Towards a critical complex systems approach to public health

There are growing calls for public health research to eschew linear models of cause and effect typical of clinical research, and instead to embrace approaches that recognise the manifold, multi-layered and interdependent processes that produce patterns of poor health (Rutter et al., 2017). This recent ‘complexity turn’ in public health is overdue, desperately needed and to be welcomed. However, there are some real risks that limit its potential to realise significant advances in our understanding of, and action on, health inequalities. One is the enduring gap between knowledge and implementation. As Grant and Hood (2017) note, in their paper in this issue, ‘[m]uch of the literature … is already known to public health experts but does not seem to be regularly used’. Public health researchers and professionals largely understand that the causes of ill health are complex and structural, and that evidence for how to address inequalities will require a range of different methods. Yet an overview of health promotion campaigns and the published journal evidence on ‘what works’ would suggest we remain deeply wedded to linear models and individualistic interventions. At the sharp end of public health practice, there are of course well-rehearsed difficulties of translating system-level understanding into what we do with clients and communities. An example comes from Powell et al. (2017) in this issue. They found that even where health promotion staff have sophisticated understandings of the social determinants of health, institutional structures can force a ‘drift’ back into individualistic framings of health.

More significantly, though, we suggest some risks to the success of the ‘complexity project’ from the marginalisation of both critical scholarship and community voices in public health.

First, there is the danger that many bodies of relevant theory that could usefully guide system analysis are overlooked. A characteristic of complex systems, which advocates of this approach are often at pains to point out, is their ‘emergent’ nature. Rutter et al. (2017), for example, note that ‘emergence describes the properties of a complex system that cannot be directly predicted from the elements within it and are more than just the sum of its parts’. This has been long recognised in decades of social and political theory that identifies repeating patterns at the level of systems. So, while the specifics of how things play out will vary across time and place, there are nevertheless important ‘demi-regularities’ and often predictable winners and losers. De Maio et al. (2017), in this issue, demonstrate how important it is to both look at the system level, and the particularities, to unpack causal effects. They analyse the very different associations between racial/ethnic minority segregation and low birth weight across two contrasting cities – Chicago and Toronto. This analysis demonstrates forcefully the need to move beyond ethnicity as if it were an attribute of individuals, and towards a more complex appreciation of racism, as an attribute of systems. Integrating social and political theory into complex systems work in public health should increase efficiency and ensure that key system elements and processes that (re)create health inequalities are clearly in view.

Social science has a long history of exploring how micro-, meso- and macro-level system structures and processes interact to explain the patterning of poor health and wellbeing and the reproduction of inequalities over time. Using social and political theory should help to move beyond ‘programme theory’ that is limited to explicating the components of interventions and how they interplay with immediate contextual characteristics, to a more holistic understanding of how deliberate action unfolds on the ground and ripples out producing both anticipated and unexpected effects; in Penny Hawe et al. (2009) words, to understand ‘events in systems’.

The second risk we identify relates to the large body of systems-informed public health practice that is already ongoing, but often under-reported and marginalised in writing about complexity. The evidence on offer for public health practice in leading journals does indeed continue to be dominated by linear thinking and controlled designs, but there are many missed opportunities to document and share learning about systems from real-world public health practitioners and local communities. As just one example, Social Action for Health’s campaign ‘Does Hackney Make you Fat’? (Social Action for Health [SAH], 2017) shows how a campaign can address the social determinants of health in an inclusive way, recognising that the proper level of action is not the individual but the broader system, yet working with communities and individuals in inclusive and respectful ways. Such embedded organisations exemplify the process of ongoing effort to learn about the system and adapt their approaches to secure positive outcomes for local communities.

Third, there is the risk that the system boundaries are mis-specified and that key elements – the evidence-generators and policy-makers – are left out. Rather than presenting evidence generation and application as distinct from, they must be seen as integral to, these complex health-patterning systems. Our systems analyses must be alert to: how public health problems are defined; how research questions are framed; what types of evidence are demanded; and what response options are on the table. The lack of attention to wider societal processes, and predominant focus on individual ‘life-style’ behaviours, as causes of health inequalities in recent years is not simply a product of medical models of evidence generation. Rather, this epistemological stance has tended to coalesce with an ideological position that locates the roots of disadvantage with individual traits and diverts attention away from policy solutions that are unpalatable to those in powerful positions. As Grant and Hood (2017) note on the inter-linkages between policy and research, the ‘mutually beneficial relationship between researchers and policy-makers’ can mean explanations of system processes remain partial and superficial, as complexity and contingency are inevitably rather more challenging to present than simple linear causes.

Heading off these three dangers should help to produce a more critical approach to complex systems research and practice in public health. Renewed calls for systems approaches to public health are energising. To ensure they translate into meaningful policy and action on the social determinants of health, we must: harness insights of generations of social and political science on structure; ensure that our evidence-based is informed and enriched by community action as well as published articles; and take a critical view on the system, including and interrogating ideological and political actors.

Sarah Salway
School of Health & Related Research, University of Sheffield, UK
[email protected]
Judith Green
School of Population Sciences & Health Services Research, King’s College London, UK
[email protected]