Discovering new pathways toward integration between health and sustainable development goals with natural language processing and network science

Globalization and Health

Research

Thomas Bryan Smith¹  Raffaele Vacca²  Luca Mantegazza³  Ilaria Capua⁴ 

[1] Bureau of Economic and Business Research, University of Florida, PO Box 117148, nd Ave Ste 150, 32611, Gainesville, FL, USA;Department of Social and Political Sciences, University of Milan, Milan, Italy;One Health Center of Excellence, IFAS, University of Florida, Gainesville, FL, USA;One Health Center of Excellence, IFAS, University of Florida, Gainesville, FL, USA;Johns Hopkins University, SAIS Europe, Bologna, Italy;

关键词: Sustainable development goals;    One health;    Natural language processing;    Network science;    Topic modeling;   

DOI  :  10.1186/s12992-023-00943-8

 received in 2023-02-20, accepted in 2023-05-30,  发布年份 2023

来源: Springer

【 摘 要 】

BackgroundResearch on health and sustainable development is growing at a pace such that conventional literature review methods appear increasingly unable to synthesize all relevant evidence. This paper employs a novel combination of natural language processing (NLP) and network science techniques to address this problem and to answer two questions: (1) how is health thematically interconnected with the Sustainable Development Goals (SDGs) in global science? (2) What specific themes have emerged in research at the intersection between SDG 3 (“Good health and well-being”) and other sustainability goals?MethodsAfter a descriptive analysis of the integration between SDGs in twenty years of global science (2001–2020) as indexed by dimensions.ai, we analyze abstracts of articles that are simultaneously relevant to SDG 3 and at least one other SDG (N = 27,928). We use the top2vec algorithm to discover topics in this corpus and measure semantic closeness between these topics. We then use network science methods to describe the network of substantive relationships between the topics and identify ‘zipper themes’, actionable domains of research and policy to co-advance health and other sustainability goals simultaneously.ResultsWe observe a clear increase in scientific research integrating SDG 3 and other SDGs since 2001, both in absolute and relative terms, especially on topics relevant to interconnections between health and SDGs 2 (“Zero hunger”), 4 (“Quality education”), and 11 (“Sustainable cities and communities”). We distill a network of 197 topics from literature on health and sustainable development, with 19 distinct network communities – areas of growing integration with potential to further bridge health and sustainability science and policy. Literature focused explicitly on the SDGs is highly central in this network, while topical overlaps between SDG 3 and the environmental SDGs (12–15) are under-developed.ConclusionOur analysis demonstrates the feasibility and promise of NLP and network science for synthesizing large amounts of health-related scientific literature and for suggesting novel research and policy domains to co-advance multiple SDGs. Many of the ‘zipper themes’ identified by our method resonate with the One Health perspective that human, animal, and plant health are closely interdependent. This and similar perspectives will help meet the challenge of ‘rewiring’ sustainability research to co-advance goals in health and sustainability.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

40517_2023_259_Article_IEq46.gif

Fig. 4

40517_2023_259_Article_IEq18.gif