{"683240":{"#nid":"683240","#data":{"type":"news","title":"New Dataset Makes Health Chatbots Like Google\u0027s MedGemma More Mindful of African Contexts","body":[{"value":"\u003Cp\u003EA groundbreaking new medical dataset is poised to revolutionize healthcare in Africa by improving chatbots\u2019 understanding of the continent\u2019s most pressing medical issues and increasing their awareness of accessible treatment options.\u003C\/p\u003E\u003Cp\u003E\u003Ca href=\u0022https:\/\/afrimedqa.com\/\u0022\u003E\u003Cstrong\u003EAfriMed-QA\u003C\/strong\u003E\u003C\/a\u003E, developed by researchers from Georgia Tech and Google, could reduce the burden on African healthcare systems.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe researchers said people in need of medical care file into overcrowded clinics and hospitals and face excruciatingly long waits with no guarantee of admission or quality treatment. There aren\u2019t enough trained healthcare professionals available to meet the demand.\u003C\/p\u003E\u003Cp\u003ESome healthcare question-answer chatbots have been introduced to treat those in need. However, the researchers said there\u2019s no transparent or standardized way to test or verify their effectiveness and safety.\u003C\/p\u003E\u003Cp\u003EThe dataset will enable technologists and researchers to develop more robust and accessible healthcare chatbots tailored to the unique experiences and challenges of Africa.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EOne such new tool is Google\u2019s\u0026nbsp;\u003Ca href=\u0022https:\/\/medgemma.org\/\u0022\u003E\u003Cstrong\u003EMedGemma\u003C\/strong\u003E\u003C\/a\u003E, a large-language model (LLM) designed to process medical text and images. AfriMed-QA was used for training and evaluation purposes.\u003C\/p\u003E\u003Cp\u003EAfriMed-QA stands as the most extensive dataset that evaluates LLM capabilities across various facets of African healthcare. It contains 15,000 question-answer pairs culled from over 60 medical schools across 16 countries and covering numerous medical specialties, disease conditions, and geographical challenges.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ETobi Olatunji and Charles Nimo co-developed AfriMed-QA and co-authored a paper about the dataset that will be presented at the\u0026nbsp;\u003Ca href=\u0022https:\/\/2025.aclweb.org\/\u0022\u003E\u003Cstrong\u003EAssociation for Computational Linguistics (ACL)\u003C\/strong\u003E\u003C\/a\u003E conference next week in Vienna.\u003C\/p\u003E\u003Cp\u003EOlatunji is a graduate of Georgia Tech\u2019s\u0026nbsp;\u003Ca href=\u0022https:\/\/omscs.gatech.edu\/\u0022\u003E\u003Cstrong\u003EOnline Master of Science in Computer Science (OMSCS) program\u003C\/strong\u003E\u003C\/a\u003E and holds a Doctor of Medicine from the College of Medicine at the University of Ibadan in Nigeria. Nimo is a Ph.D. student in Tech\u2019s School of Interactive Computing, where he is advised by School of IC professors \u003Ca href=\u0022https:\/\/mikeb.inta.gatech.edu\/\u0022\u003E\u003Cstrong\u003EMichael Best\u003C\/strong\u003E\u003C\/a\u003E and \u003Ca href=\u0022https:\/\/www.irfanessa.gatech.edu\/\u0022\u003E\u003Cstrong\u003EIrfan Essa\u003C\/strong\u003E\u003C\/a\u003E.\u003C\/p\u003E\u003Ch4\u003E\u003Cstrong\u003EFocus on Africa\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003ENimo, Olatunji, and their collaborators created AfriMed-QA as a response to MedQA, a large-scale question-answer dataset that tests the medical proficiency of all major LLMs. That includes Google\u2019s Gemini, OpenAI\u2019s ChatGPT, and Anthropic\u2019s Claude, among others.\u003C\/p\u003E\u003Cp\u003EHowever, because MedQA is trained solely on the U.S. Medical License Exams, Nimo said it is not adequate to serve patients in underdeveloped African countries nor the Global South at-large.\u003C\/p\u003E\u003Cp\u003E\u201cAfriMed-QA has the contextualized and localized understanding of African medical institutions that you don\u2019t get from Med-QA,\u201d Nimo said. \u201cThere are specific diseases and local challenges in our dataset that you wouldn\u0027t find in any U.S.-based dataset.\u201d\u003C\/p\u003E\u003Cp\u003EOlatunji said one problem African users may encounter using LLMs trained on MedQA is that they may advise unfeasible treatments or unaffordable prescription drugs.\u003C\/p\u003E\u003Cp\u003E\u201cYou consider the types of drugs, diagnostics, procedures, or therapies that exist in the U.S. that are quite advanced. These treatments are much more accessible, for example in the US, and Europe,\u201d Olatunji said. \u201cBut in Africa, they\u2019re too expensive and many times unavailable. They may cost over $100,000, and many people have no health insurance. Why recommend such treatments to someone who can\u2019t obtain them?\u201d\u003C\/p\u003E\u003Cp\u003EAnother problem may be that the LLM doesn\u2019t take a medical condition seriously if it isn\u2019t predominant in the U.S.\u003C\/p\u003E\u003Cp\u003E\u201cWe tested many of these models, for example, on how they would manage sickle-cell disease signs and symptoms, and they focused on other \u201cmore likely\u201d causes and did not rank or consider sickle cell high enough as a possible cause,\u201d he said. \u201cThey, for example, don\u2019t consider sickle-cell as important as anemia and cancer because sickle-cell is less prevalent in the U.S.\u201d\u003C\/p\u003E\u003Cp\u003EIn addition to sickle-cell disease, Olatunji said some of the healthcare issues facing Africa that can be improved through AfriMed-QA include:\u003C\/p\u003E\u003Cul\u003E\u003Cli\u003EHIV treatment and prevention\u003C\/li\u003E\u003Cli\u003EPoor maternal healthcare\u003C\/li\u003E\u003Cli\u003EWidespread malaria cases\u003C\/li\u003E\u003Cli\u003EPhysician shortage\u003C\/li\u003E\u003Cli\u003EClinician productivity and operational efficiency\u003C\/li\u003E\u003C\/ul\u003E\u003Ch4\u003E\u003Cstrong\u003EGoogle Partnership\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003EMercy Asiedu, senior author of the AfriMed-QA paper and research scientist at Google Research, has dedicated her career to improving healthcare in Africa. Her work began as a Ph.D. student at Duke University, where she invented the Callascope, a groundbreaking non-invasive tool for gynecological examinations\u003C\/p\u003E\u003Cp\u003EWith her current focus on democratizing healthcare through artificial intelligence (AI), Asiedu, who is from Ghana, helped create a research consortium to develop the dataset. The consortium consists of Georgia Tech, Google, Intron, Bio-RAMP Research Labs, the University of Cape Coast, the Federation of African Medical Students Association, and Sisonkebiotik.\u003C\/p\u003E\u003Cp\u003ESisonkebiotik is an organization of researchers that drives healthcare initiatives to advance data science, machine learning, and AI in Africa.\u003C\/p\u003E\u003Cp\u003EOlatunji leads the Bio-RAMP Research Lab, a community of healthcare and AI researchers, and he is the founder and CEO of Intron, which develops natural-language processing technologies for African communities.\u003C\/p\u003E\u003Cp\u003EIn May, Google released MedGemma, which uses both the MedQA and Afri-MedQA datasets to form a more globally accessible healthcare chatbot. MedGemma has several versions, including 4-billion and 27-billion parameter models, which support multimodal inputs that combine images and text.\u003C\/p\u003E\u003Cp\u003E\u201cWe are proud the latest medical-focused LLM from Google, MedGemma, leverages AfriMed-QA and improves performance in African contexts,\u201d Asiedu said.\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u201cWe started by asking how we could reduce the burden on Africa\u2019s healthcare systems. If we can get these large-language models to be as good as experts and make them more localized with geo-contextualization, then there\u2019s the potential to task-shift to that.\u201d\u003C\/p\u003E\u003Cp\u003EThe project is supported by the\u0026nbsp;\u003Ca href=\u0022https:\/\/www.gatesfoundation.org\/\u0022\u003E\u003Cstrong\u003EGates Foundation\u003C\/strong\u003E\u003C\/a\u003E and\u0026nbsp;\u003Ca href=\u0022https:\/\/www.path.org\/\u0022\u003E\u003Cstrong\u003EPATH\u003C\/strong\u003E\u003C\/a\u003E, a nonprofit that improves healthcare in developing countries.\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EResearchers introduced a new dataset aimed at improving health chatbots like Google\u0027s MedGemma by better accounting for cultural, linguistic, and contextual factors specific to African settings.\u0026nbsp;\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"A new dataset, AfriMed-QA, was created by researchers at Georgia Tech and Google to improve health chatbots like Google\u0027s MedGemma, making them more aware of African healthcare realities."}],"uid":"36530","created_gmt":"2025-07-23 15:32:10","changed_gmt":"2025-07-23 16:34:15","author":"Nathan Deen","boilerplate_text":"","field_publication":"","field_article_url":"","location":"Atlanta, GA","dateline":{"date":"2025-07-23T00:00:00-04:00","iso_date":"2025-07-23T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"677474":{"id":"677474","type":"image","title":"AdobeStock_181202044.jpeg","body":null,"created":"1753284749","gmt_created":"2025-07-23 15:32:29","changed":"1753284749","gmt_changed":"2025-07-23 15:32:29","alt":"AfriMed-QA","file":{"fid":"261376","name":"AdobeStock_181202044.jpeg","image_path":"\/sites\/default\/files\/2025\/07\/23\/AdobeStock_181202044.jpeg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2025\/07\/23\/AdobeStock_181202044.jpeg","mime":"image\/jpeg","size":95803,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2025\/07\/23\/AdobeStock_181202044.jpeg?itok=s52m9aW9"}}},"media_ids":["677474"],"groups":[{"id":"47223","name":"College of Computing"},{"id":"1188","name":"Research Horizons"},{"id":"50876","name":"School of Interactive Computing"}],"categories":[],"keywords":[{"id":"187915","name":"go-researchnews"},{"id":"9153","name":"Research Horizons"},{"id":"192863","name":"go-ai"},{"id":"193860","name":"Artifical Intelligence"},{"id":"187812","name":"artificial intelligence (AI)"},{"id":"194391","name":"AI in Healthcare"},{"id":"184331","name":"access to healthcare"},{"id":"1724","name":"african"},{"id":"169137","name":"chatbot"},{"id":"193556","name":"large language models"},{"id":"190091","name":"Google AI"}],"core_research_areas":[{"id":"193655","name":"Artificial Intelligence at Georgia Tech"},{"id":"39501","name":"People and Technology"}],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}