Data Governance Challenges in Artificial Intelligence-Enabled Healthcare in Low- and Middle-Income Countries
Elona Erezi
*
Department of Public Health, Faculty of Basic and Applied Biological Science, Ahmadu Bello University, Zaria, Kaduna State, Nigeria.
Kehinde Jonathan Irhodia
Department of Biotechnology, School of Life Science, Federal University of Technology, Akure, Ondo, Nigeria.
Hussaini Abba Disa
Department of Dental Surgery, State specialist Hospital, Damaturu, Yobe State, Nigeria.
Osefanmen Matthew Enosolease
School of Medicine, College of Medical Sciences, University of Benin, Benin City, Edo State, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Background: Artificial intelligence (AI) has the potential to transform healthcare delivery in low- and middle-income countries (LMICs), where the disease burden from preventable causes is high and healthcare resources are severely limited. Although major technological breakthroughs have occurred, systemic, regulatory, and infrastructural barriers continue to hinder the implementation of AI-powered health solutions in these settings.
Objectives: This review aims to examine and synthesise key data governance (DG) challenges and concerns in the context of health systems in LMICs, develop an analytical framework for assessing readiness to address data governance issues in AI-enabled health systems, and identify key areas for data governance policy interventions and future research.
Methods: A structured narrative review of peer-reviewed literature from PubMed, Scopus, and Web of Science, together with relevant grey literature sources from January 2015 to December 2023, was conducted. AI governance, digital health infrastructure, digital health data regulation, and algorithmic accountability in LMICs were considered. Findings were analysed thematically to synthesise governance concerns across key governance dimensions.
Results: Seven major governance challenge domains were identified: (1) weak and fragmented regulatory frameworks; (2) poor data quality and limited interoperability; (3) inadequate patient data protection mechanisms; (4) algorithmic bias due to undersampling of local populations; (5) inadequate digital health infrastructure; (6) limited digital literacy among healthcare professionals; and (7) ethical tensions related to consent, privacy, and community trust. Cross-cutting themes of power asymmetry in global AI development and donor dependency in digital health financing were also identified.
Conclusions: The promise of AI for LMIC health systems needs to be supported by robust, place-sensitive governance structures that extend beyond technology alone. Before equitable AI adoption can be pursued, investments in regulatory capacity, data systems, algorithmic accountability mechanisms, and stakeholder engagement are critical preconditions. A Governance Readiness Framework is proposed to support policymakers, health ministries, and development partners.
Keywords: Artificial intelligence, data governance, healthcare systems, low- and middle-income countries, digital health policy, algorithmic bias