AMR Insight Mapping (AIM) - AIM Solution

The AIM solution is a dynamic AMR surveillance dashboard, tracking bacterial AMR and antimicrobial use trends in communities and hospitals. It employs innovative data sources, including wastewater surveillance, enriched with genomic characterization. The dashboard, integrated with validated machine learning algorithms, is deployed on the national surveillance system, detecting hotspots.  

Professor Nazir Ismail 

The prevalence of community antimicrobial resistance (AMR) in South Africa is largely unknown, with empirical treatment for bacterial infections and limited testing. Community resistance emergence is evolving, and monitoring is inadequate. While hospital-based AMR systems exist, community-based monitoring is lacking, requiring inclusion in the current AMR dashboard.  

Testing communities universally for AMRs is impractical, necessitating a targeted approach. Urinary tract infections (UTIs), common bacterial infections, show emerging resistance, particularly in Escherichia coli and Klebsiella pneumoniae, top contributors to global AMR deaths. Temporal-spatial monitoring focusing on these pathogens, relevant antibiotics, and indicator mechanisms (ESBLs and CREs) is essential. However, testing all UTIs may not be sustainable. Using novel sources for this purpose, such as wastewater for AMR and antimicrobial use (AMU) processed with machine learning, would be as effective and cheaper.  

The solution requires a testing ground for validating the predictions. Gauteng, South Africa's smallest, most densely populated province, faces alarmingly high bacterial AMR rates, HIV (13%), TB (245/100k), and non-communicable diseases (22% hypertensive cases nationally). Diverse housing, from suburbs to slums, poor sanitation, and a large migrant population present challenges. An innovative solution is imperative for immediate action, serving as a model for national and regional expansion.  

The AIM Solution focuses on informing public health policy and practices. There are several target audiences for our solution.  

The scientific community working on AMR and AMU who require a better understanding of the prevalence of AMR and accurate AMU measurement in communities to undertake appropriate action to combat AMR.  

Public health policymakers on AMR. Leads for the Centre for Healthcare-Associated Infections, Antimicrobial Resistance and Mycoses at the National Institute of Communicable Diseases (NICD) are the critical liaisons with the Department of Health (DoH) and collaborators on this project.  

NICD is mandated to undertake surveillance, support policy development, and together with the DoH guide public health responses and clinical practice. The dashboard builds on their previous work, and the outputs will be transferred to the public-facing NICD website. In addition, we will advocate with local doctors and nurses on the AMR agenda.

Gauteng communities are the ultimate target audience directly affected by AMR. We will engage communities through educational offerings throughout the roll-out of the AIM solution, including digital social media posters/videos in local languages, pamphlet handouts at clinic visits, team members will participate in interviews on local radio and television stations and potential collaboration with Internews (Trinity Challenge Partner).   

The AIM Solution will provide several public goods benefiting the well-being of the Gauteng community and, ultimately, South Africa and beyond. The finalized product will be publicly available, and the underlying methodology and code will be accessible for use in other settings, especially in low- and middle-income countries. 

One primary public good is the AIM solution dashboard, a free-to-use platform. This dashboard will be accessible to key stakeholders, including the Department of Health (DOH), NICD, Department of Water and Sanitation, WHO, Public Health Managers, Researchers, and the general public. The design ensures integration into the current NICD surveillance platform, ensuring sustainability and ongoing impact. 

The comprehensive data within the AIM Solution will lead to multiple peer-reviewed publications. These publications will contain detailed methodology and data tables, facilitating replication in other settings and supporting potential future research. Additionally, validation data for novel wastewater surveillance methods and the potential value of different data sources will be included. 

Furthermore, the surveillance-generated data will be shared internationally, contributing to reporting initiatives such as the WHO GLASS surveillance.

The AIM solution aims to create tangible impact by providing early insights into temporal and spatial patterns of bacterial AMR and AMU at community level which has yet to be achieved. By leveraging machine learning models and geospatial maps, we can predict AMR patterns and potential over-use of antibiotics. This information will empower local health authorities to implement targeted measures, such as infection control interventions, public health campaigns or antibiotic stewardship programs, to curb the inappropriate use of antibiotics and the spread of AMR. 

The impact of the AIM Solution is twofold: firstly, it enables proactive management of AMR and inappropriate AMU, helping to target interventions in appropriate geographic locations. This benefits the entire community by safeguarding limited public health resources. Underserved and vulnerable populations stand to gain the most, as they often face limited access to healthcare resources. Our solution, tailored for community-level implementation, is a cost-effective and accessible tool for health authorities to address AMR and AMU issues, ensuring equitable health outcomes for all. This approach aligns with global efforts to combat antimicrobial resistance and has the potential to transform community healthcare practices in South Africa and other low- and middle-income countries. 

Over the next year, AIM Solution aims to expand its influence through collaborations with key partners, raising awareness, and advocating for the antimicrobial resistance (AMR) agenda within local communities. Utilizing existing health infrastructure and data, we will enhance our machine learning models iteratively based on historical and prospective data, ensuring adaptability to diverse community needs and improved prediction accuracy for targeted community investigations. 

Within the 3-year project duration, we strive to establish strategic partnerships with national and international health organizations, integrating our technology into public health systems like NICD. The Microbiology Department’s infection prevention and control unit that collaborates closely with the Gauteng Department of Health, will be activated to investigate events and provide valuable feedback to the model. Active engagement with the media and policymakers is planned to promote the integration of our approach into public health strategies. 

To optimize scalability, we will invest in user-friendly interfaces, develop training programs, and empower local health authorities. Regular impact assessments and the sharing of success stories will be conducted to instil confidence, encouraging national and global adoption for a transformative impact on societal health. 

To gauge success of the AIM Solution, we will focus on the accuracy and reliability of our predictions as the primary measure. Specific indicators include: 

Prediction Accuracy: Evaluate the precision of our machine learning models by comparing predicted AMR and AMU patterns with subsequent clinical, laboratory and consumption data. Accuracy will look at spatial differences between sites, and temporal trends within sites. 

Timeliness of Predictions: Assess the lead time our model provides compared to traditional clinical data collection methods. A shorter lead time indicates the tool's effectiveness in offering timely insights for proactive decision-making. 

User Adoption: Measure the acceptance and utilization of our solution by local health authorities. High adoption rates signify the tool's practicality and value in community-level settings. 

Community Awareness: Implement surveys and interviews to gauge community awareness of AMR and AMU issues over time. Increased awareness indicates the success of our tool in promoting public health education. 

These metrics will provide a foundation for refining the AIM Solution and implementation. As we progress, we will incorporate additional impact measures, informed by our initial findings, to guide our pathway towards achieving broader societal transformation over the subsequent years.

Financially, adequate funding is crucial for executing the project. Incremental optimization can enhance efficiency and lead to cost reductions. The components are intertwined, posing challenges in eliminating any single element. However, each component possesses standalone viability, mitigating risks as their results remain independently usable. Leveraging existing infrastructure, laboratories, data systems, and wastewater surveillance programs minimizes implementation costs.  

In addressing human resource challenges, the partnerships we have ensures access to skilled personnel. A specific issue is skills in data science, and our collaboration with a Data Science consultancy specializing in Predictive Analytics and advanced data services provides us with a team of data experts. 

Cultural and educational barriers include community misunderstanding of Antimicrobial Resistance (AMR) due to language barriers and limited awareness, especially in South African communities. Integrating educational components into clinical data collection will address the gap and ensure data quality. 

Ethical, regulatory, and administrative clearance processes will run concurrently with preparatory work to save time. Our good relationships with relevant authorities will facilitate time to attain approvals. Compliance with data protection regulations is critical. Laboratory data is strictly governed, with collaborators receiving only anonymized data and robust measures in place against inadvertent breaches.  

The AIM Solution team’s application to The Trinity Challenge is motivated by a strong alignment with the Trinity Challenge’s foundational principles of inclusivity, collaboration, and innovation in addressing antimicrobial resistance (AMR) at the community level. 

Inclusivity: The Trinity Challenge's commitment to inclusivity resonates with the AIM Solution’s goal of providing accessible, accurate and valuable data solutions for Gauteng/South African communities. By participating, we seek to amplify our reach, ensuring that underserved populations benefit from our predictive tool, thereby addressing disparities in healthcare access. 

Collaboration: The challenge's emphasis on cross-sectoral partnership is pivotal in overcoming data barriers. By engaging with The Trinity Challenge network, we aim to foster collaborations across healthcare, data science and environmental institutions, promoting data-sharing agreements and creating a collective effort to combat AMR. 

Innovation: The Trinity Challenge provides an ideal platform for showcasing the innovative aspects of the AIM solution. By participating, we hope to leverage the challenge's recognition and resources to refine and expand our tool, pushing the boundaries of innovation in the fight against AMR. 

By aligning with The Trinity Challenge's principles, we anticipate overcoming critical barriers, fostering meaningful collaborations, and accelerating the impact of our solution on a global scale. 

Google – Geospatial mapping and citizen trend data  

Facebook – citizen and Social Media data  

Discovery Health – SA private sector clinical data  

Internews – for the community education component