Supporting Evidence-Based Policy: a longitudinal study of AMR risk behaviours among livestock keeping communities in India and Kenya

Anti-microbial resistance (AMR) is set to rival climate change as one of the largest barriers to health and wellbeing in the 21st century (Aarestrup, 2015; O'Niel, 2015; Woolhouse et al., 2015). And like climate change, AMR is likely to have a preferential impact on communities of the poor (Heffernan, 2017). Yet, to date, the inter-relationship between poverty and AMR has not been well explicated. This is particularly true with regard to animal agriculture. Indeed, while 2/3rds of the global poor are livestock keepers (Lin and Heffernan, 2009a, 2010) very little is known about the factors facilitating AMR among this population. We do know however, that AMR emerges out of the social, cultural and behavioural milieu in which disease and healing is situated. And in this case the environment in which livestock are produced. Therefore, the aim of the project is to explore three inter-related drivers: social, behavioural and environmental to the emergence of AMR among two communities of the livestock dependent poor: Maasai pastoralists in Kenya and subsistence dairy producers in Orissa, India. Both of these communities had participated in a large-scale study on animal healthcare service delivery funded by DFID in 1999 (Heffernan and Misturelli, 2000; Heffernan, 2001). In this manner, the research will draw on a historical dataset to compare a variety of parameters important to the emergence of AMR. Ranging from animal healthcare seeking behaviours/preferences to livestock-based livelihoods to access to service providers to farmer of understanding disease etiologies and access to livestock pharmaceuticals and antibiotic use, including dosages and timing over a 17-year period. Thus the study offers a unique opportunity to explore behavioural change over time. Such an historical assessment is beneficial on two levels: first, the strength and directionality of these forces on AMR emergence can be measured and thereby, better understood and second, the analysis offers a unique opportunity to support evidence-based decision-making.

The risk of the rapid ascendancy of AMR as a global threat is that the development community will formulate policies on a weak evidential base. Equally problematic, much of the current research exploring AMR focuses on bio-prospecting in one-off exercises. Therefore, it is likely that this approach will create an evidence base biased towards single outcomes. And in this manner, related decision-support tools will be limited in their ability to predict change by the singular, deterministic nature of the underpinning data (Heffernan and Yu, 2010; 2007; Yu and Heffernan, 2009). To address this issue, the project will create and deliver the 'AMR forecaster' an easy to use ranking and weighting framework based on the longitudinal dataset. The aim is to enable policy makers, researchers and practitioners to use the simple on-line tool to assess the relative risk of AMR emergence among the communities involved.

Finally, however, we know that changing behaviour regarding the use of antibiotics is imperative both in the global North and South (O'Neil, 2015; Woolhouse et al., 2013; 2015). Previous studies have shown that children can be effective entry points for livestock-related knowledge at the community level (LDG, 2011). Therefore, the project team will produce and disseminate learning material on AMR to local school children, and measure the transfer of key messages or 'memes' on wider households members.

The project has two sets of beneficiaries: direct and indirect. Equally, in each of the countries involved, project benefits will be derived on three levels: community, household and institutional.

The direct beneficiaries, at the community level, include the livestock-keeping households, who will participate in the project. Households will benefit from an increasing awareness of the cost of AMR on their livelihoods and the behavioural shifts required to prevent it. Children of these households will further benefit from the school outreach element. Indeed, by engaging schools in the project, wider impacts at the community level are likely to accrue over years to come. Further, by measuring the ability of children to foster new knowledge pathways at the household level with the consequent ability of AMR 'memes' to flow across these pathways, the project has the potential to alter development praxis. Local NGO and CBO partners will also gain by alternately engaging with vs. driving the iterative and reflexive learning processes fostered by the project. It is anticipated that a greater understanding of the behavioural drivers to AMR among this group will be used to inform future projects and programmes. Indirect beneficiaries include the wider population of community members and school-going children who will be exposed to both project findings and learning materials.

Across both countries, 30 communities and 300 households will directly participate in the study. Further, within each of the study sites, 10 primary schools will be targeted for engagement activities. As such, the total population of community-level beneficiaries is estimated to be 6,500 individuals. Project activities are estimated to indirectly impact an additional 3,000 individuals in each of the countries involved. Therefore, the total number of direct and indirect beneficiaries at the community level is estimated at over 9,000 individuals.

Direct beneficiaries at the institutional level are the policy makers and planners who will utilize the decision-support tool. Use of the forecasting tool will better support evidenced based policy-making and resource allocation at the institutional level. In this manner, national-level gains will be accrued by relevant and targeted interventions to decrease the emergence of AMR. Equally, armed with this knowledge, national-level policy makers will ultimately be able to better inform donor-led projects and programmes of the social and behavioural forces impacting AMR among the poor. Policy change and related impacts however generally occur over a longer time period than the 24 months of the pump-priming grant. Thus, future policy-level benefits, while expected, are largely outwith the ability of the project team to specifically define and measure.

By disseminating project results via a variety of AMR networks, the project findings will reach decision-makers at the global level. Currently UN agencies such as WHO, OIE and FAO are leading global action against AMR. However, as noted above, most of the activities are focusing on commercial livestock systems, providing critical evidence of a crucial but largely overlooked population, has the potential to raise issues related to poverty and AMR up the agenda for global action.