AI in Health Research
Machine learning models that detect disease earlier, personalise treatment, and accelerate drug discovery
Overview
Artificial intelligence is no longer a speculative promise in medicine — it is a working tool in radiology departments, drug discovery labs, and primary care clinics around the world. The scale of published research has grown by roughly 400 percent since 2015, spanning diagnostic imaging, clinical decision support, robotic surgery, pharmacovigilance, and remote patient monitoring. What distinguishes the current wave from earlier cycles of enthusiasm is the breadth of real-world deployment: regulatory agencies on every continent have now approved AI-based medical devices, and several have reached routine clinical use.
The stakes are considerable. Healthcare systems face converging pressures — ageing populations, rising chronic disease burden, workforce shortages, and cost constraints that show no sign of easing. AI cannot solve these problems alone, but targeted applications are already demonstrating measurable improvements in diagnostic speed, treatment precision, and operational efficiency. The key question has shifted from whether AI works in medicine to where it works best, under what conditions, and for whom.
This repository tracks peer-reviewed research, regulatory decisions, and deployment outcomes across the full spectrum of AI health applications. Each entry is sourced, contextualised, and accompanied by plain-language summaries that make findings accessible to clinicians, policymakers, and patients alike.
Key Breakthroughs
Retinal Imaging for Cardiovascular Risk
Researchers at Google Health demonstrated that deep learning models can predict cardiovascular risk factors — including blood pressure, age, and smoking status — from retinal fundus photographs alone. Published in Nature Biomedical Engineering, the study achieved area-under-curve scores comparable to traditional lipid panels, opening the door to non-invasive screening during routine eye examinations.
AlphaFold and Protein Structure Prediction
DeepMind's AlphaFold 2 solved one of biology's longest-standing challenges: predicting protein three-dimensional structures from amino acid sequences. The system achieved a median Global Distance Test score of 92.4 at the CASP14 competition, a leap considered equivalent to decades of incremental progress. The AlphaFold Protein Structure Database now contains over 200 million predictions, accelerating drug target identification and enzyme engineering worldwide.
AI-Assisted Radiology Triage
The UK National Health Service deployed AI-driven chest X-ray triage across 20 hospital trusts in 2023, prioritising scans suggestive of lung cancer, tuberculosis, or COVID-19 pneumonia. A Lancet Digital Health evaluation found the system reduced median report turnaround from 8.2 days to 1.1 days for critical findings, without increasing false-positive rates. Radiologists retained full diagnostic authority; the model served as a pre-read flagging tool.
Antibiotic Discovery via Deep Learning
A team at MIT used a graph neural network to screen over 107 million chemical compounds for antibiotic activity, identifying halicin — a structurally novel molecule effective against drug-resistant strains including Acinetobacter baumannii and Mycobacterium tuberculosis. Published in Cell, the work illustrated how AI can expand the search space far beyond what traditional high-throughput screening permits, discovering molecules in regions of chemical space that human intuition would not explore.
Continuous Glucose Monitoring with Predictive Alerts
Closed-loop insulin delivery systems — often called artificial pancreas devices — now incorporate machine learning models that predict glucose excursions 30 to 60 minutes ahead. The FDA-approved MiniMed 780G and Tandem Control-IQ systems have demonstrated a 15 to 25 percentage-point increase in time-in-range (70–180 mg/dL) in randomised controlled trials, reducing both hyperglycaemic and hypoglycaemic episodes for Type 1 diabetes patients.