Cardiovascular diseases are a major and growing health challenge in Ethiopia, yet access to specialist cardiac care remains extremely limited. A population of around 130 million people is served by only about 30 cardiologists. In many hospitals, particularly outside major urban centers, physicians and other healthcare professionals therefore have to assess patients without immediate access to specialist expertise.
An electrocardiogram (ECG) is one of the most important and affordable tools for detecting heart conditions. However, recording an ECG is only the first step: interpreting it correctly requires considerable expertise. Where cardiologists are unavailable, potentially serious conditions can be overlooked or diagnosed too late.
Our project AI-SUSTAINS CVD Care addresses this gap by bringing AI-supported ECG interpretation directly to healthcare professionals in Ethiopian hospitals.
The system analyses standard 12-lead ECGs within seconds and provides clinicians with decision support for several common and potentially serious cardiac conditions. It acts as an additional pair of expert eyes: the healthcare professional remains responsible for the diagnosis and treatment decision, while the AI helps identify abnormalities that might otherwise be missed.
The project is implemented by the German nonprofit MI4People, together with the Armauer Hansen Research Institute (AHRI) in Ethiopia and Technische Hochschule Mittelhessen (THM) in Germany. The partnership combines AI expertise, medical research and deep integration into the Ethiopian healthcare system.
From a successful pilot to 100 hospitals: This is not a project starting from scratch.
In a first pilot phase, supported by the Bavarian State Chancellery, the AI-powered ECG solution was introduced in nine hospitals across Ethiopia. More than 30 healthcare professionals were trained, and approximately 7,200 ECG recordings from Ethiopian patients were collected within only a few months.
The pilot demonstrated that the technology can be integrated into real clinical workflows and used by healthcare professionals in Ethiopian hospitals. It also provided an important lesson for the next phase: medical AI developed primarily with data from Europe or North America cannot simply be transferred to other populations without careful local validation and adaptation.
We now want to take the decisive next step: scale the solution from nine pilot hospitals to approximately 100 healthcare facilities across Ethiopia and adjust our AI algorithms to local population.
At this scale, the infrastructure could enable approximately 600,000–800,000 AI-supported ECG assessments per year, dramatically extending access to cardiac diagnostic support without requiring a cardiologist to be physically present in every hospital.
What the next phase will achieve: Scaling is about much more than installing software. Our goal is to establish a sustainable clinical and technological infrastructure that Ethiopian healthcare professionals can operate and further develop locally.
The scale-up will therefore focus on several interconnected areas:
1. Expand access to AI-supported cardiac diagnostics
We will deploy the AI-ECG solution to approximately 100 hospitals and healthcare facilities, including facilities outside the largest urban centers. This will give substantially more physicians access to rapid AI-supported ECG interpretation when specialist expertise is not immediately available.
2. Train hundreds of healthcare professionals
Technology creates impact only when people can use it confidently. We plan to train more than 300 healthcare professionals in the use of the system and its integration into everyday clinical workflows. Local training structures and knowledge transfer will help ensure that expertise remains within the Ethiopian healthcare system.
3. Build AI that works for the population it serves
One of the most important challenges in medical AI is the lack of representative health data from African populations. Most existing ECG AI systems have been developed primarily using datasets from Western countries.
The approximately 7,200 ECGs already collected during the pilot provide a strong starting point. During the scale-up, additional anonymized ECG data will be collected, carefully annotated with Ethiopian medical experts and experts form other African countries and used to evaluate and improve the AI models.
This allows us to develop a system that is not simply imported into Ethiopia but is validated and continuously improved using Ethiopian clinical data.
4. Create an open resource for global health research
Where ethically and legally possible, anonymized data and resulting research resources will be made openly available. Our ambition is to contribute one of the most significant openly accessible ECG resources focused on African populations.
The impact can therefore extend far beyond our own project. Researchers, universities, startups, and nonprofit technology initiatives around the world could use these resources to develop and evaluate more representative medical AI systems.
5. Generate evidence for sustainable adoption
The scale-up will systematically evaluate diagnostic performance, usability, clinical integration and impact. This evidence is essential for moving from an innovative pilot toward long-term integration into the healthcare system.
Together with our Ethiopian partners, we will investigate how AI-supported ECG interpretation can strengthen clinical decision-making, support healthcare workers, reduce unnecessary referrals and improve access to specialist-level diagnostic support.
A healthcare project enabled by AI – not an AI experiment: Our objective is not to introduce technology for its own sake.
The core problem is a shortage of specialist medical expertise. Training enough cardiologists to provide nationwide coverage will take many years. AI cannot replace those specialists, but it can help distribute part of their diagnostic expertise much more widely.
A physician working far away from a specialist cardiac center should be able to record an ECG and receive immediate support in identifying potentially dangerous abnormalities. Patients who require specialist attention can then be referred more effectively, while many routine assessments can be handled locally.
In this way, AI can help healthcare professionals make scarce specialist resources available where they are needed most.
Designed for long-term sustainability: Sustainability is a core principle of this initiative.
The project is being developed together with Ethiopian institutions rather than deployed as an externally controlled technology. AHRI leads clinical implementation and works closely with hospitals and national healthcare stakeholders. Training and knowledge transfer strengthen local capacity, while the technical infrastructure is being designed so that it can increasingly be operated locally.
Our long-term ambition is for the solution to become part of a sustainable Ethiopian digital-health ecosystem rather than remain dependent on a temporary international project.
At the same time, the technology and knowledge created through the project can provide a blueprint for other countries facing similar shortages of medical specialists.
From 9 hospitals to 100 – and potentially far beyond: The first phase showed that AI-supported ECG diagnostics can work in Ethiopian hospitals. The next challenge is to demonstrate that it can work at scale.
Expanding from nine hospitals to approximately 100 facilities would transform the project from a promising pilot into healthcare infrastructure capable of supporting hundreds of thousands of cardiac assessments every year.
Every contribution helps us move toward that goal: deploying the technology to more hospitals, training healthcare professionals, improving and validating the AI with local data, maintaining the necessary digital infrastructure and building the evidence required for sustainable nationwide adoption.
Our vision is simple: a patient's chance of receiving a timely and accurate assessment of a potentially dangerous heart condition should not depend on whether a cardiologist happens to be nearby.
With AI-SUSTAINS CVD Care, we want to make high-quality cardiac diagnostic support accessible to healthcare professionals across Ethiopia and create an open, scalable model that can ultimately benefit underserved communities far beyond Ethiopia.