Multimodal Interoperable Common data model for AMR and One Health

Awareness-driven surveillance coupled with digital technologies has great potential to reduce the misuse of drugs in underprivileged communities. To empower them with linguistically-agnostic-conversational-systems to facilitate automatic medication management through only physicians, low-cost AMR surveillance, identify hot spots during environmental shifts, and build city-level forecasting models for one health.

1. Dr Rintu Kutum, PhD, Department of Computer Science, Ashoka University

2. Dr Saloni Kamboj, MBBS, MS, DNB, AIIMS, New Delhi

3. Dr Matthew Louis Robinson, MD, Johns Hopskin Medicine, JHU, USA

Although it is widely accepted that One Health approaches are needed to address AMR, current data collection and analysis across antimicrobial use and AMR prevalence is fragmented. For example, estimates of antibiotic use in food animals are not tied at a local level to AMR prevalence in human communities. Indian and global frameworks to address AMR discuss the importance of individual components of the AMR challenge but do not benefit from an understanding of their connectedness. The most widely cited data describing antimicrobial use in India and globally derives from commercial data sources which aggregate pharmaceutical sales records (Klein EY, PNAS 2018), but does not include indications for antibiotic use, making it difficult to craft policies to curtail use. Although approximately ¾ of all antimicrobial use is in food animals , modeling of antibiotic use in animals has required complex estimations based on sparse data (Van Boeckel TP, Science 2017).  Similarly, the estimation of community AMR prevalence in India derives from analysis of microbiology data from networks of flagship hospitals (ICMR). Without knowing why antibiotics are used and how the consequences of antibiotic use relate to AMR, it is difficult to craft policies and interventions to reduce AMR.

Marginalised and weaker sections of the society in Delhi NCR region (Delhi and Sonipat). We seek to support them with mobile and digital technologies designed along with and for them via free health services through camps and awareness

List of Public Good

- Linguistic-agnostic-conversational mobile app for AMR awareness 

- White paper or peer-review publication

- To build privacy-preserved and resource-limiting setting deployment of open-sourced large language models for health

Underprivileged people

Over the next year, we are planning to deploy the application and understand the build the interoperable standards to harmonized hotspots of drug usage from the communities along with comorbidities. Meanwhile, we will be collecting the environmental data, surveillance data of AMR.

None

Financial barriers

To build deployable robust systems, we would require qualified software engineers, and other stakeholders to work together. 

We require high compute infrastructure for multiple iterations of fine-tuning open source large language models. We are planning to integrate health-ops (dev-ops working for healthcare). 

Financial support

Yes we would like to collaborate with mined organizations to provide health services to underprivileged individuals, awareness about health, ecosystems and one health