Cost-Effectiveness of Integrating Artificial Intelligence-Based Retinal Photographic Cardiovascular Disease risk Assessment with Diabetic Retinopathy Screening Programmes: A Multi-Country Analysis
Author(s)
Lei Zhang, PhD;
Monash University, School of Translational Medicine, Clayton, Australia
Monash University, School of Translational Medicine, Clayton, Australia
Presentation Documents
OBJECTIVES: We aim to evaluate the cost-effectiveness of integrating AI-based retinal photographic screening for CVD risk with DR in diabetic individuals in Australia, Singapore and the United Kingdom (UK).
METHODS: We constructed a hybrid decision tree/Markov model to simulate the disease progression of DR and CVD in eligible diabetic populations for a lifetime in the three countries, respectively. A cost-effectiveness analysis was performed from the health provider’s perspective. Interventional scenarios included AI-based DR screening only, and AI-based coupled DR and CVD screening. Each approach was stratified based on four age groups (all eligible age groups, 40/45, 75, and 40/45-74) and three AI screening coverages (20%, 50% and 80%), resulting in 24 interventional scenarios for each country.
RESULTS: In Australia (1.28m people with diabetes), compared to the status quo (A$71.02b lifetime costs with 12,985,022 QALYs), implementing AI-based DR screening would be cost-saving in all interventional scenarios. Adding CVD to DR screening would further enhance the cost-effectiveness. Particularly, conducting a systematic screening of combined DR and CVD at a coverage of 80% in all adults with diabetes would be the most cost-saving scenario, saving the healthcare system A$1.45b, and gaining 102,430 QALYs throughout the lifetime. In Singapore (341,680 individuals with diabetes), implementing AI-based DR screening only would gain more health benefits but not be cost-effective compared to the status quo (S$16.50b with 3,364,515 QALYs). However, adding CVD to DR screening would be cost-effective. In the UK, the status quo would cost the healthcare system £48.89b throughout the lifetime of 2.82m individuals with diabetes, with accumulated 29,179,895 QALYs. In comparison, implementing AI-based DR screening would be cost-saving for all age groups and screening coverages. Adding CVD screening would further increase health benefits while remaining cost-effective.
CONCLUSIONS: Integrating AI-based retinal photographic CVD risk screening to DR screening showed to be cost-effective in Australia, Singapore and the UK.
METHODS: We constructed a hybrid decision tree/Markov model to simulate the disease progression of DR and CVD in eligible diabetic populations for a lifetime in the three countries, respectively. A cost-effectiveness analysis was performed from the health provider’s perspective. Interventional scenarios included AI-based DR screening only, and AI-based coupled DR and CVD screening. Each approach was stratified based on four age groups (all eligible age groups, 40/45, 75, and 40/45-74) and three AI screening coverages (20%, 50% and 80%), resulting in 24 interventional scenarios for each country.
RESULTS: In Australia (1.28m people with diabetes), compared to the status quo (A$71.02b lifetime costs with 12,985,022 QALYs), implementing AI-based DR screening would be cost-saving in all interventional scenarios. Adding CVD to DR screening would further enhance the cost-effectiveness. Particularly, conducting a systematic screening of combined DR and CVD at a coverage of 80% in all adults with diabetes would be the most cost-saving scenario, saving the healthcare system A$1.45b, and gaining 102,430 QALYs throughout the lifetime. In Singapore (341,680 individuals with diabetes), implementing AI-based DR screening only would gain more health benefits but not be cost-effective compared to the status quo (S$16.50b with 3,364,515 QALYs). However, adding CVD to DR screening would be cost-effective. In the UK, the status quo would cost the healthcare system £48.89b throughout the lifetime of 2.82m individuals with diabetes, with accumulated 29,179,895 QALYs. In comparison, implementing AI-based DR screening would be cost-saving for all age groups and screening coverages. Adding CVD screening would further increase health benefits while remaining cost-effective.
CONCLUSIONS: Integrating AI-based retinal photographic CVD risk screening to DR screening showed to be cost-effective in Australia, Singapore and the UK.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
EE273
Topic
Economic Evaluation
Topic Subcategory
Cost/Cost of Illness/Resource Use Studies
Disease
SDC: Cardiovascular Disorders (including MI, Stroke, Circulatory), SDC: Diabetes/Endocrine/Metabolic Disorders (including obesity)