Language and geographical bias limits global health research
Research in global health must expand its scope beyond English language publishing to ensure inclusivity, writes Abdourahmane Ndong.
Each time I engage with global health research I am reminded of how many studies never reach the global spotlight or appear in indexed databases because of the language they are written in or the journal in which they were published. This persistent exclusion weakens the principles of evidence-based medicine and highlights the need for greater inclusivity in research.1 For a truly global evidence base, research must urgently expand its linguistic and geographic scope.
A substantial weakness arises from the dominance of publications in English, with estimates suggesting that as much as 98% of scientific output is published in this language.234 When research in non-English languages is published, it is often overlooked—even when it’s rigorously conducted and highly relevant. Language restrictions influence which studies are cited, included in clinical guidelines, and used to inform global health policies, especially in biomedical areas with strong regional variations in health systems or disease burden.5
Some researchers argue that including only English language research does not have a substantial impact and can be a reasonable choice when resources and time are limited. But this approach promotes uniformity and a lack of inclusivity, limiting diverse perspectives, missing key global insights, and reinforcing global power imbalances in research.67 Current guidelines, including those provided by PRISMA, do not require authors to include studies in multiple languages.8 As author Ngũgĩ wa Thiong’o reminds us, every language “carries its own memory of the world.”9
Beyond language, a critical geographic bias also persists. A substantial body of research, particularly from low and middle income countries or specific regions in Africa, is published in local or national journals that are often not indexed in major international databases.10 A study in 2023 showed that of 560 active African biomedical journals, only 96 (17.14%) were included in at least one of four major international databases (Web of Science, Scopus, Embase, MEDLINE).11 In comparison, an analysis of biomedical journals indexed in Directory of Open Access Journals reported that over 75% of such journals globally were listed in at least one of four major biomedical databases.12
This lack of indexing of African journals frequently stems from organisational, financial, or human resource limitations, rather than reflecting a lower quality of publication.
I have been part of several teams conducting systematic reviews on topics related to health in sub-Saharan Africa. We wanted to find out if relevant data existed that had been written in different languages and published in non-indexed journals, but which offered valuable insights for policy and clinical decision making. We therefore made it mandatory not to apply language restrictions and to include at least one language other than English, translating search keywords and eligible articles using available platforms.1314 We also included African databases such as African Journals OnLine and African Index Medicus and used supplementary sources when necessary.1516 Checking the first pages of results on Google Scholar or ResearchGate may still be used as a supplementary source to identify grey literature or unpublished studies,1718 as well as directly contacting editors from local journals in specific subjects.19 This approach was effective in increasing study inclusion in most reviews. In several cases, more than half of the articles included were identified through these additional search strategies. Importantly, this enabled the generation of evidence in research areas where the published literature had previously suggested that data from African settings were scarce.
Our experiences taught us that several approaches can meaningfully improve representation and language and geographic bias and barriers can be overcome with deliberate action. Indexing systems must expand their coverage of high quality regional journals, and authors and journals should justify language restrictions and support multilingual dissemination. Funding mechanisms need to include resources for developing and strengthening local journals so that they can meet indexing standards. Empowering local journals is essential to enhance their visibility, and training researchers to manage and develop such journals will be critical for sustained impact.
Authors and editors must be aware of their responsibility to support geographic and language inclusion and balance this with the extra work, time, and resources needed to facilitate it. We can leverage responsible use of artificial intelligence tools to reduce the practical barriers that once justified language exclusions. Translation that used to be costly, time consuming, and logistically challenging is now much easier and more accessible.20 Large language models can help translate search queries and articles.2122
If global health research is to guide equitable clinical practice and policy, it must evolve to reflect the diversity of scientific production. Greater multilingual and geographic inclusion will help build a truly global and representative knowledge base.
Ndong A. Language and geographical bias limits global health research BMJ 2026; 392 :s436 doi:10.1136/bmj.s436