Cardiovascular diseases Cardiovascular diseases
Conditions
Interventions
AI Powered ECG for Africa Platform with Clinical Decision Support
Sponsors
MI4People
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: Patients of any age or sex suspected of having cardiovascular disease. Patients undergoing a 12-lead ECG at one of the participating hospitals. Patients (or their legal guardians) who provide informed consent to participate in the study. Patients whose ECG data can be uploaded to the AI-powered ECG4Africa platform for analysis.
Exclusion criteria
Exclusion criteria: Patients with incomplete or poor-quality ECG recordings that cannot be analyzed by the AI platform. Patients who decline or are unable to provide informed consent. Patients already enrolled in another interventional cardiovascular study that may interfere with ECG interpretation. Patients with conditions that prevent standard 12-lead ECG acquisition (e.g., severe skin conditions, limb amputations preventing electrode placement).
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Diagnostic accuracy of the AI-powered ECG4Africa system compared with senior cardiologist interpretations, measured by sensitivity, specificity, and overall agreement in detecting cardiovascular abnormalities from 12-lead ECGs. | — |
Secondary
| Measure | Time frame |
|---|---|
| Turnaround time for ECG interpretation: Time from ECG acquisition to clinical decision support provided by the AI system. Clinician adherence to AI recommendations: Proportion of AI suggestions followed by physicians in clinical decision-making. Feasibility and usability of the AI system: Assessed through clinician feedback and system usage metrics. Capacity building outcomes: Number of clinicians trained and level of confidence in interpreting AI-assisted ECGs. Data completeness and quality: Proportion of ECGs successfully uploaded and analyzed on the AI platform. | — |
Countries
Ethiopia
Contacts
Public ContactTsegab Alemayehu Bukate
Clinical Researcher Armauer Hansen Research Institute
Outcome results
None listed