Cost-Consequences Analysis (CCA) of Echocardiography Digital Artificial Intelligence (AI) Monitoring Application Developed at Hosmartai (Horizon 2020 Funded)
Author(s)
Chatzikou M1, Latsou D2, Apostolidis G3, Charisis V3, Hadjidimitriou S3, Hadjileontiadis L4, Pagkourelias E5, Vassilikos V5
1Pharmecons Easy Access LtD, Rafina, UK, 2Pharmecons Easy Access LtD, York, UK, 3Signal Processing & Biomedical Technology Unit, Department of Electrical & Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece, 4Department of Biomedical Engineering, Khalifa University, Abu Dhabi, Abu Dhabi, United Arab Emirates, 53rd Cardiology Department, Hippokrateion Hospital, Medical School, Aristotle University of Thessaloniki, Thessaloniki, Greece
Presentation Documents
OBJECTIVES: Echocardiography (ECHO) is a type of ultrasound scan used for examining cardiac function and morphology. Estimation of both measurements from ECHO scans is usually performed semi-manually by a cardiologist and requires a non-negligible amount of time, while accuracy depends on the cardiologist’s experience and the quality of scans, leading to intra- and inter-observer variability. The study aimed at evaluating the economic and clinical performance of an AI-based tool that automatically estimates left ventricle ejection fraction (LV-EF) and left ventricle global longitudinal strain (LV-GLS) from ECHO scans.
METHODS: A micro-costing analysis, based on costs of the Greek healthcare system, was performed to estimate the duration of LV-EF measurement by a high and low experienced physicians as well as the cost of the technology development, maintenance and infrastructure. The comprehensive selection of Key Performance Indicators was performed, to capture the incremental difference between the current practice and the new AI technology and enable the performance of the cost-consequence analysis.
RESULTS: The annual cost of use of the new echocardiography technology costs more than current practice (€9.409 vs. €2.116), mainly attributed to the cost of development of AI technology. The diagnostic accuracy for a low experienced physician<5years was 0.80 with the new AI technology vs. 0.54 of current practice. Similarly, for high experienced physicians>5years 0.82 vs. 0.64 at current practice. The duration of LVEF measurement by a low experienced physician was 4minutes vs. 1,5 with the new AI technology, similarly for highly experienced physician 3minutes vs. 1,5. The system usability score of the new technology was 75%.
CONCLUSIONS: Artificial intelligence (AI) could support cardiologists in LV function assessment with automatic and consistent estimation of LV-EF and LV-GLS, while reducing the time of ECHO examination and the variability of measurements. New AI technologies seem to be value for money options for healthcare systems.
Conference/Value in Health Info
Value in Health, Volume 27, Issue 12, S2 (December 2024)
Code
MT8
Topic
Medical Technologies
Topic Subcategory
Diagnostics & Imaging
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Medical Devices