Skip to main navigation Skip to search Skip to main content

Medical Large Language Models and Multi-Modal AI in Diabetes

Project: Applied Research

Project Details

Abstract

Type 2 diabetes (T2D) is a chronic metabolic disorder and one of the leading causes of disability and mortality worldwide [1]. In 2025, approximately 589 million people were living with diabetes, and global prevalence is projected to rise by nearly 45% by 2050 [2]. The Middle East/North Africa region has the highest regional prevalence of diabetes and the highest diabetes-related mortality globally [2]. Qatar has a high obesity rate [3], with 65–70% of obese adults and driving the country’s escalating T2D burden. In 2012, the Qatar STEPwise Survey reported a T2D prevalence of 16.7% among Qataris. By 2023, this number rose to 18.1% among Qataris and 15.4% among all adults in Qatar [4,5]. The 2025 IDF Diabetes Atlas estimated the prevalence of T2D among adults aged 20–79 at 24.6%, placing Qatar among the top 10 affected countries globally [2]. T2D drives complications across multiple organs. In Qatar, T2D accounts for approximately 50% of dialysis-requiring renal disease and 70% of stroke and transient ischemic attack cases [6]. Annual healthcare expenditure for diabetic patients reached 1.8 billion QAR in 2015 and is projected to rise to 5.0 billion QAR by 2030, a nearly threefold increase in just 15 years [3]. These obesity-associated diabetic complications arise from gene–environment interactions, where epigenetic mechanisms link unhealthy diets and lifestyle factors to metabolic dysregulation [7,8].

Submitting Institute Name

Hamad Bin Khalifa University (HBKU)
Sponsor's Award NumberHBKU-INT-VPR-JFI-02
Proposal IDHBKU-OVPR-SIC-01-18
StatusActive
Effective start/end date3/09/26 → 2/09/28

Primary Theme

  • Artificial Intelligence

Primary Subtheme

  • AI - Healthcare

Secondary Theme

  • None

Secondary Subtheme

  • None

Keywords

  • None

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.