Ministry Rolls Out AI Tool To Support Frontline Health Workers In Remote Areas | The Reporter Ethiopia

The Ministry of Health, in collaboration with the US-based Last Mile Health, has launched HEP Assist, an artificial intelligence-powered assistant designed to support health extension workers (HEWs) serving in remote parts of the country.

The launch was accompanied by a media training and field visit in Arba Minch, where journalists were introduced to the pilot implementation of the tool. The initiative is currently being rolled out across seven regions and 15 woredas.

The pilot areas include Oromia, Sidama, Afar, South Ethiopia, Central Ethiopia, and Harari regions, as well as the Dire Dawa city administration. According to Dr. Ruth Diriba, a senior program manager at Last Mile Health, said these locations were selected due to their limited access to higher-level health facilities and the increased need for clinical decision support among HEWs.

HEP Assist uses a Retrieval-Augmented Generation (RAG) system—an AI framework that enhances the accuracy of large language models by retrieving relevant, up-to-date information from curated sources. In this case, the system draws exclusively from 100 Ministry of Health guidelines.

Before the introduction of HEP Assist, HEWs relied on calling health professionals at nearby health centers for guidance on complex cases. However, this system proved unreliable due to network disruptions, busy lines, and inconsistent medical advice.

Data from Last Mile Health indicates that out of 18,000 calls made to call agents, only 7,000 were successfully answered, largely due to the competing responsibilities of health professionals.

Since its introduction in January 2025, HEP Assist has facilitated over 6,000 consultations—initially through call agents and more recently through direct use by HEWs, according to officials.

The tool is accessible via internet-enabled devices and requires an email login. HEWs can submit queries using voice or text in multiple languages, including Amharic, Afarigna, Oromiffa, Tigrinya, Somali, and English. The system then provides concise, guideline-based responses with cited references.

The platform does not require patient-identifying information, helping maintain confidentiality. Developers are also working on an offline version to address connectivity challenges in remote areas, according to program managers.

In Mirab Abaya Woreda, Gamo Zone, South Ethiopia, HEWs report significant improvements in service delivery. The woreda includes hard-to-reach kebeles such as Lao Tirga, located far from urban centers with limited infrastructure.

Aynet Adalo, a health extension worker in the area, said the tool has reduced unnecessary referrals and improved her ability to manage cases independently. She recalled treating a woman experiencing prolonged and heavy menstrual bleeding following contraceptive use—cases that previously would have required referral to a health center.

“Now we can access the necessary guidance immediately and provide care at the health post level,” she said.

Supervisors also report reduced patient load at referral facilities. Belay Osso, a health center supervisor in the woreda, noted a decline in referrals from the three kebeles and four health posts using the tool. He added that eight HEWs have been trained to use HEP Assist, with health centers continuing to provide technical and logistical support, including electricity and internet access.

Gemechis Melkamu, digital health lead at the Ministry of Health, said the initiative aligns with the country’s broader digital transformation agenda under Digital Ethiopia 2030.

He noted that digital tools such as the Electronic Community Health Information System (ECHIS) are already operational in over 9,000 health posts nationwide. HEP Assist, currently in its pilot phase, is expected to expand further following evaluation.

“The plan is to scale this solution across all woredas after assessing its performance,” he said.