Modeling Research Topics for Artificial Intelligence Applications in Medicine: Latent Dirichlet Allocation...

Area of Science:

• Bibliometrics
• Artificial Intelligence in Medicine
• Health Informatics

Background:

• Rapid advancements in AI technologies present numerous applications in medicine and healthcare.
• A significant gap exists in comprehensive reporting on AI's productivity, workflow, topics, and research landscape within this field.

Purpose of the Study:

• To evaluate the global scientific publication development related to AI in medicine.
• To construct interdisciplinary research topics on the theory and practice of AI in medicine from 1977 to 2018.

Main Methods:

• Bibliographic data and abstract content from 1977-2018 were sourced from the Web of Science database.
• Latent Dirichlet Allocation (LDA) was used for research topic classification.
• Principal Component Analysis (PCA) identified the construct of the research landscape.

Main Results:

• AI applications primarily impact clinical settings (prognosis, diagnosis, surgery, rehabilitation) and data science (precision medicine).
• Ethical and legal issues, particularly data privacy, are prominent in AI policy-making.
• AI adoption is limited in resource-poor settings due to infrastructure and human resource constraints.

Conclusions:

• AI in medicine is rapidly growing, focusing on clinical practices, materials, and policies.
• AI has the potential to reduce health inequalities between developing and developed countries.
• Technology transfer and support are crucial for advancing AI in healthcare in developing nations.