( 2005) Reassembling the Social. An Introduction to Actor-Network-Theory . Oxford University Press, New York. Latour B. ( 2007) L’espoir de Pandore. Pour une vision réaliste de l’activité scientifique . La Découverte, Paris. Lehoux P. ( 2006) The Problem of Health Technology. Policy Implications for Modern Health Care Systems . Routledge, New York. Mol A. ( 1999) Ontological politics: a word and some questions. In Law J., Hassard J. (eds), Actor Network Theory and After . Blackwell Publishers, Oxford, pp 74– 89. (Sociological Review Monograph Series). Petticrew M. ( 2011) When are complex interventions ’complex’? When simple interventions ‘simple’?. European Journal of Public Health , 21, 4, 397– 399. Google Scholar CrossRef Search ADS PubMed  Pluye P., Potvin L., Denis J.L. ( 2004) Making public health programs last: Conceptualizing sustainability. Evaluation and Program Planning , 27, 121– 133. Google Scholar CrossRef Search ADS   Poland B., Frohlich K.L., Cargo M. ( 2008) Context as a fundamental dimension of health promotion program evaluation. In Potvin L., McQueen D. (eds), Health Promotion Evaluation Practices in the Americas . Springer, New York, pp. 299– 317. Google Scholar CrossRef Search ADS   Potvin L., Haddad S., Frohlich K.L. ( 2001) Beyond process and outcome evaluation. A comprehensive approach for evaluating health promotion programmes. In Rootman I., Goodstadt M., Hyndman B., McQueen D.V., Potvin L., Springett J., Ziglio E (eds), Evaluation in Health Promotion. Principles and Perspectives . European series, No 92. WHO Regional Publications, Copenhague, pp. 45– 62. Potvin L. ( 2005) Integrating social theory into public health practice. American Journal of Public Health , 95, 4, 591– 595. Google Scholar CrossRef Search ADS PubMed  Potvin L., Gendron S., Bilodeau A., Chabot P. ( 2005) Social theory and the renewal of the foundations of contemporary public health practice. American Journal of Public Health , 95, 591– 595. Google Scholar CrossRef Search ADS PubMed  Potvin L., McQueen D. ( 2008a) Introduction. In Potvin L., McQueen D. (eds) Health Promotion Evaluation Practices in the Americas. Values and Research . Springer, New York, pp. 25– 45. Google Scholar CrossRef Search ADS   Potvin L., McQueen D. ( 2008b) Health Promotion Evaluation Practices in the Americas. Values and Research . Springer, New York. Potvin L., Bisset S. L., Walz L. ( 2010) Participatory action research: Theoretical perspectives on the challenges of researching action. In Bourgeault Y., Dingwall R., de Vries R. (eds) The Sage Handbook of Qualitative Methods in Health Research . Sage, London, pp. 433– 453. Google Scholar CrossRef Search ADS   Potvin L., Clavier C. ( 2012) La théorie de l’acteur-réseau. In Aubry F., Potvin L. (eds) Construire l’espace socio-sanitaire. Expériences et pratiques de recherche dans la production locale de la santé . Presses de l’Université de Montréal, Montréal, pp. 75– 98. Potvin L., Clavier C. ( 2013) Actor-Network Theory: The governance of intersectoral initiatives. In Clavier C., De Leeuw E. (eds.) Health Promotion and the Policy Process . Oxford University Press, Oxford, pp. 82– 103. Google Scholar CrossRef Search ADS   Ramiller N.C. ( 2007) Constructing safety: System designs, system effects, and the play of heterogeneous interests in a behavioral health care setting. International Journal of Medical Informatics , 76S, S196– S204. Google Scholar CrossRef Search ADS   Roke C., Cloatre E., Dingwall R. ( 2012) The regulation of nicotine in the United Kingdom: How nicotine gum came to be a medicine, but not a drug. Journal of Law and Society , 39, 1, 39– 57. Google Scholar CrossRef Search ADS PubMed  Rushton C. ( 2014) Whose place is it anyway? Representational politics in a place-based health initiative. Health and Place , 26, 100– 109. Google Scholar CrossRef Search ADS PubMed  Stoopendaal A., Bal R. ( 2013) Conferences, tablecloths and cupboards: How to understand the situatedness of quality improvements in long-term care. Social Science & Medicine , 78, 78– 85. Google Scholar CrossRef Search ADS   Shiell A., Hawe P., Gold L. ( 2008) Complex interventions or complex systems? Implications for health economic evaluation. British Medical Journal , 336, 1281– 1283. Google Scholar CrossRef Search ADS PubMed  Sénécal G. ( 2012) L’acteur social: L’expérience du quartier Villeray à Montréal. In Aubry F., Potvin L. (eds), Construire l’espace sociosanitaire. Expériences et pratiques de recherche dans la production locale de la santé . Presses de l’Université de Montréal, Montréal, pp. 141– 162. Timpka T., Nordqvist C., Lindqvist K. ( 2009) Infrastructural requirements for local implementation of safety policies: the discordance between top-down and bottom-up systems of action. BMC Health Services Research , 9, 45– 53. Google Scholar CrossRef Search ADS PubMed  Van Herzele A., Dendoncker N., Acosta-Michlik L. ( 2011) Mobilisation capacity for agri-environmental management. Journal of Environmental Management , 92, 1023– 1032. Google Scholar CrossRef Search ADS PubMed  Willis K., Daly J., Kealy M., Small R., Koutroulis G., Green J. et al.   ( 2007) The essential role of social theory in qualitative public health research. Australian and New Zealand Journal of Public Health , 31, 438– 443. Google Scholar CrossRef Search ADS PubMed  Young D., Borland R., Coghill K. ( 2010) An actor-network theory analysis of policy innovation for smoke-free places: Understanding change in complex systems. American Journal of Public Health , 100, 7, 1208– 1217. Google Scholar CrossRef Search ADS PubMed  © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Health Promotion International Oxford University Press

Unpacking complexity in public health interventions with the Actor–Network Theory

Loading next page...
 
/lp/ou_press/unpacking-complexity-in-public-health-interventions-with-the-actor-ZdikBdZVKA
Publisher
Oxford University Press
Copyright
© The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0957-4824
eISSN
1460-2245
D.O.I.
10.1093/heapro/daw062
Publisher site
See Article on Publisher Site

Abstract

Abstract This article proposes a sociologically informed theoretical and methodological framework to address the complexity of public health interventions (PHI). It first proposes three arguments in favour of using the Actor–Network Theory (ANT) for the framework. ANT: (1) deals with systems made of human and non-human entities and proposes a relational view of action; (2) provides an understanding of the intervention–context interactions and (3) is a tool for opening the intervention’s black box. Three principles derived from ANT addressing theoretical problems with conceptualisation of PHI as complex systems are proposed: (1) to focus on the process of connecting the network entities instead of their stabilised form; (2) both human and non-human entities composing networks have performative capacities and (3) network and intervention shape one another. Three methodological guidelines are further derived: (1) the researcher’s task consists in documenting the events that transform the network and intervention; (2) events must be ordered chronologically to represent the intervention’s evolution and (3) a broad range of data is needed to capture complex interventions’ evolution. Using ANT as a guide, this paper helps reconcile technicist and social views of PHI and provides a mean to integrate process and effect studies of interventions. Actor–Network theory, complex system perspective, public health intervention research Introduction Actor–network theory (ANT) has been developed by Bruno Latour, Michel Callon and others since the 1980s. Originally, ANT was developed as a sociology of science focussing on understanding the production of scientific facts and technology. Because this theory developed a renewed vision of the social world, it evolved into a grand sociological theory known by the name of sociology of translation or ANT. Latour (2005) proposed an in-depth presentation of this theory in his classic book Reassembling the social. ANT is now recognised as a useful conceptual tool for appraising complex situations and analysing the production of change. The theory is increasingly used to understand change in complex systems such as health care, where it has been employed to study the introduction of technology or of new strategies for service improvement (Ramiller, 2007; Cresswell et al., 2010, 2011; Bleakley, 2012; Stoopendaal and Bal, 2013). It offers a powerful tool for opening up the black box of public health interventions as it provides a means for mapping the genesis of interventions in the form of a networking process (Bisset and Potvin, 2007; Timpka et al., 2009; Young et al., 2010; Duff, 2011; Roke et al., 2012; Rushton, 2014), while showing how effects are produced (Correa and Domènech, 2013). More than an analytical tool, ANT is a methodology for understanding the world (Latour, 2005; Broer et al., 2010a; Bleakley, 2012). Given that it achieves its full heuristic potential when applied to concrete situations, we thought that if we could successfully apply it to the analysis of population health intervention it would provide useful (and much needed) theoretical and methodological frameworks for population health intervention research (PHIR). In this article, we set out a conceptual framework and methodological guidelines we have derived from ANT for investigating PHI based on both our own field studies on public health interventions and other researchers’ studies that have employed ANT to unpack the black box of complex interventions. We first justify the use of ANT for the analysis of PHI; we then present our ANT-based framework for (complex) population health interventions, followed by a series of methodological guidelines derived from the framework. Finally, we discuss the usefulness of this work to public health research and practice. The appropriateness of applying ANT to the analysis of PHI PHI addressing social determinants of health at a population level is still an innovative field in public health practice. It challenges traditional public health knowledge, especially in the formulation of intervention theories that needs to incorporate social determinants and actor mobilization for social change (Potvin et al., 2005). Such interventions are complex systems (Hawe, 2015) that mobilise actors from various sectors (e.g. education, transportation, work, urban planning, welfare) and spheres (public, community, philanthropic, private), at the community and government levels, together with a variety of non-human entities such as specialised knowledge, resources, communication technology and so forth. Such disparate assemblages of entities are called hybrid actor–networks (Callon et al., 2001). Interventions of this type must not be exclusively conceived of as techno-scientific objects, meaning knowledge-based activities developed and implemented in response to problems (Hawe, 2015). Instead, they must be conceived of as evolving objects that take form and transform in response to the contexts in which they are implemented (Potvin et al., 2001; Poland et al., 2008; Hawe, 2015). This view is in line with the conceptualization of interventions as systems, which focuses on the appraisal of the whole rather than its parts, emphasizing relationships, interactions and feedback loops between the various components and between the system and its context (Hawe et al., 2009). Although it is always possible to focus on specific areas or subsystems, such as the different levels of action, strategies and resources involved, conceiving PHIs as systems provides a framework for problematizing relationships between interventions and their contexts, and how such relationships transform them both (Poland et al., 2008; Potvin and McQueen, 2008a; Shiell et al., 2008; Hawe et al., 2009; Potvin and Clavier, 2012). Focusing on the dynamic and recursive interactions between an intervention and its context leads to examining the connections between the various elements of the intervention and context, the network they form and the evolution of that network. This causes the observer to pay attention to interactions between the various entities—human and non-human—comprising such networks, their converging and diverging interests, their mutual influence, and the various actions undertaken by intervention providers or beneficiaries. From this perspective, intervention constitutes a strategy for enabling actors to change roles or create new ones, establish or strengthen connections within and between existing networks, create new networks and mobilise new resources, with the aim of producing changes in context. PHIs addressing social determinants of health at a population level are still in need of theoretical general frameworks to provide adequate support for innovation and transform practices. Meanwhile, leaders and researchers in the field have called for theories elucidating the roles of social actors, their agency and associations, and the opportunities and constraints that exist in the contexts in which they interact (Potvin, 2005; Willis et al., 2007; Hawe et al., 2009). Contemporary social theories offer relevant frameworks for addressing the shortcomings in public health interventions aimed at producing social change, especially the under-theorization responsible for slowing down the production of general knowledge about the processes at work. Three arguments support the pertinence of using ANT to analyse PHIs as processes of social change. Firstly, ANT suggests a relational view of action; secondly, it conceives of the context as being defined by the actors and their actions; and thirdly, it allows for the investigation of how effects are produced. ANT is a powerful theoretical tool for studying PHIs in that it proposes a relational view of action. It locates humans together with material and technical objects inside a complex web of associations. It conceives of actors as performing actions in association with other actors. Focusing on actors, ANT encompasses and scrutinises the large network of connections supporting or restraining action. This perspective is well suited to complex population health interventions that often involve partnership and intersectional action, while assembling a wide range of entities, whether material—such as technical devices and resources—or social, such as organizations (Bilodeau et al., 2011; Potvin and Clavier, 2013). It is not the actors per se that comprise the focal point of ANT but rather the connections between them through which they act. Interventions are thus seen as systems of actions in the form of networks (Latour, 2005; Cresswell et al., 2010, 2011; Van Herzele et al., 2011). This is in conformity with Hawe et al. (2009)’s view of interventions as networks of events. The second argument in favour of ANT is that it promotes the conceptualization of context in relation to networks of actors and their actions. Social actors are constantly positioning themselves and their actions in relation to their context, identifying the universe in which they act, its contours, scale, and make-up, all of which evolve through their action. The researcher’s task therefore consists of following and documenting the evolution and constant reassembling of network components (Latour, 2005). Instead of context being conceived of as a backdrop for networks, intervention contextualization is understood as being the by-product of networks (Andrews et al., 2013). Thus, a researcher employing ANT would view contextualisation (i.e. the intervention in context) as a jumping-off point for understanding a situation, conceiving of it in terms of connectedness and the web of connections linking the entities involved in the situation. In network analysis, entities only acquire meaning and importance with regard to their connections (Latour, 2005; Callon and Feray, 2006; Andrews et al., 2013). Contextualisation thus consists of identifying the various relevant components in a situation (social actors and their roles, physical objects in the form of material and financial resources, and symbolic objects in the form of scientific or experiential knowledge, values, norms and rules), analysing their attributes, strategic positions and power relationships, points of convergence and divergence with regard to the situation in question, and what is at stake for them, and distinguishing areas of stronger and denser connections. The third argument for the relevance of ANT to the study of PHIs is that it is a powerful theoretical system for opening up the black box of interventions (Bleakley, 2012). It allows for investigation of the production of effects, i.e. the complex process by which PHIs act and transform the context in which they are implemented. Documenting how intervention effects are produced remains a challenge for public health intervention research (Potvin and McQueen, 2008b). ANT offers a solution to this problem in that it provides an appropriate analytical framework for capturing actions operated through interventions, and makes it possible to reconstruct the various connections through which interventions change contexts and impact population health. The ANT-based analytical approach helps illustrate how effects are produced by connecting various entities (social actors, ideas, facts, resources) that act collectively in a network, and then document the chain of events that produces new connections leading to observed effects (Callon and Ferrary, 2006; Cresswell et al., 2010, 2011; Correa and Domènech, 2013). Paying close attention to the processes involved in producing effects contributes to a more comprehensive view of the complexity of PHI. ANT assumes that relationships among and between actors (both human and non-human) in an emerging network are mostly unpredictable and affect social environment in ways that are often unexpected (Cresswell et al., 2011). ANT however suggests that strengthening connections within a network leads to a better organised network and that the predictability of a network’s behaviour (effects) increases with its level of organisation (Latour, 2005). ANT-based framework for public health intervention research Among the numerous principles founding ANT we have identified three as the most important to the study of PHIs. First, ANT focuses on the connections between heterogeneous entities that make up a situation rather than on independently existing entities. It is concerned with the process of assembling social connections rather than their stabilised form. Thus, it considers ‘an actor [to be] what is made to act by many others’ (Latour, 2005: 46). Actors can only act in association with others and in constellations that give them the possibility of acting (Mol, 1999; Latour, 2005). Therefore, in the context of PHI, ANT concentrates on the processes responsible for producing the effects of interventions, based on the premise that observable changes are, in fact, the end result of completed processes. For example, the development of a tobacco control policy requires actors from different domains to connect with one another and mobilise scientific evidence in advocacy actions in order to influence the public discourse in such a manner as to promote the restriction of smokers freedom so as to convince policy makers to act accordingly (Breton et al., 2008). Second, one of the foundational proposition in ANT is the principle of symmetry which essentially rejects the dualism between human and non-human entities. As a consequence, ANT affords human and non-human entities in networks the same treatment. This means that both human and non-human entities are considered to be capable of agency, depending on how they are associated within a given network. Non-human entities’ agency is granted to them by human entities that activate them within a network (Potvin and Clavier, 2012). Consequently, ANT examines what objects make humans do as well as what humans make objects do, using the same conceptual repertory (Callon, 1986; Latour, 1987; Latour, 2005; Callon and Ferrary, 2006). Thus, in an area such as PHI, in which social actors (humans, organizations, institutions, etc.) are at the forefront, ANT would highlight both the performative capacities of non-humans entities, such as scientific evidence, expert knowledge, financial resources and regulations, as well as human agencies in shaping PHI. For example, recommendations concerning the plain packaging of tobacco products are based on evidence that cigarette packages themselves influence choices among certain groups (British Heart Foundation, 2011). Third, another consequence of the principle of symmetry is that humans and technologies they produce are intertwined, and they cannot be understood independently from one another. Applying this principle to public health intervention as technology (Lehoux, 2006), this means the sociotechnical network and the intervention shape one another. The intervention is made possible only to the extent that it is supported by a sociotechnical network. In turn, the network is gradually constructed around the intervention. The intervention develops through the construction, expansion and strengthening of the network. (Callon and Latour, 2006). Thus, ANT offers a new way of understanding innovation in PHI, as well as their sustainability and dissemination. For example, the routinization of a programme within an organization requires linkages between elements from the programme and the environment to be strengthened at the expense of other competing events (Pluye et al., 2004). Central to ANT is the concept of translation, which refers to the linkage operation that connects disparate entities—humans, things, new ideas, interests, values, specialised and lay knowledge, financial resources, institutions, regulations, etc.—in a situation. These hybrid collections of connected entities are called sociotechnical networks. Translation is the process by which networks are created, expand, and act. They act through the creation, reconfiguration, strengthening of connections and expansion of a sociotechnical network that connects new actors, resources and knowledge, resulting in the development of innovative solutions. The translation process consists of four non-linear operations (Callon, 1986). Problematization involves the identification and connection of entities that are relevant to the situation at hand. It consists of assigning the entities roles and interests related to the situation. It also includes identifying problems, as well as potential solutions and controversies. The latter means the confrontation of differing views that actors in a given situation may hold and that are tied to their role and interest. Problematization evolves over time through interactions and controversies within the network (Broer et al., 2010a). Interessement is the development of convergence in a network through the negotiation of roles and interests (Akrich et al., 2002). It involves actors deploying strategies and mobilizing new entities in order to rally others to their own problematization. Enrolment is the product of successful interessement. It is the alignment of actors’ interests based on their acceptance of roles as defined through negotiations. Enrolment implies certain shifts in actors’ positions (actor displacements) and the closure of controversies. Mobilization consists of achieving a critical mass of connected actors able to coordinate their actions, thereby testing the solidity of those connections and the capacity of spokespersons to act and speak on behalf of the broader network. A certain level of solidity is required for a network to acquire sustainability and the capacity to act (Potvin and Clavier, 2012). Once actors have been enrolled, the network may stabilise for a certain amount of time and produce results (Latour, 2005; Andrews et al., 2013; Stoopendaal and Bal, 2013). In order for networks to operate and produce adequate solutions, new roles must be proposed to relevant entities and accepted by them. Commitment to new roles involves actors assuming new identities, which necessarily produces controversy (Potvin and Clavier, 2013). The notion of controversy is central to the translation process. Controversies link together and enmesh the techno-scientific and political contents that make up the issues facing actors (Lascoume, 2002). Solving controversies is essential in order to develop collaboration within a network. It is the closure of controversies among actors that creates agreement among them and strengthens the network, stabilizing the intervention. Controversies are solved through translation by the addition of knowledge, other viewpoints and argumentative elements, as well as by the strengthening of existing connections and the enrolment of relevant new actors bringing new knowledge and resources necessary for action (Sénécal, 2012; Potvin and Clavier, 2013). Solving controversies consists of concerned actors striving to elaborate new, more robust solutions that would allow them to improve their cooperation while satisfying their own interests, at least in part (Callon, 1986; Callon, 1988; Bilodeau et al., 2002; Latour, 2005; Akrich et al., 2006). It should be noted that different networks pursuing disparate or opposite goals may compete for allies, institutional endorsement, public support and resources, at a particular time and place (Young et al., 2010). ANT also includes the concepts of mediator and of displacement and inscription into social or material forms that act as intermediaries in sociotechnical networks. Translation is operated by mediators, who are actors having the capacity to displace other actors thereby creating new connections that prompt changes in the network’s entities. Displacements are the result of negotiations and persuasive actions through which actors manage to gain adhesion from others, reorganise connections, and shape the network (Latour, 1987; Correa and Domènech, 2013). As opposed to mediators, whose actions transform networks, intermediaries are entities that ‘transport meaning or force without transformation’ (Latour, 2005: 39). Such inscriptions—i.e. documents, devices, trained individuals or money (Callon, 1991; Stoopendaal and Bal, 2013)— circulate among the network’s actors, stabilizing agreement and making innovations sustainable and capable of diffusion and dissemination. Dissemination involves connecting the interventions with other networks in which they are to be adopted (Callon and Latour, 1986). ANT-derived methodological guidelines As shown above, ANT provides a powerful theoretical framework that offers solid foundations for interpreting interventions as they are developed, interact with their context, produce effects, and are sustained and disseminated. It may be used as a methodology for opening up the black box of PHI in that it elucidates the way in which a wide range of entities are linked together and displaced to produce social change. Three methodological guidelines have been derived from ANT as relevant to PHIR: developing a system of inscription to follow actor networks, adopting a chronological ordering of events and making practical decisions during the field work, and using a broad range of data sources. Firstly, the researcher’s task consists of following and documenting transformations in sociotechnical networks, their actions, and the context in which this occurs, and analysing how the actors themselves define what is going on (Callon, 1986; Latour, 1987; Broer et al., 2010a, b; Potvin and Clavier, 2013). Drawing on the concept of inscription (Callon, 2006; Latour, 2007; Potvin and Clavier, 2012; Stoopendaal and Bal, 2013), the researcher develops a system by which data and initial interpretations and analysis may be recorded in a material form that aggregates and synthesises information. Such system must include specific tools designed for following the evolution of the actor–networks, their intervention and context. These systems of inscription capture reality and help create meaning and knowledge. They shape the interpretation of situations under analysis by organizing data in the sense of ANT. They make it easier to manage and synthesise vast amounts of information, and permits thoughts, analysis under construction and primary results to be expressed in compact form. This type of system also facilitates the sharing of thoughts with stakeholders and practitioners during collaborative research or for collective reflection. The second methodological guideline is the adoption of a chronological ordering of events, as ANT focuses on processes and processes evolve over time. This makes it possible to record PHIs as they develop (for data collected prospectively) or have developed (for data collected retrospectively). Consequently, it entails adopting the position of ‘the researcher following the thing’ (Cresswell et al., 2011; Van Herzele et al., 2011), i.e. making practical decisions about objects to be investigated and methods during field work rather than only a priori, based on the evolution of the process and the research objectives. The third guideline is the use of a broad range of data at different system levels to capture the complex ways in which PHIs evolve, instead of planned data collection such as planned one-shot interviews with stakeholders—as Latour stated: ‘Everything is data’ (2005: 133). The various recommended ethnographic methods and types of longitudinal data are as follows: (i) observational notes taken during events (meetings, planning days, team-building sessions); (ii) logbooks kept by board members, managers or practitioners throughout the intervention; (iii) interviews conducted at appropriate times; (iv) government documents, internal administrative documents (strategic plans, progress reports) and intervention tools; (v) media (local and national newspapers, websites) and ephemera, such as letters, mails, contracts, meeting agendas and reports. Implications A systemic perspective that conceives of PHIs as transformative processes in complex systems opens up new possibilities for the creation of scientific knowledge relevant to public health planning. Having the necessary conceptual and methodological frameworks for capturing and interpreting observable manifestations of ongoing processes is a prerequisite for achieving this aim (Potvin et al., 2010; Petticrew, 2011). The initial challenge is the development of a theoretical understanding of public health intervention, how it works and produces change (Datta and Petticrew, 2013). ANT offers a comprehensive framework for the analysis of complex systems of change. It provides a framework based on a general social theory for interpreting empirical data and making sense of PHIs. The theoretical and methodological frameworks presented in this article help capture the observable manifestations of social processes that take place during PHI, connecting them to a larger body of knowledge on similar social processes. This perspective is highly relevant in light of repeated calls for more comprehensive (high level) theories of change pertaining to such interventions (Larsen, 2001; Barnes et al., 2003; Auspos and Kubisch, 2004; Potvin, 2005; Potvin et al., 2005; Willis et al. 2007; Hawe et al., 2009). Anticipating and observing changes at the various levels at which they occur is another challenge to understanding and theorizing complex systems of actions. Such complex systems have properties that pertain to the system as a whole, and data must be gathered on multiple levels with tools designed for this purpose. This type of methodology is considered a prerequisite for the study of such interventions (Shiell et al., 2008; Datta and Petticrew, 2013), and equipping PHIR with a system of inscription such as that proposed by ANT would be an appropriate means of achieving this end. Such systems are comprised of diverse sources of data concerning the sociotechnical network, the intervention and the context in which it is implemented. All of these are necessary to capture changes in the system as a whole. Furthermore, complex systems of change require research designs that focus on the intermediate changes that act as preconditions of system-level change, such as movements in the positions of key actors in the network (Shiell et al., 2008). This prerequisite is adequately met by ANT-derived theoretical and methodological frameworks, especially the emphasis on the processes by which changes are produced. Finally, to address adaptivity and feedback processes in complex systems requires an adaptive methodological approach in order to follow actors as they redefine their action as well as any planned or unintended consequences (Shiell et al., 2008; Datta and Petticrew, 2013). The methodological principles of ANT foster adaptive and longitudinal methodologies consisting, on the one part, of following actors and their actions so as to capture the chains of events that occur in complex systems, regardless of whether the events result in changes, planned or otherwise; and on the other, of making decisions regarding methods during field research rather than simply a priori. This sort of research position is well suited to the less predictable nature of change in complex systems, as well as to the different levels on which such change can take place. Adaptivity and feedback processes in complex systems underscore the fact that knowledge about interventions is generated by practitioners as well as by intervention researchers. Complex systems are governed by feedback that may transform interventions. Agents in these systems are thoughtful, and capable of self-monitoring. They correct their courses of action in response to feedback (Hawe, 2015). Thus, research methodologies that are founded on the co-production of knowledge are more appropriate to capturing this reality than traditional designs that suppose a distant researcher. This implies a reciprocal practitioner-researcher relationship consistent with the researcher’s epistemological position. Conclusion Many difficult issues about PHIR are still being debated within both researcher and practitioner communities. This article helps advance these debates by proposing a theoretical and methodological framework derived from ANT that makes it possible to overcome two important traditional cleavages in public health intervention research. The first is between a technicist view (intervention as a programme, a technology or a set of products) and a social view of PHI (interventions as relationships, social actors and resources, power structures, ideas and set of values) (Hawe, 2015). By proposing a sociotechnical view of interventions, ANT illuminates the process by which social actors and technical entities are linked in action networks. In such sociotechnical networks technical entities have agency in the same sense as social actors, in that they can be credited with being the source of actions. The second cleavage is between process and effect studies of interventions, which are traditionally conducted separately using different methodologies, making it difficult to establish empirical and logical connections between the two (Craig et al., 2008). ANT helps reconcile these two types of study by supporting a research approach that employs the same conceptual repertory and methodological principles, whether it be in the identification and observation of transformations in the sociotechnical network and in the intervention it supported, or in the context in which intervention is implemented. This type of approach makes it possible to capture data starting from the first significant changes in processes right through to the end results, expected or otherwise, that transform contexts. Ethical approval Ethical approval is not required because this article is a theoretical and methodological one. Funding information This work was supported by the Canadian Institutes of Health Research, Operating Grant: Programmatic Grants to Tackle Health and Health Equity 2011–2016 [ROH 115211]. Louise Potvin holds the Canada Research Chair in Community Approaches and Health Inequalities (CIHR 950-228295). Acknowledgements This article has benefited from fruitful discussions with colleagues, notably Professor Sylvie Gendron from the University of Montréal, Chantal Lefebvre and Marilène Galarneau, research professionals at the Centre de recherche Léa-Roback sur les inégalités sociales de santé de Montréal. The authors wish to express their thanks to Elizabeth Reeve, Certified Translator, for proofreading this article. References Akrich M., Callon M., Latour B. ( 2002) The key to success in innovation. Part I: The art of interessement. International Journal of Innovation Management , 6, 187– 206. Google Scholar CrossRef Search ADS   Akrich M., Callon M., Latour B. ( 2006) Sociologie de la traduction. Textes fondateurs . École des Mines, Paris. Google Scholar CrossRef Search ADS   Andrews G.J., Evans J., McAlister S. ( 2013) Creating the right therapy vibe: Relational performances in holistic medicine. Social Science & Medicine , 83, 99– 109. Google Scholar CrossRef Search ADS   Auspos P., Kubisch A.C. ( 2004) Building Knowledge About Community Change. Moving Beyond Evaluation . The Aspen Institute, New York. Barnes M., Matka E., Sullivan H. ( 2003) Evidence, understanding and complexity. Evaluation in non-linear systems. Evaluation , 9, 3, 265– 284. Google Scholar CrossRef Search ADS   Bilodeau A., Chamberland C., White D. ( 2002) L’innovation sociale, une condition pour accroitre la qualité de l’action en partenariat dans le champ de la sante publique. Revue canadienne d’évaluation de programme , 17, 2, 59– 88. Bilodeau A., Galarneau M., Fournier M., Potvin L. ( 2011) L’Outil diagnostic de l’action en partenariat: fondements, élaboration et validation. Canadian Journal of Public Health , 102, 4, 298– 302. Google Scholar PubMed  Bisset S.L., Potvin L. ( 2007) Expanding our conceptualisation of program implementation: lessons from the genealogy of a school-based nutrition program. Health Education Research , 22, 737– 746. Google Scholar CrossRef Search ADS PubMed  Bleakley A. ( 2012) The proof is in the pudding: Putting Actor-Network-Theory to work in medical education. Medical Teacher , 34, 462– 467. Google Scholar CrossRef Search ADS PubMed  Breton E., Richard L., Gagnon F., Jacques M., Bergeron P. ( 2008) Health promotion research and practice require sound policy analysis models: the case of Quebec’s Tobacco Act. Social Science & Medicine , 67, 11, 1679– 1689. Google Scholar CrossRef Search ADS   British Heart Foundation. ( 2011) The Plain Truth: How Tobacco Plain Packaging will Save Lives . British Heart Foundation, London, UK. Broer T., Nieboer A.P., Bal R.A. ( 2010a) Opening the black box of quality improvement collaborative: an Actor-network theory approach. BMC Health Services Research , 10, 265– 274. Google Scholar CrossRef Search ADS   Broer T., Nieboer A.P., Bal R.A. ( 2010b) Quest for client autonomy in improving long-term mental health care. International Journal of Mental Health Nursing , 19, 385– 393. Google Scholar CrossRef Search ADS   Callon M., Latour B. ( 1986) Les paradoxes de la modernité. Comment concevoir les innovations? Perspective et Santé , 36, 13– 25. Callon M. ( 1986) Some elements of a sociology of translation: domestication of the scallops and the fishermen of St. Brieux Bay. In Law J. (ed), Power, Action and Belief: A New Sociology of Knowledge? . Routledge & Kegan Paul, London and Boston, pp. 196– 233. (Sociological Review Monograph Series, 32). Callon M. ( 1988) Introduction. In Callon M. (ed) La science et ses réseaux. Genèse et circulation des faits scientifiques . La Découverte, Paris, pp. 7– 32. Callon M. ( 1991) Techno-economic networks and irreversibility. In Law J. (ed.) A Sociology of Monsters: Essays on Power, Technology and Domination . Routledge, London and New York, pp. 132– 161. (Sociological Review Monograph Series, 38). Callon M., Lascoumes P., Barthe Y. ( 2001) Agir dans un monde incertain. Essai sur la démocratie technique . Le Seuil, Paris. Callon M. ( 2006) Sociologie de l’acteur réseau. In Akrich M, Callon M, Latour B. (eds), Sociologie de la traduction. Textes fondateurs . École des mines, Paris, pp. 267– 276. Google Scholar CrossRef Search ADS   Callon M., Latour B. ( 2006) Le grand Leviathan s’apprivoise-t-il?. In Akrich M., Callon M. (eds), Sociologie de la traduction. Textes fondateurs . École des Mines, Paris, pp. 11– 32. Google Scholar CrossRef Search ADS   Callon M., Ferrary M. ( 2006) Les réseaux sociaux à l’aune de la théorie de l’acteur-réseau. Sociologies pratiques , 2, 13, 37– 44. Google Scholar CrossRef Search ADS   Correa G., Domènech M. ( 2013) Care networking: A study of technical mediations in a home telecare service. International Journal of Environmental Research and Public Health , 10, 3072– 3088. Google Scholar CrossRef Search ADS PubMed  Craig P., Dieppe P., Macintyre S., Mitchie S., Nazareth I., Petticrew M. ( 2008) Developing and evaluating complex interventions: the new Medical Research Council Guidance. British Medical Journal , 337, 979– 983. Google Scholar CrossRef Search ADS   Cresswell K.M., Worth A., Sheikh A. ( 2010) Actor-Network Theory and its role in understanding the implementation of information technology developments in healthcare. BMC Medical Informatics and Decision Making , 10, 67– 78. Google Scholar CrossRef Search ADS PubMed  Cresswell K.M., Worth A., Sheikh A. ( 2011) Implementing and adopting electronic health record systems: How actor‐network theory can support evaluation. Clinical Governance: An International Journal , 16, 4, 320– 336. Google Scholar CrossRef Search ADS   Datta J., Petticrew M. ( 2013) Challenges to evaluating complex interventions: a content analysis of published papers. BMC Public Health , 13, 568– 586. Google Scholar CrossRef Search ADS PubMed  Duff C. ( 2011) Networks, resources and agencies: On the character and production of enabling places. Health and Place , 17, 149– 156. Google Scholar CrossRef Search ADS   Hawe P., Shiell A., Riley T. ( 2009) Theorising interventions as events in systems. American Journal of Community Psychology , 43, 267– 276. Google Scholar CrossRef Search ADS PubMed  Hawe P. ( 2015) Lessons from complex interventions to improve health. Annual Review of Public Health , 36, 307– 323. Google Scholar CrossRef Search ADS PubMed  Larsen P.D. ( 2001) From programme theory to constructivism: On tragic, magic and competing programmes. Evaluation , 7, 3, 331– 350. Google Scholar CrossRef Search ADS   Lascoume P. ( 2002) Chapitre 7. De l’utilité des controverses sociotechniques. Journal International de Bioéthique , 3, 68– 79. Google Scholar CrossRef Search ADS   Latour B. ( 1987) Science in Action: How to Follow Scientists and Engineers through Society . Harvard University Press, Cambridge. Latour B. ( 2005) Reassembling the Social. An Introduction to Actor-Network-Theory . Oxford University Press, New York. Latour B. ( 2007) L’espoir de Pandore. Pour une vision réaliste de l’activité scientifique . La Découverte, Paris. Lehoux P. ( 2006) The Problem of Health Technology. Policy Implications for Modern Health Care Systems . Routledge, New York. Mol A. ( 1999) Ontological politics: a word and some questions. In Law J., Hassard J. (eds), Actor Network Theory and After . Blackwell Publishers, Oxford, pp 74– 89. (Sociological Review Monograph Series). Petticrew M. ( 2011) When are complex interventions ’complex’? When simple interventions ‘simple’?. European Journal of Public Health , 21, 4, 397– 399. Google Scholar CrossRef Search ADS PubMed  Pluye P., Potvin L., Denis J.L. ( 2004) Making public health programs last: Conceptualizing sustainability. Evaluation and Program Planning , 27, 121– 133. Google Scholar CrossRef Search ADS   Poland B., Frohlich K.L., Cargo M. ( 2008) Context as a fundamental dimension of health promotion program evaluation. In Potvin L., McQueen D. (eds), Health Promotion Evaluation Practices in the Americas . Springer, New York, pp. 299– 317. Google Scholar CrossRef Search ADS   Potvin L., Haddad S., Frohlich K.L. ( 2001) Beyond process and outcome evaluation. A comprehensive approach for evaluating health promotion programmes. In Rootman I., Goodstadt M., Hyndman B., McQueen D.V., Potvin L., Springett J., Ziglio E (eds), Evaluation in Health Promotion. Principles and Perspectives . European series, No 92. WHO Regional Publications, Copenhague, pp. 45– 62. Potvin L. ( 2005) Integrating social theory into public health practice. American Journal of Public Health , 95, 4, 591– 595. Google Scholar CrossRef Search ADS PubMed  Potvin L., Gendron S., Bilodeau A., Chabot P. ( 2005) Social theory and the renewal of the foundations of contemporary public health practice. American Journal of Public Health , 95, 591– 595. Google Scholar CrossRef Search ADS PubMed  Potvin L., McQueen D. ( 2008a) Introduction. In Potvin L., McQueen D. (eds) Health Promotion Evaluation Practices in the Americas. Values and Research . Springer, New York, pp. 25– 45. Google Scholar CrossRef Search ADS   Potvin L., McQueen D. ( 2008b) Health Promotion Evaluation Practices in the Americas. Values and Research . Springer, New York. Potvin L., Bisset S. L., Walz L. ( 2010) Participatory action research: Theoretical perspectives on the challenges of researching action. In Bourgeault Y., Dingwall R., de Vries R. (eds) The Sage Handbook of Qualitative Methods in Health Research . Sage, London, pp. 433– 453. Google Scholar CrossRef Search ADS   Potvin L., Clavier C. ( 2012) La théorie de l’acteur-réseau. In Aubry F., Potvin L. (eds) Construire l’espace socio-sanitaire. Expériences et pratiques de recherche dans la production locale de la santé . Presses de l’Université de Montréal, Montréal, pp. 75– 98. Potvin L., Clavier C. ( 2013) Actor-Network Theory: The governance of intersectoral initiatives. In Clavier C., De Leeuw E. (eds.) Health Promotion and the Policy Process . Oxford University Press, Oxford, pp. 82– 103. Google Scholar CrossRef Search ADS   Ramiller N.C. ( 2007) Constructing safety: System designs, system effects, and the play of heterogeneous interests in a behavioral health care setting. International Journal of Medical Informatics , 76S, S196– S204. Google Scholar CrossRef Search ADS   Roke C., Cloatre E., Dingwall R. ( 2012) The regulation of nicotine in the United Kingdom: How nicotine gum came to be a medicine, but not a drug. Journal of Law and Society , 39, 1, 39– 57. Google Scholar CrossRef Search ADS PubMed  Rushton C. ( 2014) Whose place is it anyway? Representational politics in a place-based health initiative. Health and Place , 26, 100– 109. Google Scholar CrossRef Search ADS PubMed  Stoopendaal A., Bal R. ( 2013) Conferences, tablecloths and cupboards: How to understand the situatedness of quality improvements in long-term care. Social Science & Medicine , 78, 78– 85. Google Scholar CrossRef Search ADS   Shiell A., Hawe P., Gold L. ( 2008) Complex interventions or complex systems? Implications for health economic evaluation. British Medical Journal , 336, 1281– 1283. Google Scholar CrossRef Search ADS PubMed  Sénécal G. ( 2012) L’acteur social: L’expérience du quartier Villeray à Montréal. In Aubry F., Potvin L. (eds), Construire l’espace sociosanitaire. Expériences et pratiques de recherche dans la production locale de la santé . Presses de l’Université de Montréal, Montréal, pp. 141– 162. Timpka T., Nordqvist C., Lindqvist K. ( 2009) Infrastructural requirements for local implementation of safety policies: the discordance between top-down and bottom-up systems of action. BMC Health Services Research , 9, 45– 53. Google Scholar CrossRef Search ADS PubMed  Van Herzele A., Dendoncker N., Acosta-Michlik L. ( 2011) Mobilisation capacity for agri-environmental management. Journal of Environmental Management , 92, 1023– 1032. Google Scholar CrossRef Search ADS PubMed  Willis K., Daly J., Kealy M., Small R., Koutroulis G., Green J. et al.   ( 2007) The essential role of social theory in qualitative public health research. Australian and New Zealand Journal of Public Health , 31, 438– 443. Google Scholar CrossRef Search ADS PubMed  Young D., Borland R., Coghill K. ( 2010) An actor-network theory analysis of policy innovation for smoke-free places: Understanding change in complex systems. American Journal of Public Health , 100, 7, 1208– 1217. Google Scholar CrossRef Search ADS PubMed  © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Journal

Health Promotion InternationalOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off