Food systems are notoriously complex. Many actors are involved, from subsistence farmers through to multinational corporations with more economic power than many small nations, and from informal ambulant vendors through supranational bureaucracies. All of them interact with each other directly and indirectly, amplifying (sometimes, dampening) each other’s actions.
Policymakers seek to cut through this complexity, identifying and prioritising actions that lead to more favourable outcomes for the stakeholder groups that they care about. Two recent, and very different, publications move us forward on this path, both focusing on food systems actions that could improve nutritional outcomes.
Corinna Hawkes and colleagues have combed 45 expert reports on food systems written since 2013 and extracted from them the list of actions believed to lead to better nutritional outcomes. They arrived at a list of 42 consolidated actions in the domains of: agriculture; international trade; research, processing and technology; supply chain infrastructure; investment and finance; public institutions; business incentives; regulations and laws; education and public awareness, and national guidelines. For each action, the potential impact was clarified using an Impact Pathway analysis.
Nick Moore and colleagues at 3ie, on the other hand, have been back to the source literature, reviewing 1,838 impact evaluations and 178 systematic reviews. Their aim was not to estimate the strength and direction of different associations, but rather to help the food systems community understand where we do and where we do not have sufficient evidence to guide action. They look at interventions in the following domains: production; distribution and storage; processing and packaging; food loss and waste management; affordability; availability; promotion and labelling; women’s empowerment, and information and behaviour change communication. Because they required impact evaluations, they tended to exclude interventions at the level of the enabling environment, which were included in the Hawkes et al. paper. Their findings are summarised in a clear and compelling interactive evidence gap map.
What do these two excellent and highly complementary resources tell us?
Firstly, the field of food systems research is exploding. The number of relevant impact evaluations identified by the team at 3ie went from essentially zero in 2000 to nearly 2000 studies 20 years later! This means that as in other areas of science, knowledge is constantly evolving and the ability to synthesise disparate evidence is perhaps the most precious skillset that we have.
Secondly, the evidence base is very uneven. The 3ie team find, for example, that there were 300 relevant impact evaluations looking at the association between information/Behaviour Change Communication interventions and outcomes of diet quality and adequacy, versus just one looking at the impact of food loss and waste management interventions on any nutritional outcomes. Whatever you feel about the criteria than Moore and colleagues used to limit the scope of their analyses, this vast differential is food for thought, and an important caveat to the work of Hawkes and colleagues - clearly, not all of the 42 prioritised actions in that source are supported by the same base of robust evidence.
Thirdly, the complexity of food systems should not prevent us from moving forward with the identification of priority actions. In the year of the United Nations Food Systems Summit, we note remarkable consensus in the community about: what the linkages between food systems and nutrition look like (generally building on the work of the High Level Panel of Experts), what the relevant universe of possible actions includes, and where we can start even while more evidence is evolving.
At the Global Alliance for Improved Nutrition (GAIN) we aim to build on these knowledge products and add to them based on our own work with partners in the field. We are working on decision-making tools such as the Food Systems Dashboard, but also grappling with some of the hard systems issues such as how to reconcile nutritional and environmental objectives? One thing is for sure—five years from now, the evidence landscape will look very different and we will still be striving for that all-important synthesis!