AMRSense: Empowering Communities with a Proactive OneHealth Ecosystem

AMRSense, our socio-technological innovation will 

1. build a network of community health workers in two Indian states, 

2. empower with AI-assisted data recording,

3. integrate antibiotic sales, consumption, surveillance trends,

4. federate predictive analytics across OneHealth ecosystem,

5. implement a multidimensional AMR scorecard at scale, 

to close the capture, analyze, engage loop for tackling AMR.

• Overall Team lead: Dr. Tavpritesh Sethi (IIIT-Delhi, India)
• Organizational Leads: 
  • Dr. Tavpritesh Sethi (IIIT-Delhi, India)
  • Dr. Harpreet Singh (ICMR. India)
  • Dr. Tikesh Bisen (CHRI-PATH, India)
  • Mr. Nikhil Doegar (Tata 1mg.com, India)

The specific problem we are addressing with AMRSense is the engagement, motivation, and training of community health workers (CHWs) in antimicrobial resistance (AMR) surveillance and management, compounded by the lack of a comprehensive data ecosystem, and analytics capabilities. India alone, with more than 900,000 ASHA workers encounters hurdles such as limited awareness, insufficient training, and low motivation, and a conspicuous gap in AMR data collection at the community level. Hence the absence of evidence-based AMR management. AMRSense tackles this multifaceted problem through four major components:

• Community Engagement: Building a network of CHWs, empowered with AI-assisted data recording tools to simplify data collection and improve accuracy.

• Data Integration: Integrating antibiotic sales, consumption, and WHONet-compliant surveillance data through open-source tools and APIs to create a unified AMR data ecosystem.

• Predictive Analytics: Federating predictive analytics across the OneHealth ecosystem to triangulate integrative insights on AMR. 

• AMRaura will be a multidimensional AMRaura scorecard to monitor and evaluate AMR trends, guiding targeted interventions and demonstrating the tangible benefits of data collection efforts.

Rooted in empirical research, AMRSense aims to enhance CHW engagement, motivation, and data collection capabilities while establishing a robust data ecosystem and analytics framework for AMR surveillance and management.

AMRSense has three major target audience categories: 

1. Community Health Workers (CHWs) at the forefront of healthcare delivery in LMICs.

2. Microbiology Labs and District Hospitals routinely generating massive data on Antimicrobial Susceptibility Testing, often lost in the absence of open, interoperable digital solutions.

3. Policymakers, district-level officials, and directors of AMR Stewardship programs.

AMRSense addresses the needs of these categories by 

1. building digital communities and camaraderie through social badges, recognition, and leaderboards to enhance motivation and engagement.

2. empowerment with AI-assisted tools for efficient, federated and secure data recording aligned with national and international digital frameworks such as Ayushman Bharat Digital Mission (ABDM) and Fast Health Interoperable Resources (FHIR).

3. showcasing the value proposition of data through analytics, scorecards, interactive dashboards. 

Over the years, the AMRSense team has engaged in understanding and addressing these needs and gaps. ICMR's surveillance network has created a WHONet-compliant prototype of a common data model(CDM). IIIT-Delhi has developed AI models, apps, and dashboards for hotspot identification, trends, scorecards, and the impact of AMR on SDG achievement. CHRI, the Indian affiliate of PATH engaged with community health workers and authorities. Together, we will be conducting further surveys, interviews, and focus group discussions for design, validation, testing, and feedback on AMRSense.

1. Community-Centric Social Network Model: a replicable model for identifying and engaging focal points for awareness for public health initiatives, open for adaptationion in other public health initiatives.
2. Open-Sourced Data Ecosystem: modular, federated data ecosystem using APIs that will be globally accessible and can be adapted and used by other countries. 
3. Integrative modeling framework. Open-sourced modeling platform similar to this platform (IFHP) was created by PI's lab. Exploration of non-personal open-data and community-contributed models will be encouraged through a datahub and modelzoo.
4. AMRaura Scorecard: a tool for policymakers, healthcare providers, and researchers useful for integrative modeling, monitoring trends, guiding interventions, and resource allocation. 
5. AMR Awareness, Education Resources, AMR Playbook: a suite of awareness and education materials to enhance understanding and promote responsible antibiotic use among healthcare providers, the general public, to guidelines and best practices for AMR management.
6. Knowledge Communities and Campaigns: campaigns, workshops, health talks, and training modules, enhancing AMR awareness.
7. Research and Knowledge Sharing: publish findings in peer-reviewed publications,  white papers 

By providing these public goods, AMRSense aims to improve global health security, support evidence-based decision-making, and foster a collaborative approach to addressing the critical challenge of antimicrobial resistance.

We aim to create a tangible impact through a cohesive series of activities, outputs, and outcomes:

1. Empowering community health workers (CHWs) with AI-assisted tools and a federated data ecosystem for efficient AMR data collection.

   • Output: Improved quality and reliability of AMR surveillance data.
   • Outcome: Enhanced capacity for healthcare providers and policymakers to monitor AMR trends and implement targeted interventions.

2. Conducting education campaigns and training modules to raise awareness about responsible antibiotic use among healthcare providers and the public.

   • Output: Increased awareness and understanding of AMR.
   • Outcome: Reduced misuse of antibiotics, contributing to a decrease in the spread of AMR.

3. Implementing the AMRaura scorecard to integrate diverse datasets and provide insights into AMR dynamics.

   • Output: A tool for monitoring AMR trends and guiding interventions.
   • Outcome: Informed decision-making leading to more effective AMR management strategies.

4. Utilizing advanced analytics to process integrated data for identifying AMR trends and predicting resistance hotspots.

   • Output: Actionable insights for AMR management.
   • Outcome: Data-driven policy formulation and resource allocation for AMR containment.

AMRsense is supported by our background work, evidence on strengthening systems with data, and the success of initiatives such as ICMR IAMRSN and mMitra. 

Over the next one year, AMRSense aims to scale its impact by:

1. Expansion and integration into the District-hospital based surveillance in 700+ district hospitals under an MoU between ICMR, AIIMS, and NHSRC. 

2. Develop plugins for social initiatives of corporate partnerships such as Tata 1mg.com

3. Strengthening National Partnerships: Leveraging our lead-role in the Delhi Science and Technology Cluster (DRIIV), we will liaise with government agencies, non-governmental organizations, and private sector partners to leverage their networks and resources for wider deployment.

In the next three years, the goal is to:

1. Policy Integration: Work closely with policymakers to integrate AMRSense's data-driven insights into national AMR surveillance and management strategies.
2. Technology Enhancement: Through inter-cluster collaboration currently existing with Bangalore and Pune clusters, and with organizations leading AMR diagnostics such as CCAMP, we plan to onboard new technology innovations into AMRSense's open-source ecosystem

3. Global Outreach: The presence of PATH in nearby LMICs such as Bangladesh will be instrumental in creating a global outreach for AMRSense. We will explore opportunities to adapt and implement AMRSense in other countries facing similar AMR challenges, particularly in the South East Asia region, leveraging India's experience as a model.

A Knowledge, Action, and Practice approach will be used to assess the success of the solution. Baseline data for the following specific and measurable indicators will be tracked over time to measure success against impact goals

1. Community Engagement: We will assess the level of community engagement and participation of healthcare workers (HCWs) and frontline workers (FLWs) in AMR surveillance and awareness efforts. The increase in the number actively using AMRSense and participating in awareness activities will be recorded.

2. Quality and Quantity of Data Collected: Number of District hospitals onboarded; overall volume, completeness, and accuracy of data collected by HCWs, FLWs, and laboratories using AMRSense. Randomly selected subsets will be evaluated for the accuracy of the recorded antibiotic prescriptions and AMR cases.

3. Usage of AMRaura Scorecard: adoption and utilization of the AMRaura scorecard by IAMRSN sites, district hospitals, and district health officials.  We will record the number of policy consults arising from the use of the AMRaura scorecard.

4. Awareness and Behavior Change: We will assess the effectiveness of education campaigns and training modules in raising awareness and promoting responsible antibiotic use. A reduction in inappropriate antibiotic prescriptions, measured through pre- and post-campaign surveys will indicate behavior change.

1. Financial Constraints: Effective engagement with health care workers(HCWs) and frontline workers(FLWs) requires significant financial resources. Additionally, our social badges proposed to motivate communities through health services add further costs. To build and sustain the momentum during the execution, we will roll out AMRSense in three phases: 

   1. Ignite: Stakeholder engagements, gap understanding, community education, human-centered design iteration.

   2. Overdrive: pilot AMRSense at the block level and building of social recognition and rewards. 

   3. Autopilot: Identification of community and lab champions for targeted interventions such as awareness campaigns and data pioneers.

2. Technical Challenges: 
   1. Privacy. AMRSense will not store individual data on our servers. Federated aggregated data to assess the AMR "aura" at the block level instead.
   2. Interoperability: a common data model and WHONet-compliant APIs for data capture. 
3. Cultural and Educational barriers: resistance to technology adoption will managed by early engagement with HCWs and FLWs, human-centered design, and iterative development.

4. Policy and Regulatory Hurdles: Collaboration with ICMR on the team will help understand and navigate the globally shifting landscape of AI and patient privacy. 

5. Market Challenges: Establishing the value proposition of AMRSense through pilot projects, dashboards, AMR Scorecards, and integration with pharmacy solutions will build credibility.

The Trinity Challenge aligns perfectly with our prior collective work and mission to combat antimicrobial resistance at the ground level in LMICs. AMRSense empowers data collection and leverages AI and analytics to improve AMR surveillance at scale, which aligns perfectly with The Trinity Challenge's objectives. In general, AMRSense resonates with the Challenge's focus on harnessing the power of data to tackle AMR and health emergencies.

Barriers The Trinity Challenge Can Help Overcome:

1. Financial Constraints: Adoption of digital solutions requires an initial phase of value-building exercise with frontline workers. The Challenge provides a unique opportunity to overcome this financial hurdle and enable a self-driven data ecosystem extending beyond the pilot regions proposed.
2. Technical Challenges: Access to The Trinity Challenge's network of experts and resources can help us refine our strategy, ensuring the robustness and scalability of our solution.
3. Policy and Regulatory Hurdles: The Challenge can facilitate connections with policymakers and regulatory bodies, aiding us in navigating legal frameworks and advocating for policies that support AMR surveillance and management.
4. Market Challenges: The visibility and credibility gained through The Trinity Challenge can help us establish the value proposition of AMRSense, attracting additional partners and stakeholders for wider adoption.

1. The World Health Organization (WHO):

   • Value Proposition: Provides global leadership in AMR surveillance and policy guidance.
   • How It Helps: Access to global AMR guidelines and networks, enabling AMRSense to align with international standards and expand its impact.

2. The Centers for Disease Control and Prevention (CDC):

   • Value Proposition: Expertise in infectious disease surveillance and response.
   • How It Helps: integration and refinement of AMRSense surveillance tools and methodologies.

3. ReAct – Action on Antibiotic Resistance:

   • Value Proposition: Advocacy and policy development focused on AMR.
   • How It Helps: Assists in raising awareness about AMRSense and advocating for policies that support its adoption and scaling.

4. Global Antibiotic Research and Development Partnership (GARDP):

   • Value Proposition: Research and development of new antibiotics.
   • How It Helps: Provides insights into emerging antibiotic treatments and resistance patterns, enriching AMRSense's data ecosystem.

5. Fleming Fund:

   • Value Proposition: Financial support for AMR surveillance capacity-building in LMICs.
   • How It Helps: Offers funding opportunities to expand AMRSense's implementation in LMICs, enhancing its reach and impact.

6. Médecins Sans Frontières (MSF):

   • Value Proposition: Medical programs addressing AMR in LMICs.
   • How It Helps: Shares practical insights and experiences from AMR management in resource-limited settings, informing AMRSense's effective approach to address AMR challenges.