Skip to main content

Primary supervisor

Jesmin Nahar

Background

Diabetes can lead to serious complications, including peripheral neuropathy, peripheral arterial disease and diabetic foot disease, which may increase the risk of foot ulcers and lower-limb amputation. Early identification of high-risk patients can support timely clinical intervention and preventive care.

Research Aim

This research aims to investigate Artificial Intelligence (AI) and Machine Learning (ML) methods for early prediction and risk stratification of diabetes-related complications, with a specific focus on diabetic foot disease and lower-limb amputation risk.

Data and Methodology

The study will use relevant electronic health records (EHRs), glycaemic measures/logs and clinical history of patients with diabetes. Standard machine-learning techniques, including ensemble models, will be investigated to classify the risk of peripheral neuropathy and/or peripheral arterial disease.

Key Tasks

The project will involve:

  • Preprocessing and preparing EHR and clinical data.
  • Selecting and engineering relevant clinical features.
  • Training and comparing standard machine-learning/ensemble models.
  • Evaluating models using precision, recall, F1-score and precision-recall curves.
  • Identifying the most important clinical predictors associated with vascular and diabetic foot complications.

Expected Outcome

The project aims to develop an interpretable ML-based risk-classification approach that can help identify patients at higher risk of diabetic foot complications and support earlier intervention and preventive clinical care. The AI system will be considered a decision-support tool rather than a replacement for clinical judgement.

Aim/outline

Aim

To investigate the use of Artificial Intelligence (AI) and Machine Learning (ML) for early prediction and risk stratification of diabetes-related complications, with a particular focus on diabetic foot disease, peripheral neuropathy, peripheral arterial disease and lower-limb amputation risk.

Outline

The research will:

  • Analyse electronic health records (EHRs), glycaemic measures/logs and clinical history of patients with diabetes.
  • Preprocess and prepare clinical data for machine-learning analysis.
  • Identify relevant clinical features and potential predictors of diabetic foot complications.
  • Develop and compare standard machine-learning and ensemble models for risk classification.
  • Evaluate model performance using precision, recall, F1-score and precision-recall analysis.
  • Identify the most important clinical predictors associated with peripheral neuropathy, peripheral arterial disease and vascular complications.
  • Explore how AI-based risk stratification could support early identification, preventive care and timely clinical intervention.

Required knowledge

Required Knowledge

Students should have a basic to intermediate understanding of:

  • Artificial Intelligence and Machine Learning
  • Python and/or R programming
  • Data cleaning, preprocessing and feature engineering
  • Supervised machine learning, particularly classification
  • Basic statistics and data analysis
  • Model evaluation, including precision, recall, F1-score and precision-recall curves
  • Basic knowledge of healthcare, diabetes or clinical data is desirable but not essential
  • An interest in applying AI/ML to healthcare and clinical decision support