By Nichola Naylor

National Action Plans (NAPs) for Antimicrobial Resistance (AMR) highlight the intended objectives a country has in relation to tackling AMR, and gives an overview of how these objectives are intended to be met over a given time period. The World Health Organisation houses a library of NAPs, available here: Library of national action plans (who.int).

There are often calls for more epidemiological and economic evaluations of interventions aimed at tackling AMR, to subsequently lead evidence-based policy making in relation to future NAPs. In order to fill this gap in relation to farm-level interventions and their links with antimicrobial usage (AMU) and the subsequent impacts on AMR-associated costs, we first must understand what are the key interventions/categories of interventions we should be evaluating. To have a more detailed understanding of the landscape of priority farm-level interventions that could relate to AMU, from the perspective of policy makers, such NAPs were reviewed for the country case studies of SEFASI. Prior to the first SEFASI Knowledge Hub meeting (see blog here), we wanted to understand what potential interventions related to farm-level antimicrobial usage policies could be found in SEFASI case-study countries. The following documents were included for review:

• For Denmark: The Danish Veterinary and Food Administration's national action plan for antibiotic resistance in production animals and food (pdf)
• For England: Tackling antimicrobial resistance 2019–2024 - The UK’s five-year national action plan (pdf)
• For Senegal: Plan d'action national multisectoriel de surveillance et de lutte contre les résistances aux antimicrobiens (pdf)

They covered overlapping but slightly different timelines, with Senegal’s previous NAP used (covering 2018-2022), the UK’s NAP which is due for an update in 2024(covering 2019 - 2024) and Denmark’s finishing in 2023 (2021 - 2023).  The specific NAP objectives that the interventions fell under, their direct description within the NAPs, the intended population of interest and mentioned measurements of impact were extracted. Whether these are related to more generic categorisations of farm-level interventions (e.g. biosecurity, vaccine, hygiene, stewardship), and whether these are considered AMU-sensitive or AMU-specific (reduced AMU indirectly or directly respectively), were then noted.

The related objectives pulled from these NAPs are available in Table 1, at the end of this blog. Note that additional information presented (such as intervention and population descriptions) might be subjective interpretation, or an umbrella-term/summary of the information provided within the NAPs - please refer to the NAP documents for specific details regarding each if in doubt.

Regarding target type, all countries had a mix of both types (Figure 1). Senegal and Denmark focus more on specific-AMU interventions predominantly, whereas the UK focuses more on AMU-sensitive interventions, in relation to AMU that's associated with farms. Box 1 presents an overview of specific interventions found across the NAPs. Examples of where an intervention would be classed as both would be if it mentions the promotion of infection, prevention and control measures and stewardship, within the same aim. 

Figure 1: NAP Objectives: Antimicrobial Sensitive or Specific?

The intended impact described in each NAP ranged from very specific and measurable, to broader AMR goals. In terms of measurements of impact, antimicrobial use was the most common measure mentioned throughout the NAPs, though the Denmark NAP did also include monitoring of knowledge and awareness, hygiene and epidemiology of AMR. The UK NAP also mentioned that understanding cost-effectiveness at the farm-level will be key. For example, Denmark goals include to “achieve a reduction of 2 per cent per year (2019–2022) in the use of antibiotics for pigs, and maintain or reduce the use of antibiotics for other livestock species. Additionally, to maintain low use for production animals of those antibiotics that are critically important for treating humans (2019 level)”, which are specific and give baseline indicators and timelines.

Whilst Senegal had more general aims, with less stated impact measures, these aims and objectives were linked to exact, defined, activities linked to each objective. These also had a clear timeline of when these were planning to be executed over the period, with set budgets for each activity. For example, for “3.3. Strengthen the institutional and Regulatory process” in relation to the sale/provision of antimicrobials, an action of organizing 20 control missions including 10 in human health and 10 in animal health per inspection of antibiotic use (breeding and fishing) was listed.

Interventions found within the NAPs

• animal housing design and build
• group management
• minimum movement and mixing
• biosecurity and hygiene
• freezing quality colostrum
• guidance
• communication
• establishment and promotion of best practice
• animal health planning interventions
• promote vaccine use
• gene therapies
• nutrition
• targeted interventions based on animal disease data
• training; teaching
• code of practice
• voluntary targets
• 6-week structured programme for targeted antibiotic use + biosecurity
• vaccine
• benchmarking
• regulation on sales and market authorization
• behaviour change
• regulation on antibiotic sales (prescription)
• r&d initiatives
• access initiatives

Although more general “infection, prevention and control” interventions, which we deemed AMU-sensitive for this review, have the potential for reducing drug resistant infections and antibiotic use, there is still lots unknown on the comparative and combined impacts of sensitive and specific interventions. NAPs will be implementing lots of things at once, making the quantification of specific impacts difficult using real-world evidence, though we hope to fill some of these gaps through SEFASI research. This can be seen in our hypothetical, scenario testing approach in our recently published paper “AHHME: A model for estimating the holistic cost-effectiveness of antimicrobial resistance interventions in food animal production” available here https://www.sciencedirect.com/science/article/pii/S2352771423001490.

Whilst the NAP development continues across these countries for the next round, as does research into the potential impact of interventions, it is key that we feed both of these streams into each other. In SEFASI we are trying to do this with the continued involvement of our policy and research networks and our Knowledge Hub members. Look out for the next meeting (11th January) and subsequent blogs early next year!

Table 1. A Summary of farm-level antibiotic usage objectives in SEFASI-related NAPs 

By Joshua Aboah and Michel Dione

Antimicrobial resistance (AMR) is a global public health threat that necessitates comprehensive strategies tailored to specific contexts. Addressing the complex drivers of AMR requires the collaborative participation of all stakeholders involved in the value chain. Globally, the poultry production sector is a high consumer of antibiotics and high levels of AMR have been found in bacteria isolated in poultry farms [1]. Yet, in Senegal, the poultry industry plays a vital role in food security and economic development. A recent article shows that there is evidence that awareness raising and biosecurity improvements could encourage prudent use of antibiotics, and that biosecurity and vaccination could to some extent replace antibiotic use as productivity-enhancing and disease management tools in broiler farms. Finally, issues of farm antimicrobial stewardship must be considered at the structural level, with farm behaviours contingent on interaction with state and private stakeholders [2].

Therefore, addressing AMR challenges by reducing antibiotic use in the poultry sector could potentially be done with little impact on production and could go a long way to supporting long term food and health security of the country. However, understanding how to combine this information to effectively combat AMR within the poultry sector is complex. System dynamics modelling has emerged as a useful tool to fill this gap. This blog post highlights how system dynamics modelling is being applied to understand and mitigate AMR in Senegal's poultry value chain.

System dynamics modelling provides a holistic and dynamic approach to understanding complex systems and their interdependencies. By integrating various factors, feedback loops, and causal relationships, system dynamics models simulate the behaviour of a system over time and identify leverage points for intervention. This modelling approach facilitates a participatory model building process, allowing stakeholders to co-conceptualise problems and highlight the various interactions that contribute to systemic issues.

As part of the SEFASI project’s research activities, a two-day workshop was organised by scientists from the International Livestock Research Institute’s epidemiologist led by Dr Michel Dione, in Senegal in October 2022. The workshop aimed to structurally co-design the model structure to capture all parts of the poultry value chain with key stakeholders. Using a group model building process, the stakeholders were guided to map the pathways for AMR, identify the unintended consequences of activities to tackle AMR by key chain actors and propose strategies to combat this challenge. This collaborative workshop provided a platform for stakeholders to come together and contribute their expertise in conceptualising and designing a model to help address AMR [3]. The output from the workshop will be the foundation for developing a quantitative model and a user-friendly web-based interface that will serve as a decision-making tool for stakeholders.

System dynamics modelling enables the simulation of the effects of interventions and policies on AMR dynamics within poultry value chains. By quantifying potential outcomes, stakeholders can make informed decisions regarding the efficacy of different intervention strategies in mitigating AMR before implementing them. This proactive approach empowers stakeholders to optimise policies and allocate resources effectively.

In conclusion, it is worth reiterating that the threat of AMR in Senegal’s poultry value chain demands a multi-faceted approach that harnesses innovative tools and strategies. By utilising the system dynamics modelling approach, stakeholders in Senegal’s poultry industry can develop targeted strategies to reduce antimicrobial use, improve farming practices, enhance veterinary services, and promote responsible antimicrobial stewardship. Moreover, by optimising policies through scenario testing and evaluation, the stakeholders are able to evaluate the trade-offs between public health, economic considerations, and food security objectives. By embracing this modelling approach with early engagement and co-design, stakeholders in Senegal’s poultry industry can work together towards sustainable practices, ensure food safety, protect public health, and safeguard the future of poultry production in Senegal.

References

[1] Nhung NT, Chansiripornchai N, Carrique-Mas JJ. Antimicrobial Resistance in Bacterial Poultry Pathogens: A Review. Front Vet Sci. 2017 Aug 10;4:126. doi: 10.3389/fvets.2017.00126. PMID: 28848739; PMCID: PMC5554362.

[2] Emes E, Faye A, Naylor N, Belay D, Ngom B, Fall AG, Knight G, Dione M. Drivers of Antibiotic Use in Semi-Intensive Poultry Farms: Evidence from a Survey in Senegal. Antibiotics (Basel). 2023 Feb 24;12(3):460. doi: 10.3390/antibiotics12030460. PMID: 36978328; PMCID: PMC10044536.

[3] Aboah J, Ngom B, Emes E, Fall AG, Seydi M, Faye A and Dione M (2023) Mapping the effect of antimicrobial resistance in poultry production in Senegal: an integrated system dynamics and network analysis approach. Front. Vet. Sci. 10:1189109. doi: 10.3389/fvets.2023.1189109

By Gwen Knight

The first Knowledge Hub meeting of the SEFASI project took place on the 1^st March 2023. In this meeting we held discussions in English and French with Knowledge Hub members on a variety of questions to inform the direction of SEFASI decision making around analysis and modelling. To do this, we had a survey that we went through in the meeting and asked members to fill in in their own time with their opinions on key questions. A big thanks to all participants and respondents!

The majority of the 36 survey respondents were based in Senegal (13) or the UK (12), reflecting two of the key settings for the SEFASI project. A similar spread of participants was seen in those attending the meeting. We had a good spread of expertise, with the majority reporting expertise in epidemiology, farming/agriculture, veterinary or microbiology with a good spread across both human and animal health in both countries. We were biased to most of the participants being from academia, but did have NGO representation in Senegal. Interestingly, 21 of the respondents said they worked in the “One Health” sector.

To inform our data analysis, the first questions asked about key One Health bacteria and antibiotics to focus on. Across Senegal and the UK, there were commonalities in that E. coli and Salmonella came up as the top mentioned bacteria, and fluoroquinolones and the beta-lactams as the top mentioned drugs. However, there was a huge range of mentioned bacteria and drugs, and it is hard to know the denominator for comparison.

For the model scenarios, it would be key to know baseline resistance trend scenarios. During the meeting the inability to put a number on this was discussed and it came across in the free text from survey respondents. However, we had a clear result that a decrease would be unlikely (Figure 1). One of the aims of SEFASI is to raise awareness of modelling and the complexity of capturing intervention impact - we hope that questions like this, that are difficult to answer, increase the awareness of the community about the kind of assumptions that models have to make in order to evaluate interventions.

Figure 1: Knowledge Hub responses to likely future trends in resistance over the next 10 years

For interventions, we asked each country about the AMU-specific and AMU-sensitive interventions in their National Action Plans [see other blog for details of the interventions]. Here we found that none were assessed as being harmful but that “Voluntary Usage Targets” and “Surveillance data and benchmarking” or “Research and Development into alternatives to antibiotics” were dominated by “not effective” or “neutral”. Those where “very effective” or “effective” dominates were deemed to be “Infection, prevention and control”, “More targeted antibiotic use” and “behaviour change”. Interestingly, “surveillance data and targeted health interventions” was deemed “effective” whilst only benchmarking was neutral suggesting that in England, the assessment is that merely reporting data is not enough and that the data has to guide specific interventions. In Senegal, similarly none were harmful and all were seen as effective or very effective.

Finally, we asked about understanding what output measures we should evaluate intervention impact on. We had a very diverse selection here, with no clear priorities in the One Health setting in either the UK or Senegal, and a similar distribution across the two countries though the majority of the Senegal responses set “Antibiotic Residues” being key. There was a substantial focus on animal welfare and mortality as well as infection incidence in humans and antibiotic consumption. Economic indicators that ranked the highly were the monetary cost-benefit and individual income effects, whilst GDP was not often selected.

Conclusions

Next steps for Knowledge Hub development

The Knowledge Hub is currently biased towards academic partners, with some NGO connections from Senegal, and few participating members from Denmark. These biases need to be addressed to ensure we have good prior information for our parameters and analysis going forward.

We asked some potentially “unanswerable” questions (e.g. on future trends) and had this feedback from participants. In future surveys we need to ensure that we use these as learning tools: to predict intervention impact we have to know something about what happens if we don’t intervene. This could potentially be addressed by motivating examples from the modelling process.

Our experts did not clearly differentiate the interventions in terms of their effectiveness, suggesting that we have to explore all interventions from the NAPs. However, we have to reach a balance between model complexity to account for variation in interventions whilst being able to capture a range. Many of the impacts are already included in AHHME, but we don’t have an environment component so will have to work on adjusting this angle in our work.

Next steps for SEFASI

Our Knowledge Hub participants provided key inputs for moving our SEFASI research forward such as key pathogens and antibiotics to focus on, and trends in resistance. These have been integrated into the data analysis and guide our decision making as to intervention analysis.

By Ross Booton

Dr. Ross D. Booton previously worked on antimicrobial resistance (AMR) in Thailand from a One Health perspective, where human health is co-dependent on animal health and the environment. He did this with colleagues from the One Health drivers of AMR in rural Thailand (OH-DART) group.

He used a novel mathematical model framework of gut colonisation with extended-spectrum beta-lactamase (ESBL)-producing bacteria and transmission to project the reduction in human antimicrobial resistance over 20 years (2020-2040) for each One Health driver, including individual transmission rates between humans, animals and the environment, and to estimate the long-term impact of the Thai National Strategic Plan (2017) on Antimicrobial Resistance intervention.

His model predicts that human usage of antibiotics was the most important factor in reducing the colonisation of humans with resistant bacteria. The Thai National Strategic Plan on AMR was projected to reduce human colonisation by 6.0-18.8%, with more ambitious targets (30% reductions in human antibiotic usage) increasing this to 8.5-24.9%.

This model provides a simple framework to explain the mechanisms underpinning antimicrobial resistance, suggesting that future interventions targeting the simultaneous reduction of transmission and usage of antimicrobials would help to control antimicrobial resistance more effectively in Thailand.

For more information: Booton RD, et. al. One Health drivers of antibacterial resistance: Quantifying the relative impacts of human, animal and environmental use and transmission. One Health. 2021. 10.1016/j.onehlt.2021.100220.

Application to SEFASI

Ross will use this model framework to predict the potential impact of One Health interventions in England, Denmark, and Senegal. He plans to develop his model and parameterise to each of England, Denmark, and Senegal’s specific antimicrobial resistance settings. Together with collaborators in these 3 countries, he will estimate the value of intervention packages that relate to human, animal, or environmental sources. He will do this in the context of E. coli resistance to third generation cephalosporins. This theoretical modelling work could help understand which intervention would yield the greatest impact in each country and help decision makers to effectively plan national strategic action plans for AMR in the immediate future.