“Imagine a world where diagnosing viral infections like common cold, influenza, COVID-19, measles or chickenpox is fast and affordable”
Published in Research Data
Feb 25, 2025
Lecturer and Researcher, Aksum University
Liked by India Ambler
Read the paper
Viral infections have always been a major public health concern and the COVID-19 pandemic made it clear how vital it is to have a reliable, fast and efficient method of diagnosing infections. In many parts of the world, especially in regions with strained healthcare systems, diagnosing viral infections often involves expensive, time-consuming processes that are out of reach for many people.
In Tigray, the healthcare system has faced severe challenges due to a shortage of trained medical professionals, limited medical supplies, and the ongoing effects of conflict. In this context, the need for an efficient, low-cost, and easily accessible diagnostic system is more critical than ever. That’s why we turned to AI.
The Birth of CSDIS: An AI-Driven Solution
The idea behind the CSDIS prototype was born out of a desire to bridge the gap between modern healthcare technologies and the realities of healthcare in resource-limited settings. We wanted to develop a system that could assist healthcare workers in diagnosing viral infections by analyzing patient symptoms and providing accurate predictions.
CSDIS uses a rule-based approach, meaning it operates by following a predefined set of rules based on knowledge gathered from healthcare professionals and existing medical literature. These rules help the system identify key symptoms and differentiate between similar diseases. For example, COVID-19 shares symptoms with other diseases like the common cold and influenza, making it challenging for healthcare professionals to diagnose correctly, especially without extensive testing.
To build CSDIS, we worked closely with healthcare professionals in Tigray. We gathered data from patients, including symptoms and medical history, to create a robust knowledge base that the system could use to make accurate predictions. The prototype was designed to be user-friendly, allowing healthcare workers to input patient symptoms and receive an instant diagnosis, along with treatment recommendations.
The Methodology Behind the System
To ensure the accuracy of CSDIS, we used a dataset of 1,500 patient cases, which was divided into five categories: common cold, influenza, COVID-19, measles, and chickenpox. These diseases were selected based on their prevalence in Ethiopia and the fact that they share symptoms with COVID-19.
The system was tested using 70% of the dataset for training and 30% for testing. The results were impressive: CSDIS achieved an accuracy rate of 98% in identifying these diseases, with a 96% acceptance rate from local healthcare workers. This is significant because it means that even in a resource-constrained environment like Tigray, healthcare professionals can rely on AI to make accurate diagnoses.
The CSDIS Prototype in Action
One of the standout features of CSDIS is its architecture. The system includes a user interface that supports multiple languages, including Tigrigna and Amharic, which are spoken in the Tigray region. This makes the system accessible to a wider population, ensuring that language barriers don’t prevent people from benefiting from this technology.
The system also includes an inference engine, which uses the knowledge base to identify the most likely diagnosis based on the patient’s symptoms. Additionally, a learning and memorization (LM) component allows the system to update itself as new information becomes available, ensuring it stays current with the evolving nature of viral diseases like COVID-19.
For example, as new variants of COVID-19 emerge, CSDIS can learn from past cases and incorporate new data into its decision-making process. This adaptability is crucial for maintaining the system’s relevance and effectiveness over time.
Real-World Impact: Evaluating CSDIS
To evaluate the effectiveness of CSDIS, we conducted tests in three health centers in Tigray: Wukro-Maray, Selekleka, and Gendebta. The results were overwhelmingly positive. Healthcare workers found the system to be user-friendly, accurate, and highly valuable in diagnosing viral infections.
In one instance, a healthcare worker in a remote health center was able to diagnose a suspected case of Influenza in minutes, using only the system’s recommendations and the patient’s symptoms. This saved valuable time and helped ensure that the patient received the right treatment as quickly as possible.
Additionally, the system was evaluated based on user acceptance, with local healthcare workers providing feedback on its usability. The results showed that CSDIS was well-received, with a high acceptance rate from the community. Healthcare workers appreciated how the system could assist them in making informed decisions, especially in areas where they might not have specialized knowledge in diagnosing viral infections.
Looking Forward: The Future of AI in Healthcare
The success of CSDIS in Tigray highlights the transformative potential of AI in healthcare, particularly in underserved regions. By leveraging AI technology, we can make healthcare more accessible and efficient, even in the face of resource shortages.
The impact of CSDIS extends beyond just diagnosis; it also provides healthcare workers with a tool that can guide treatment decisions, improving patient outcomes. With further development, the system could be expanded to diagnose and treat a wider range of diseases, becoming an essential tool in global healthcare.
As AI technology continues to evolve, the potential for systems like CSDIS to revolutionize healthcare worldwide is limitless. By making use of data-driven approaches, we can create solutions that adapt to local needs, ensuring that no one is left behind, regardless of where they live.
Conclusion
The story behind this research is one of innovation, collaboration and determination. By combining AI technology with local healthcare knowledge, we have developed a tool that can help healthcare workers in Tigray and potentially other parts of Ethiopia and the world make more accurate, faster diagnoses of viral infections. This could save lives, reduce the spread of diseases and ultimately improve healthcare delivery in resource limited settings.
We hope that this research serves as a model for future projects that use technology to address global health challenges. In a world that is increasingly interconnected, AI-powered healthcare solutions like CSDIS could be the key to a healthier, more equitable future for all.
Lecturer and Researcher, Aksum University
A dedicated teaching and research professional with a strong focus on leveraging AI technologies for healthcare, agriculture and manufacturing solutions. His approach is methodical, guided by a well-structured project timeline that includes stages like planning, tool development, data collection, analysis and dissemination. https://et.linkedin.com/in/dawit-teklu-weldeslasie