An integrated surveillance network for antimicrobial resistance, India

Abstract Objective To assess the preparedness of veterinary laboratories in India to participate in an integrated antimicrobial resistance surveillance network and to address gaps in provision identified. Methods The Indian Council of Medical Research and the Indian Council of Agricultural Research collaborated: (i) to select eight nationally representative veterinary microbiology laboratories whose capacity for participating in an integrated antimicrobial resistance surveillance network would be assessed using a standardized tool; (ii) to identify gaps in provision from the assessment findings; and (iii) to develop a plan, and take the necessary steps to address these gaps in consultation with participating organizations. Findings The main gaps in provision identified were: (i) a lack of dedicated funding for antimicrobial resistance surveillance; (ii) the absence of standard guidelines for antimicrobial susceptibility testing; (iii) a shortage of reference strains for testing and quality assurance; and (iv) the absence of mechanisms for sharing data. We addressed these gaps by creating a veterinary standard operating procedure for antimicrobial susceptibility testing, by carrying out a validation exercise to identify problems with implementing the procedure and by conducting capacity-building workshops for veterinary laboratories. Conclusion Antimicrobial resistance surveillance networks depend on the availability of accurate, quality-controlled testing. The challenges identified in creating an integrated surveillance network for India can be overcome by developing a comprehensive plan for improving laboratory capacity in human, veterinary and environmental sectors that is supported by the necessary funds. The study’s findings may provide guidance for other low- and middle-income countries planning to develop a similar network.


Introduction
The misuse and overuse of antimicrobials in both humans and animals are major drivers of antimicrobial resistance. [1][2][3] Several clinically important antibiotics are used extensively in food-producing animals, either for metaphylaxis, prophylaxis or promoting growth, 4,5 and these animals and their food products are recognized as prominent routes of human exposure to foodborne pathogens. Moreover, there is a risk that resistant microbial strains or genes will be transmitted to humans and enter the environment. 5,6 Given the lack of antimicrobial resistance surveillance systems in the veterinary sector in low-and middle-income countries, it is difficult to quantify the contribution antimicrobial use in animals makes to the emergence of drug-resistant pathogens. The establishment of surveillance systems, as part of a holistic One Health approach to public health, is therefore key to understanding the transmission of drug resistance between different health sectors and for designing interventions. Antimicrobial resistance surveillance in animals is generally inadequate both regionally and globally. 7,8 Apart from agencies like the World Health Organization and public health bodies in Canada, the European Union, Norway and the United States of America, few organizations or countries have successfully created integrated antimicrobial resistance surveillance systems that support comprehensive antimicrobial stewardship programmes in both animals and humans and that help control antimicrobial usage and its impact on human health. 9,10 In India, the 2017 National Action Plan stated that strengthening knowledge and evidence of antimicrobial resistance through surveillance was a strategic priority. 11 Currently, however, surveillance in the country is limited to human health and there has been little progress towards expansion into animal or environmental health. 12 In India, there are 17 research institutes and 19 universities operating under the animal sciences division of the Indian Council of Agricultural Research but only a handful are engaged in research into antimicrobial resistance. In 2016, the Indian Council of Medical Research entered into an agreement with the Indian Council of Agricultural Research to support collaborative research on areas of mutual interest, including antimicrobial resistance. 13 Both organizations recognized the importance of strengthening research in, and the surveillance of, antimicrobial resistance in humans and animals and took the first steps towards developing a plan for integrated surveillance.
The aim of our study was, for the first time, to assess the preparedness of veterinary laboratories in India to participate in an integrated antimicrobial resistance surveillance network. We carried out a systematic assessment of laboratories and undertook capacity-building to address gaps in provision (Fig. 1).

Methods
We conducted a cross-sectional survey of antimicrobial susceptibility testing at eight veterinary institutes and universities in India. Participating laboratories, which represented

Preparedness
To assess laboratories' preparedness to participate in a national antimicrobial resistance surveillance network, we used the United Nations Food and Agriculture Organization's (FAO) laboratory mapping tool for antimicrobial resistance, which is part of the FAO Assessment Tool for Laboratories and AMR Surveillance Systems (FAO-ATLASS). 14 This tool is designed to assess individual laboratories in the food and agriculture sectors in: (i) funding availability; (ii) workflow management; (iii) the availability of appropriate standard operating procedures; (iv) collaboration and partnerships with other national and international laboratories; (v) relevant publications; (vi) opportunities for staff training on antimicrobial susceptibility testing; (vii) the management of biological materials, including the mode of sample identification and tracking; (viii) the archiving of bacterial isolates and the type of inventory; and (ix) preservation methods. A score for laboratory capacity (from 0 to 100%) was derived automatically for each veterinary laboratory from the completed assessment form, with a score above 70% indicating that laboratory capacity was sufficient for participation in antimicrobial resistance surveillance.

Standard operating procedure
Standardized laboratory practices are essential for accurately interpreting antimicrobial resistance data. In cooperation with the Indian Council of Agricultural Research and participating institutions, therefore, we developed an action plan and strategic framework for improving the capacity of veterinary laboratories to be implemented between August 2018 and September 2019. There were three steps: (i) preparation of a veterinary standard operating procedure; (ii) validation of the procedure by monitoring its implementation; and (iii) training laboratory personnel.
The veterinary standard operating procedure was devised through consultations with veterinary clinicians and microbiologists with the goal of enabling antimicrobial resistance patterns in animals and humans to be compared. The overarching aims were: (i) to standardize, and ensure the accuracy of, antimicrobial resistance data; and (ii) to ensure that veterinary isolates are appropriately tested for drugs commonly used in veterinary practice and that comparisons can be made directly with drug susceptibility breakpoints (which indicate whether or not a bacterial species is susceptible to antibacterials) in humans for the purpose of integrated surveillance.
The resulting veterinary standard operating procedure itemizes methods for bacterial isolation, for identifying species using biochemical tests and for the molecular characterization of five pathogens of public health importance. It also specifies antimicrobial susceptibility testing method and gives updated data on breakpoints for different antibiotics important for animal and human health. 15,16 To facilitate data collection for the purpose of integrated surveillance, we harmonized breakpoints with standard operating procedures for bacteriological assessment of human clinical samples. The incorporation of breakpoints from the Clinical and Laboratory Standards Institute's veterinary guidelines ensured that veterinary specimen isolates were interpreted correctly and were, therefore, useful for veterinary clinical practice. 17 This standard operating procedure has been accepted by the Indian Council of Agricultural Research and the use of such standard operating procedures is now obligatory for institutions belonging to the Indian Network for Fishery and Animals Antimicrobial Resistance, 18 which was established by the Indian Council of Agricultural Research.

Validation
We performed a validation exercise for the veterinary standard operating procedure to identify challenges faced by laboratories in implementing it and to address these challenges before the procedure document was finalized. Four laboratories with high scores on the preparedness assessment volunteered to participate in two rounds of validation. Laboratories were instructed to perform antimicrobial susceptibility testing and interpretation according to the procedure. In the first round, five isolates each of five pathogens were provided in 25 silica gel pouches in a blinded manner. The five pathogens, which were selected by the Indian Council of Medical Research and the Indian Council of Agricultural Research because of their importance for public health and evidence of potential transmission, were: Escherichia coli, Klebsiella spp., non-typhoidal Salmonella, staphylococci (four Staphylococcus aureus and one coagulase-negative staphylococcus) and enterococci. Four reference strains were also provided: E. coli ATCC 35218, Klebsiella pneumoniae ATCC 700603, S. aureus ATCC 25923 and Enterococcus faecalis ATCC 51299. For all five pathogens, laboratories received a list of six antibiotics for antimicrobial susceptibility testing. In the second round, the same five pathogens were provided in 13 isolates: two isolates each of E.coli, Klebsiella, non-typhoidal Salmonella and enterococci and five staphylococci isolates (three S. aureus and two coagulase-negative staphylococci).
Laboratories were scored on: (i) antimicrobial susceptibility testing; (ii) minimum inhibitory concentration findings; and (iii) susceptibility interpretation errors. Two points were awarded for correct organism identification: 1 point each for genus and species identification. For susceptibility testing, a maximum of 2 points were awarded per drug for the six antibiotics provided (maximum: 12 points). In the evaluation of antimicrobial susceptibility testing, the number of drugs a laboratory used to test each isolate was also considered (e.g. if a laboratory tested only four drugs, the accuracy of interpretation of susceptibility testing for a single isolate was calculated from a maximum of 8 points; four drugs by 2 points each). Susceptibility interpretation errors were recorded as no error (score: 2) or a minor (score: 1), major (score: 0) or very major (score: -1) error. A minor error occurred when a susceptible or resistant isolate was reported as intermediate or an intermediate isolate was reported as susceptible or resistant. A major error occurred when a susceptible isolate was reported as resistant and a very major error, when a resistant isolate was reported as susceptible. Following the validation exercise, we organized workshops for the veterinary institutes and universities to improve performance.
Integrated antimicrobial resistance surveillance, India Sonam Vijay et al.

Preparedness
The strengths and weaknesses of the eight veterinary laboratories, as evaluated using the laboratory assessment tool for antimicrobial resistance, are reported in Table 1. Scores were awarded for four areas: (i) laboratory activities; (ii) technical factors; (iii) management of data and biological materials; and (iv) quality assurance. The mean score of the laboratories across 11 categories in the four areas was 60.4% (range: 59.5-84.2). Most laboratories were strong in workflow organization (mean score: 87.5%; 95% confidence interval, CI: 77.0-97.9), antimicrobial susceptibility testing method (mean score: 80.2%; 95% CI: 72.7-87.7) and the management of biological materials (mean score: 73.3%; 95% CI: 68.4-78.2). Six of the eight laboratories reported ample availability of resources (e.g. media, reagents, equipment and facilities) for the identification and susceptibility testing of a broad range of bacterial species: they scored over 70.0%. In addition, these six laboratories routinely performed molecular characterization for over 30.0% of resistant isolates. Two laboratories (laboratories D and H) sequenced resistance genes from more than 100 isolates annually. Overall, the mean score for molecular characterization was 77.0% (95% CI: 63.3-90.7). The laboratories' scores for data management varied widely from 55.6% to 100.0% (mean: 77.8%; 95% CI: 65.4-78.2). The assessment also found there was no mechanism for data sharing with a data collection unit: only four of the eight laboratories shared data and this sharing was irregular and partial.
Five laboratories scored under 34% for the sustainability of their antimicrobial resistance assessment activities (mean: 50.0%; 95% CI: 20.4-79.6). Six of the eight scored 50% or less for documentation in the quality assurance area due to the absence of standard operating procedures (mean: 50%; 95%

Standard operating procedure
In the first validation round for the veterinary standard operating procedure, two of the four laboratories examined (laboratories A and D;  For each culture, genus identification and species identification accuracy were combined. c The accuracy of each laboratory's interpretation of susceptibility testing was calculated as follows: (i) a maximum of 2 points was awarded for each drug tested against each bacterial isolate, where interpretation errors were categorized as no error (2 points) or a minor (1 point), major (0 points) or very major (-1 point) error; and (ii) the total number of points awarded was expressed as a percentage of the maximum possible score if all drugs tested against all isolates by that laboratory were interpreted correctly. For example, if a laboratory tested six drugs against 25 cultures, the maximum possible score would be 6 x 25 x 2 = 300 points.
Note: We evaluated the four laboratories' ability to implement the veterinary standard operating procedure in two validation rounds.
Integrated antimicrobial resistance surveillance, India Sonam Vijay et al.
cussed and solutions were suggested by the experts. A second validation round was recommended. All four laboratories showed substantial improvements across all parameters in the second round ( Table 2). The laboratories correctly identified Salmonella spp. and coagulase-negative staphylococci using the veterinary standard operating procedure. In addition, three of the four laboratories scored 90% or higher for the interpretation of antimicrobial susceptibility testing findings, compared with 71.2-88.3% in first round ( Table 2).

Training workshops
Having identified the challenges faced by veterinary microbiologists during the validation exercise, training workshops were organized to improve antimicrobial susceptibility testing by ensuring it conformed with the veterinary standard operating procedure. The workshops were attended by two veterinary microbiologists and two laboratory staff from each of the eight participating institutes. Workshops took place at two sites to avoid overcrowding: the microbiology departments of the Christian Medical College, Vellore, and the Postgraduate Institute of Medical Education and Research, Chandigarh, respectively.
The 3-day workshops were held in September 2019, covered both theoretical and practical aspects of antimicrobial susceptibility testing and included a 2-day, hands-on workshop on bacterial identification and antimicrobial susceptibility testing as stipulated by the operating procedure. The workshop curriculum covered: (i) choosing the correct combination of drug and microbe; (ii) selecting the appropriate testing method; and (iii) guidance on the accurate reading and interpretation of susceptibility testing data. Participants were evaluated by questionnaire before and after training. Questionnaire findings showed that the proportion of participants who understood pathogen and species identification on antimicrobial susceptibility testing improved from 25% (55/220) before training to 60% (132/220) after.

Discussion
Surveillance of antimicrobial resistance in line with the One Health concept requires the creation of integrated national networks. Most countries have pledged to set up these networks but are facing their own individual challenges. 19 Researchers have suggested that the most common challenges are: (i) inadequate funding; (ii) poor communication between agencies; and (iii) limited participation of human and animal experts, including those in environmental health. 20 Our collaborative study identified several factors that may hamper the creation of integrated surveillance networks in India and other low-and middle-income countries: (i) a lack of dedicated funding for antimicrobial susceptibility testing; (ii) a lack of trained staff; (iii) the absence of essential quality control measures; and (iv) a failure to take into account updated guidelines and drug susceptibility breakpoints. These factors, which have been highlighted by other researchers, 21 affect the quality and accuracy of antimicrobial resistance data and the sustainability of surveillance. Most work on antimicrobial resistance at laboratories in our study was heavily dependent on short-term or project funding from national and international agencies. No funding for antimicrobial resistance surveillance had been received from the Indian Council of Agricultural Research or the government. These financial constraints can limit laboratory capacity, thereby compromising data quality. In addition, the constrained availability of resources was also reflected in a lack of essential reagents and a shortage of the reference strains needed for quality assurance.
As we recognized that the lack of a veterinary standard operating procedure for antimicrobial susceptibility testing was a major deficiency in India, we developed an operating procedure for priority pathogens that incorporated national susceptibility breakpoints. 15 In our validation exercise to determine how easily laboratories could adopt the procedure and how accurately they could apply it, all four laboratories admitted they did not follow it when identifying pathogens or interpreting the results of antimicrobial susceptibility testing. Instead, laboratories chose to follow previous procedures (e.g. by using the disk diffusion test for colistin instead of the recommended broth microdilution method), which indicated behavioural nonconformance. During the first validation round, we observed a reluctance to adopt the new procedure, which could obstruct its implementation across the country. Laboratories may, therefore, need to be convinced of its advantages. We regarded the limitations we identified through this assessment as an opportunity for improvement rather than as a setback. After laboratory staff met experts, who explained the importance of following standard operating procedures, all laboratories performed as well or better on all measures in the second validation round, which illustrates the importance of understanding laboratory staff 's concerns and explaining the usefulness of following new procedures. The failure to correctly identify susceptible or resistant isolates in the second round or to mistakenly interpret breakpoints, thereby leading to erroneous antimicrobial susceptibility testing results, indicated that laboratory staff had received inadequate training.
Our findings suggest that a national plan to build laboratory capacity is needed to ensure that the data collected by an integrated antimicrobial resistance surveillance network are of high quality (Box 1). All stakeholders should participate in the development of the plan, which should involve periodic training, quality improvement initiatives and rigorous follow-up protocols. Clearly, political commitment and sustained funding will be crucial for achieving the desired outcomes. Moreover, implementation of the plan would benefit from the Box 1. Main study findings • A One Health approach to antimicrobial resistance surveillance necessitates the creation of an integrated network. • Most data on antimicrobial resistance come from human sources as veterinary and environmental sectors have limited capacity for antimicrobial susceptibility testing. • Political commitment and leadership and multisectoral collaboration are critical for building an integrated surveillance network. • A top-down approach may not result in multisectoral collaboration or ensure the availability of resources for an integrated network. • The way forward is to identify and address challenges in implementation.
• Countries planning to create an integrated network can learn from others' experience.

Research
Integrated antimicrobial resistance surveillance, India Sonam Vijay et al.
support of trusted champions who can provide guidance and support across all sectors.
In 2013, the Indian Council of Medical Research established a national antimicrobial resistance surveillance network for human health, which today provides 15 000 United States dollars (US$) annually to participating hospitals and has invested in capacity-building and quality assurance initiatives. 12 We drafted a similar plan for an integrated surveillance programme that will collect comparable data from both veterinary and human sectors on five pathogens of public health importance (i.e. E. coli, Klebsiella, non-typhoid Salmonella, staphylococci and enterococci). We identified eight reference centres in India: four human centres for healthy and diseased human samples and four veterinary centres for healthy and diseased animal and food product samples.
Three of the five pathogens will each be assessed by one human and one veterinary laboratory, whereas staphylococci and enterococci will be assessed together by the fourth human and the fourth veterinary laboratory. These reference centres will: (i) carry out antimicrobial susceptibility testing; (ii) study resistance mechanisms; (iii) conduct in-depth molecular and transmission dynamics studies; and (iv) provide training for other laboratories. In addition, between eight and 10 potential participating centres have been identified: these centres will help in sample collection and provide isolates of the five key pathogens to the relevant reference laboratories (Fig. 3). Our experience with funding of the Indian Council of Medical Research's human antimicrobial resistance network indicates that the estimated annual cost would be US$ 55 000 for each reference centre and US$ 20 000 for each participating centre, which would cover human resources, non-recurring expenditure (e.g. for equipment), consumables, training, communication, travel, transport and other costs.
Having established a functional antimicrobial resistance surveillance system for human health, India has an opportunity to build on past experience and design an efficient integrated system that generates high-quality data on pathogens of concern for public health. In its 2017 National Action Plan, 11 the Indian government made a commitment to support antimicrobial resistance surveillance and to strengthen laboratories in human, animal and environmental health. However, this commitment has not been translated into an implementable plan with the requisite financial resources. We hope the findings of our systematic assessment and capacity-building exercise will contribute to the development of an integrated laboratory network in India. We believe our study is an important step in the creation of a long-term, integrated, antimicrobial resistance surveillance strategy linking human and veterinary sectors in the country and hope it will provide guidance to other countries planning to develop similar strategies. ■