Finalist
Student Team 13243 from Kyiv School of Economics: Inna Kucherova (Team lead) and Anna Gurina
Antimicrobial resistance (AMR) is a silent global pandemic. Currently, clinical surveillance relies on rigid “breakpoints” to classify bacteria as either susceptible or resistant. However, this binary approach creates a dangerous blind spot: it misses the phenomenon known as “MIC Creep” – a gradual, invisible increase in the Minimum Inhibitory Concentration (the amount of antibiotic needed to kill the bacteria). While the bacteria may still be technically classified as “susceptible” today, they are quietly evolving toward full resistance tomorrow.
Our proposed research bridges wet-lab epidemiology with data science by leveraging Explainable Artificial Intelligence (AI). We will analyze massive global datasets (over 1 million isolates) to build an early-warning predictive model. This tool will detect hidden patterns of MIC Creep in high-priority pathogens like Acinetobacter baumannii and Klebsiella pneumoniae, predicting resistance shifts before they result in clinical failure. How the research will impact key areas:
Help improve patient outcomes: By detecting hidden resistance trends, this model will alert clinicians when a standard antibiotic is losing its potency, even if laboratory tests still label it as “effective.” This allows doctors to adjust treatment plans preemptively, reducing the risk of treatment failure, preventing severe complications, and ultimately saving lives, especially in vulnerable groups like pediatric or ICU patients.
Strengthen stewardship: Antimicrobial Stewardship (AMS) relies on using the right drug at the right time. Our tool will prevent the blind overuse of broad-spectrum antibiotics. If the model shows that a pathogen’s MIC is creeping upward in a specific hospital, stewardship teams can temporarily restrict that drug to preserve its efficacy. Conversely, it will give clinicians the confidence to use older, targeted antibiotics if the data shows no signs of creeping resistance.
Inform public health practice: The project shifts public health surveillance from a reactive to a proactive model. Instead of merely counting how many people died from resistant infections last year, epidemiologists will have a “weather forecast” for AMR. Public health centers will be able to map geographic hotspots where resistance is incubating, allowing for targeted interventions, localized quarantines, and updated clinical guidelines before a full-scale outbreak occurs.
Strengthen health systems: AMR places a massive financial and logistical burden on healthcare systems through prolonged hospitalizations and expensive second-line treatments. By integrating AI-driven insights into national surveillance pipelines, ministries of health can optimize their healthcare budgets. Decision-makers will be able to forecast exactly which antibiotics need to be procured nationally for the next 3–5 years based on the evolving resistance landscape, reducing waste and ensuring hospitals are always stocked with effective medications.