AI POWERED TRACKERS AND SENSORS FOR EARLY AMR DATA SURVEILLANCE

We propose to develop and train a robust AI integrated machine learning model using image dataset to track and transmit real time data of unhealthy human companions and marauding wild animals spreading AMR organisms in environments. This early disease surveillance system will identify onset and bilateral transmission amongst pastoralists

Dr Joy Nyawira Riungu. Sanitation Engineering. 

Dr Joy is the Founder and Director for the Sanitation Research Centre at Meru University of Science and Technology

Globally, close to 4.9 million people die of multi drug resistant attributable illnesses each year. Achieving good health for all SDG will save lives of the estimated statistics of death projections of close to10 million persons per year by 2050. Antimicrobial resistance in wild animals poses significant threat to human and animal health within their interface environments. Companion animals, and marauding wildlife act as reservoirs/vectors for AMR microbes. Complications happen when the animals are sick and in distress, they shed antimicrobial resistant bacteria such as Salmonella in the environments, or through contamination of animal protein products. During herding and food handling these organisms end up being ingested by man and animals as they graze in the contaminated environment. In LMIC, majorly affected groups are children under five years, immunocompromised adults and the aged.  The emergence of AMR can be blamed on either animals and people, majorly resulting from under use, misuse and abuse of antimicrobial agents resulting from easy access of antibiotics from drug kiosks. This complicates management and augments spread of AMR within the community members, animal populations as well as in wildlife, thus great loses affecting the already stressed up economy.

The pastoral communities living in close interactions with wild animals are at greater risk of increased mortality due to spread and transmission of zoonotic Antimicrobial resistant bacteria.

Supporting them with a system that is able to identify onset, before subsequent spread and thus transmission of these antimicrobial resistant organism is our main foci.

We will design and provide them with the user-friendly AI Geo integrated mobile app with basic smart phones. We will train them on how to capture real time data and relay it immediately they notice any distress in either people or animal behaviors for quick intervention.

This will alleviate disease burden and reduce mortality due to spread, transmission and complications of antimicrobial resistant microorganisms within the communities. The phone will be provided to selected community members, including the herds me, community health workers and cattle owners. They will be trained to take photos of animals immediately they suspect any form of distress.

The AI developed in data sets will detect onset of AMR early enough to alleviate community suffering and massive loss of lives due to treatable bacterial strains. 

The tools shall be made available to local community especially on data pertaining transmission dynamics.

Here, the innovation will empower the community policing on AMR elimination campaigns. With the knowledge, on which antimicrobial agents to use when, it will be possible to go back to susceptible first-generation antibiotics. The community shall be in a position to identify sick animals and people on time, call for help from the public health or from veterinary officers and also from environmental scientists to bring their integrated expertise. This way mapping and intervention of zero AMR organisms in the community campaigns shall be acquirable. Once this innovation is tested and working, data shall be published in peer reviewed journals. The app shall in such a way as to relay feedback to local communities through the smartphones (already provided) in their local dialect. The information shall be in terms of sensitization videos, The herds men can watch the videos as they graze, while community health workers can make available the feedback in local radios and community education forums.

The AI integrated smart phones will be provided to herdsmen, community health workers and animal health professionals in the area. These will collect real time data.  Data relayed will include animal status immediately there is onset of distress, animal fecal matter real time images, human data especially on open defecation after start of distress and diarrhea. The data captured will include Geo parameters at that point in time. Collard tracked animal movement will be monitored via GpS coordinated which will subsequently guide data collection and analysis, both from domestic and wild animals (in the are both are often observed encroaching the interface areas because there exists no barrier (Kagendo et al 2014). Immediately data is relayed scientist will move with spread to collect and analyze samples early enough.  Data obtained shall inform clear treatment recommendations based on the best 1st generation antibiotic that will be found susceptible to the organism. Information on effective prevention and control of Spread of AMR, proper handling of animal protein products and animal handling/human behavior change, including proper use of antimicrobial agents shall be relayed back through videos and talks in local languages.  The aim will be reducing disease burden, mortality, and improve livelihoods. 

Once this prototype is tested in Samburu wildlife human interface area (foci Non typhi salmonella AMR strains), it will be further used in other similar pastoral communities in Maasai community and Tharaka communities both who similarly live at the interface areas of Maasai Mara National Reserve and Meru National Park respectively.  The innovation will thereafter be used across the country, for early detection of AMR and other various diseases in either similar setups, or different setups such as in informal settlements. It will be possible to also map out sources of abused antimicrobial agents in drug stores and notify regulators, including pharmacy and poisons board on time.

The solution can then be used in East Africa (Tanzania (interface areas of the Serengeti) and Uganda (Interface areas of queen Elizabeth National Park). Then in sub–Saharan Africa, in other low- and middle-income countries to reduce mortality and economic loses due AMR.  This once working well can be adopted globally for early AMR surveillance. 

The success shall be measured as follows. 

1. Ability of the system to capture and transmit real time images of distressed sick animals and people immediately behavior starts changing.
2. Ability to capture and transmit real time images of feces shed by the animals and people whilst in distress (Often characteristic diarrhea like fecal samples).
3. Ability to capture and transmit successfully data relating to animal behavior and GpS coordinates where data on environmental samples as well as related samples shall be collected as controls.
4. Ability to collar and geotag animals (both wild and domestic) notes to be in distress and marauding) in a view to monitor their specific movements from park areas to domestic residences and back for successful real time collection of shed fecal samples.
5.  Turnaround time after signal reception to sample collection, to laboratory analysis and feedback through the app.
6. Monitoring reduction in AMR related mortality in hospital cases (both inpatient and outpatient data shall be used in this case).
7. Monitoring animal loses (livestock and wild animals) due to AMR related diseases. 
8. Monitoring community behavior change towards antimicrobial use/abuse, and sanitary behaviors risking zoonotic disease transmission (whether AMR or susceptible strains of bacteria)

• Cultural barrier: The community where we want to test the prototype is used to movement in search of pasture for livestock for their livelihood. To counter this, we intend to move with them with our AI trained gargets. Only real time data collection shall be done using tracking systemics and GpS coordinates from relayed data.
• Insecurity: The areas are not very secure due to dangerous wildlife encroaching the interface areas: We will use community expertise and involve wildlife rangers for security while collecting samples in these areas.  Illiteracy: The level of literacy in the pastoral communities is quiet low. To counter this, we will use translators, area chiefs and community health workers for interpretation to local dialect and for acceptance in the communities.
• Financial barriers. We are hopeful to get this innovation funded through the Trinity challenge. Lack of finances has limited our thought in trying to reduce mortality due to AMR, in the community, within animals, in the environments, in food industry and also in hospitals environments.

The pastoral communities living in close interactions with wild animals are at risk of increased mortality due to spread and transmission of zoonotic diseases. The rise of antimicrobial resistance (AMR) in microbes carried by wild animals poses significant threat to human and animal health where their populations interact. Companion animals (livestock), and marauding wildlife can act as reservoirs for AMR microbes. To combat this challenge, we propose a novel early disease surveillance system that integrates GPS tracking, image analysis, and machine learning, entailing equipping wild animals with wearable GPS trackers to monitor their movements and capture images of their feces using smartphones. This data will be used to train machine learning models that can identify high-risk animals based on movement patterns and fecal characteristics, detect disease onset through analyzing GPS data and images, and monitor the potential for bilateral transmission of AMR microbes between wildlife and pastoralists within shared environments.

By applying to trinity challenge, we hope to improve antimicrobial resistant disease surveillance hence supporting affected communities with a system capable of identifying onset, before subsequent spread and thus transmission of antimicrobial resistant organism. This will alleviate disease burden and reduce mortality and complications of AMR microorganisms within the communities.

1. Bill &Melinda Gates foundation (www.gatesfoundation.org)
2. Wellcome  (www.welcome.org)
3. Ineos Oxford Institute for Antimicrobial Research (www.ineosoxford.ox.ac.uk)
4. GSK (www.gsk.com)
5. Institute for Health Metrics and Evaluation (www.healthdata.org)
6. Johns Hopkins Bloomberg School of Public Health (www.jhsph.edu)