Benchmarking national action plans on antimicrobial resistance in eight selected LMICs: Focus on the veterinary sector strategies

Background The WHO Global Action Plan on antimicrobial resistance (GAP) provides a global strategy for combating antimicrobial resistance. Context-specific national action plans (NAP) translate GAP to reflect local priorities. However, the process by which countries translate GAP into NAPs, and the resultant concordance, is not well-known. The aim of the paper is to evaluate the NAPs of eight selected low- and lower-middle income countries (LMICs) against GAP and each other to identify best practices with a focus on the veterinary sector. Methods Using the WHO GAP, and the WHO Manual for designing NAPs, we performed a policy content evaluation for: Afghanistan, Bangladesh, Ethiopia, Ghana, Nepal, Nigeria, Pakistan and Uganda. NAPs were assessed as concordant with GAP if they contained ≥80% of the recommendations. Operational and monitoring and evaluation (M&E) plans were assessed as: Specific, Measurable, Assignable, and Time-bound (or SMAT). Financing, targets and legislation for antimicrobial use reduction, and medicine quality assurance mechanisms were assessed using a constructed framework. Countries were then ranked using a scoring system to identify best practices. Results All NAPs contained ≥80% of GAP’s recommendations. Whereas Nepal’s NAP was strategic, the rest were operational and uniformly SMAT; except Afghanistan’s. The M&E plans were not all SMAT. Detailed costing and funding sources were included for only Ghana and Uganda. Quantitative target for antimicrobial use reduction was found only in Nepal’s NAP and legislation only for Bangladesh. Ghana’s and Uganda’s medicine quality assurance mechanisms were the most robust. Conclusions All NAPs were concordant with GAP. However, gaps exist in relation to M&E, diminishing the countries’ capacity to be accountable and implement corrective action if necessary. Most lacked financing plans and targets for antimicrobial use reduction. The antimicrobial quality assurances strategies are limited in most of the NAPs assessed. A mechanism by which countries can benchmark their NAP would allow identification of specific limitations and areas of best practice.

Globally, efforts to combat AMR in all sectors-human, animal, and environment, are led by the World Health Organization (WHO). In 2015, the WHO, in collaboration with the Food and Agricultural Organization and the OIE, launched the Global Action Plan on antimicrobial resistance (GAP, 2015-2019) containing five strategic objectives for AMR containment [14]. In 2016, the WHO followed up the GAP with a "how-to-do-manual" providing guidance on the development of country-level National Actions Plans (NAPs) [15]. This manual listed three core components of a NAP as: (1) a strategic plan, (2) an operational plan including a financing component, (3) a monitoring and evaluation (M&E) plan. At the United Nations (UN) high commission meeting in 2016, all countries committed to having NAPs, based on the GAP, by 2017. As of March 2019, the WHO reported 67% of the 194 UN member states had a NAP [16].
Context-specific NAPS are needed to effectively combat AMR. However, despite the launch of NAPs in many countries, little is known on how countries are translating GAP into NAPs. Most LMICs also face the problem of poor-quality or substandard and falsified antimicrobials, overall estimated at 10%-20% in the human health sector, with evidence of a similarly high prevalence in the veterinary sector in some settings [17][18][19]. There is not much information on how countries are addressing medicine quality assurance in the supply chain in the veterinary sector, especially in the context of NAPs on AMR. It has been shown that poor-quality antimicrobials contribute to AMR although the exact mechanisms are unclear [20,21]. There is a need to understand how LMICs with substantial veterinary sectors translate the GAP into NAPs. Additionally, it is necessary to know how the overall process of translating GAP into NAPs may be strengthened, if necessary, to produce context-specific policies to effectively combat AMR. While the WHO tracks self-reported plan implementation, there are scarce records of comparative content analysis of individual plans. Our manuscript seeks to fill this gap.
The aim of this paper was to perform a policy content evaluation of the NAPs of selected LMICs, with a focus on the veterinary sector. The objectives were 2-fold: (1) to benchmark policy provisions against general GAP recommendations in the veterinary sector, and (2) to identify best practices in design of NAPs from benchmarking and a cross-country comparison, with a view to providing lessons for countries designing or iterating NAPs.

Study design
The study was designed as a policy content evaluation. Two researchers independently evaluated the NAPs on antimicrobial resistance for eight LMICs. Provisions for the veterinary sector in the plans were benchmarked against the WHO GAP for concordance, as detailed in the section on Benchmarking below. In addition, a cross-country comparison of the plans was performed as described in the section titled Cross-country comparison to identify best practices.

Country selection
Eligibility criteria were low-and lower-middle income countries as defined by the World Bank with a NAP on the WHO database [22,23]. The selection criteria among eligible countries were: size of the livestock VIEWPOINTS PAPERS economy (selected countries either had high livestock (cattle, poultry, fisheries) holding by population or livestock produce, belonging in each case to the top 2 quartiles of international or regional rankings), and population (Table 1). Using these criteria, four countries each from Asia and Africa were selected: (I) Lower-middle income: Asia -Bangladesh and Pakistan; Africa -Ghana and Nigeria, (II) Low-income: Asia -Afghanistan and Nepal; Africa -Uganda and Ethiopia.
Only countries from Asia and Africa were selected because all but five low-and lower-middle income countries are in these regions. The five LMICs outside these regions are in South America and the Caribbean: Bolivia, El Salvador, Haiti, Honduras and Nicaragua, and NAPs were not available for these countries by 2018.  Benchmarking GAP: Using the GAP framework for national action, the NAP of the eight countries were benchmarked against recommended GAP actions in the veterinary sector. Overall, NAPs were assessed for concordance as previously described [38]. In brief, NAPs were assessed as compliant with GAP if they make policy provisions for at least 80% of the recommended actions for the veterinary sector for strategic objectives 1-4 in the GAP. Benchmarking was performed by a single researcher (SO).

WHO Manual:
In addition, operational and monitoring and evaluation (M&E) frameworks were assessed if they met four, out of five, SMART criteria: specific, measurable, assignable, and time-bound, as specified in the WHO Manual: the Realistic aspect was excluded because this depends largely on several factors beyond the scope of this paper. Additionally, financing information in the NAP was noted and evaluated for costing and funding source according to the WHO Manual.

Cross-country comparison
The cross-country comparison was performed by a qualitative assessment of the following parameters: 1. Targets for antimicrobial use reduction in the veterinary sector according to the recommendation of the WHO and specific targets of antimicrobial use reduction as specified in the O'Neill Report, 2015 [12,13]. 2. Legislation restricting non-therapeutic uses of antimicrobial for growth promotion, or prophylactically, including CIAs, in the veterinary sector, if any: targets without legislative backing may be difficult to enforce in certain contexts. We include legislation here as an indication of government willingness to enforce targets. This term is also used for policies restricting use in the veterinary sector for growth promotion as specified under Objective 4 of the GAP [14]. 3. Medicine quality assurance: The GAP recommends quality assurance of antimicrobial medicines in human and animal health. To operationalize the assessment of the recommendation, we used VIEWPOINTS PAPERS the medicine supply chain "framework" by Silva et al., 2019. We adapted it to describe nodes at which quality assurance mechanisms would be required [39]. Eight mechanisms covering four nodes and the presence of a National Medicines Regulatory Authority were framed as criteria for assessing the NAPs.
A scoring system was developed using a data extraction tool created on Excel in which scores were tallied for each country using the answers (Yes = 1; No = 0) to questions created for the benchmarking and cross-country comparison parameters (as described above). Countries with the highest tally for each parameter were identified as an example of best practice for the design of National Action Plans on AMR in countries with a similar context.
Evaluation for concordance with the WHO Manual and the cross-country comparisons were independently performed by two researchers (SO & IS). Results were studied for discrepancies and any differences that could not be resolved by the two were referred upwards to VJW and MHZ for resolution.

Analysis of results
Results are largely descriptive and presented in terms of the scoring system applied.

Concordance with GAP
All NAPs were found to be compliant with GAP where all plans had at least 80% of the recommended actions in the veterinary sector for nations in the GAP ( Table 2). However, most plans lacked diagnosis to guide rational prescription as a recommended action. Some countries showed more gaps in specific areas than others. For example, the NAP of Afghanistan did not include three of the recommended actions within the Infection Prevention and Control (IPC) domain. Similarly, the NAP of Pakistan did not include three actions under the Optimized use of antimicrobials domain. Despite this, within each domain, countries included the majority of GAP recommended actions for each domain.

Concordance with the WHO Manual
Operational plan: Operational plans were assessed as specific, measurable, assignable and time-bound, but for Afghanistan' s and Nepal' s (

Cross-country comparison
Targets for antimicrobial use reduction: Interventions to reduce veterinary antimicrobial use were included as plans in all NAPs, except Ghana' s. However, quantitative targets are included only in Nepal's. Half (50%) of the NAPs made no specific mention of reserve antimicrobials or CIAs as defined by WHO (11). In many cases, CIAs are included under the general antimicrobial category. Only Afghanistan, Bangladesh, Ethiopia, and Nepal expressly mention CIAs. Similarly, only 50% of the NAPs include planned interventions to ban the use of antibiotic-containing feeds.
Legislation: Legislation restricting the non-therapeutic uses of antimicrobials in livestock was found only in Bangladesh. However, there were indications of intention to make these legislations either de novo or by modification of existing legislation eg, in Pakistan. Only the NAP of Ethiopia mentions a separate veterinary agency that is assumed to monitor or ensure the quality of veterinary medicines. The NAP of Pakistan mentions a separate veterinary agency for monitoring quality, but it is not known if this body also has authority over veterinary antimicrobials.
Antimicrobial quality assurance strategies: All NAPs contained, to varying degrees, mechanisms for the quality assurance of antimicrobials ( Table 5). Only two countries met all specified criteria: Ghana and Uganda. Bangladesh and Ethiopia met all but one. Nepal' s NAP was a strategic document without detailed operational plans.

DISCUSSION
The WHO Global Action Plan on antimicrobial resistance represents the first global tool for the containment of AMR. Over time, this has been followed by several tools including the Manual for Developing NAPs [40]. A NAP, as it maps a committed plan of action agreed by stakeholders, is an essential tool in combating AMR in different national settings. There is a need for this plan to be comprehensive (ie, include all considerations for tackling AMR), and for plan implementation to be continuously evaluated, and publicly reported. This evaluation benchmarked and compared the NAP of eight selected LMICs with a significant livestock economy against the WHO' s GAP and Manual, and other parameters derived from the literature on AMR, or constructed, to identify gaps and best practices in the design of NAPs. Such an evaluation can *These are indicated as possible sources of funds. In some cases, there sources actively fund some activities, for example funding by WHO for the Antimicrobial Awareness Week programs in many countries. Inclusion here does not imply active funding by the common funding sources included here.

VIEWPOINTS PAPERS
provide useful insights for policy-makers in countries with similar settings that are designing or iterating NAPs. To the best of our knowledge, this is the first of such focused independent policy content evaluation.
This evaluation demonstrates that indeed all the studied NAPs used the GAP as a template. Without exception, all had at least 80% of all recommended actions for veterinary antimicrobial resistance containment. This was regardless of whether the NAP was a broad strategy document, as for Nepal, or included a detailed implementation or operational scheme as for all other included countries.
However, NAPs differed in terms of operational, including financial, and M&E plans as outlined in the WHO Manual; as well as in targets and legislation for antimicrobial use reduction, and medicine quality assurance mechanisms. We found that the operational plans for all countries except Afghanistan and Nepal were uniformly detailed. However, many countries have gaps in their M&E plans. Countries whose NAPs do not include an M&E framework, or whose M&E framework are not sufficiently robust could use the plans of countries such as Ghana and Pakistan as models of best practice, or could use the WHO Monitoring and Evaluation framework to design one [41]. It may also be useful if such evaluations can be made publicly available as suggested and practiced by the United States [42,43]. The WHO maintains a database of self-reported progress in the implementation of NAPs for countries [23].
Of all the countries only Nepal' s NAP contained a target for antimicrobial use reduction. In high-income countries with advanced surveillance systems, antimicrobial usage is monitored via sales or consumption data with policies designed to reduce use yielding positive results. Recently, the consumption of antimicrobials in food-production animals dropped by 32% across the European Union [44]. For most of these countries, there were/are established targets for use reduction. In Asia, the NAP of Thailand includes a target of 30% reduction in veterinary antimicrobial usage by 2021 [45]. The absence of use reduction targets in the NAP of most of the studied LMICs needs to be addressed. Targets should be included for both all antimicrobials and specifically for reserve antimicrobials or CIAs [42]. To define country-specific targets it is important to have baseline data on usage. Prospectively, countries will need to develop surveillance systems to routinely collect data on consumption.
Similarly, the lack of legislation restricting the non-therapeutic use of antimicrobials in animals needs to be addressed. It is possible that some countries such as Pakistan are in the process of doing so. It would be necessary for countries to have legislation restricting the use of antimicrobials in animals and to include these in the NAP -with incentives, alternatives and disincentives to guide farmers, veterinarians and other stakeholders. This should also be the case with antimicrobial-containing feeds. Interventions to eliminate antimicrobials from feed would also be necessary. Countries can learn from Bangladesh' s NAP which contains three legislations restricting non-therapeutic uses in animals [46].
Antimicrobial quality assurance strategies in the NAPs of the selected countries need to be strengthened. Poor-quality veterinary antimicrobials are prevalent in Ethiopia and Nigeria, for example. All countries differed on the quality score from a low score of 3 (Nepal) and 4 (Pakistan) to a high of 8 (Ghana and Uganda). The other countries had high scores of between 6 and 7. There is a need to make provisions for assuring the quality of antimicrobials throughout the supply chain. Ghana' s and Uganda' s NAP may be used as best practices in this regard. A stringent regulatory authority is one prerequisite for ensuring the quality of medicines throughout the supply chain for human and veterinary medicines [19]. Of all the studied countries, only Ethiopia had a separate agency for the control of veterinary medicines quality [47]. While the existence of a separate regulatory agency may not by itself lead to the elimination of poor-quality veterinary medicines from circulation, it does represent capacity advance in veterinary medicines regulation. A separate veterinary medicines regulatory authority also recognizes the one health approach to AMR containment.
Financing for NAP activities is poorly defined for most of the studied LMICs, with all but Afghanistan, Ghana and Uganda not including costing for proposed activities. Overall, this finding agrees with a study by Shabangu et al, 2019, who found that budgets were included in only 19% of 43 NAPs [48]. A previous study of Bangladesh showed that some veterinary-related activities are funded under other programs (the SDGs) -implying a measure of financing for these activities [38]. It should be noted though that Afghanistan like Ethiopia indicates that its proposed activities need to be prioritized before financing is decided. Thus, whilst all countries seem to recognize the need for funding NAP activities, funding decisions still need to be finalized in some countries. The lack of a clear funding structure in the NAP may be a barrier to successful implementation.
Additionally, while financing for national action plans is recognized as important, information as to total costs of implementation is scarce for many LMICs. Ghana estimated that implementation of its 5-year, 2017-2021, national action plan would cost about $21 million [49]. In the United States, funding to im-