Application of artificial intelligence methods in AMR

Expression of Interest Title: Data Challenge: Application of artificial intelligence methods in AMR
Team 8995, led by Fredrick Mutisya of Ministry of Health, Narok County, Kenya, with Rachael Kanguha, Chuka County Referral Hospital, Kenya

Summary of Research:

Harnessing Artificial Intelligence to Predict Antimicrobial Resistance Trends Summary We propose a research project aimed at employing artificial intelligence (AI) to predict trends in antimicrobial resistance. This project topic is precipitated by the growing threat of antimicrobial resistance, which stands as a significant challenge to global health. Traditional methods of prediction have been insufficiently dynamic to effectively monitor and predict these trends, leaving gaps in our understanding and ability to respond. In addition, the growing amount of genomic data is straining the capabilities of traditional modeling techniques. The study will involve the development of AI models, utilizing machine learning and deep learning techniques, to analyze existing data on antimicrobial resistance. These AI models will be trained to identify patterns and make predictions about future trends. The data sources that we hope to use are the Pfizer-Atlas data sets (antibiotics and antifungals) surveillance data set. Both of these datasets have pediatric data would be an interesting area for subgroup analysis. In addition, the large number of isolates and presence of genomic data lends itself well to machine learning and deep learning. The general approach would be to start with exploratory data analysis to check on data quality and structure. The team will then use geospatial, temporal and genomic variables from the data to train, validate and test different machine learning and deep learning models. The end goal is to showcase the best predictive model that can support healthcare professionals, researchers, and policy makers in forecasting antimicrobial resistance trends. This will enable proactive measures, early detection of emerging resistance patterns, and contribute to effective antimicrobial stewardship. The project’s success can potentially revolutionize the approach to tackling antimicrobial resistance. It can provide a model for ethically integrating AI in epidemiology and encourage data-based decision making.

Requested Datasets:

ATLAS_Antibiotics
Data Contributor: Pfizer Inc.

ATLAS_Antifungals
Data Contributor: Pfizer Inc.