Data sharing barriers and enablers for healthcare Artificial Intelligence (AI) in Low- and Middle-income countries
More details
Hide details
Saw Swee Hock School of Public Health, National University of Singapore, Singapore
National University of Singapore
Publication date: 2023-04-27
Popul. Med. 2023;5(Supplement):A573
Background: Health systems in low- and middle-income countries (LMICs) can greatly benefit from AI interventions, including electronic decision support tools and predictive analytics. Unfortunately, challenges to data sharing pose a significant obstacle to the development and use of AI in healthcare. Objective: This systematic review identified barriers to and enablers for data sharing related to AI in healthcare in LMICs. Methods: A systematic literature search was performed using PubMed, SCOPUS, Embase, Web of Science and ACM (Association for Computing Machinery) for articles describing barriers to and enablers for data sharing for AI in healthcare in LMICs. Qualitative data extraction and thematic analysis was conducted on selected studies. Barriers and enablers were classified according to a framework formulated by merging two existing framework (Panhuis 2014 and Sun and Medaglia 2019), comprising eight categories: 1) Technical, 2) Motivational, 3) Economic, 4) Political, Legal and Policy, 5) Ethical, 6) Social, 7) Organisational and Managerial and 8) Data-related. Results: The systematic search identified 2471 records, of which 22 met the eligibility criteria. The studies were from 13 countries, with the majority from Africa (12 studies, 55%) and Asia (6 studies, 27%). The most important barriers were technical (e.g. lack of hospital infrastructure) and data-related (e.g. lack of interoperability standards). Significant enablers for data sharing included political and regulatory enablers (e.g. government support, clear policies and guidelines on data sharing) and technical enablers (e.g. shifting from paper-based to electronic data collection). Conclusions: This systematic review identified various barriers and enablers relevant to LMICs. These results can inform context-specific recommendations to promote local AI development in resource-limited settings. Recommendations and best practices arising from this review can also indicate ways to mitigate common data sharing barriers for transitioning economies and health systems.
Journals System - logo
Scroll to top