health transformation
knowledge portal
Joaquim Cardoso MSc
Founder and Chief Researcher and Strategy Officer (CRSO)
Editor in Chief
April 2, 2024
What is the message?
The central message of this article is to assess the current trends and impact of artificial intelligence (AI) in the field of medicine through a bibliometric analysis of the top 100 most cited original articles.
What are the key points?
Introduction to Bibliometric Analysis: The article highlights the significance of bibliometric analysis in understanding the progress and trends in a specific field, which aids researchers and funding agencies in identifying areas for future research.
Rise of AI in Medicine: It discusses the increasing interest and adoption of AI in medicine since the 1950s, with the potential to assist in diagnostics, drug prescriptions, workload reduction, and precision medicine.
Methodology: The study conducted a bibliometric analysis using Scopus database, extracting original articles relevant to AI in medicine without specific time constraints. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed for article selection.
Results: The analysis of 100 most cited articles revealed a significant increase in research output in recent years, with most articles published between 2015 and 2020. The United States contributed the highest number of articles, followed by the United Kingdom and Germany.
Discussion: It compares the research trends in AI in medicine with other medical fields, highlighting the rapid growth and evolution of AI research. The phenomenon of “obliteration-by-incorporation” is discussed, where classic articles may not be cited as frequently due to widespread use of their findings.
Conclusion: The review identifies the current trends, contributions, and limitations in AI research in medicine, aiming to guide future research directions and funding allocation.
DEEP DIVE
Authors: Fatima Zahoor, Muhammad Abdullah, Muhammad Waleed Tahir, Asif Islam
Abstract
Objective
To assess the current trends in the field of artificial intelligence in medicine by analysing 100 most cited original articles relevant to the field.
Methods
The bibliometric analysis was conducted in September 2022, and comprised literature search on Scopus database for original articles only. Google and Medical Subject Headings databases were used as resources to extract key words. In order to cover a broad range of articles, original studies comprising human as well as non-human subjects, studies without abstract and studies in languages other than English were part of the inclusion criteria. There was no specific time period applied to the search and no specific selection was done regarding the journals in the database. The screening was done using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines to extract the top 100 most cited articles in the field of artificial intelligence usage in medicine. Data was analysed using SPSS 23.
Results
Of the 11,571 studies identified, 100(0.86%) were analysed in detail. The studies were published between 1986 and 2021, with a median of 43 citations (IQR 53) per article. The journal ‘Artificial Intelligence in Medicine’ accounted for the highest number 9(9%)) of articles, and the United States was the country of origin for most of the articles 36(36%).
Conclusion
The trends, development and shortcomings in field of artificial intelligence usage in medicine need to be understood to conduct an effective research in areas that still need attention, and to guide the authorities to direct their funding accordingly.
Introduction
Evidence-based approach is important in revolutionising the field of medicine and research. It guides the need to make changes in old practices and bring forth new methods that could benefit science and humanity. In this era of evidence-based medicine (EBM), bibliometric analysis has become an important tool that guides researchers and funding agencies about the progress made in a particular field to date and identifies the areas for future research prospects. Therefore, there is growing interest in bibliometric analysis to find the most significant and impactful papers, their authors, countries of origin and affiliations.
Bibliometric analysis is a tool which focuses on extracting data from published scientific literature to see research progress and trends in a specific field and provide physicians with the most impactful data in a particular field. It is useful not only for the researchers, but also guides funding agencies and the government that finance the researchers to allocate their funding in the most useful way.
Artificial intelligence (AI) is rapidly making its place in the field of medicine. AI means the use of machines and computers to accomplish the task that usually requires human intelligence. There has been a rise in interest in AI since the 1950s. The advocates of AI believe that it could help in diagnosing patients, prescribing drugs and determining the prognosis. It could also help to reduce the workload in hospitals and can help in precision medicine. A huge body of literature is available on the subject of AI in medicine with variations in quality. Therefore, there is a need to identify the current trends in the field that would help to guide those who have limited knowledge of this field. The current systematic review was planned to analyse the top 100 most cited original articles across the full range of specialities related to AI in medicine.
Results
Of the 11,571 studies identified, 100(0.86%) were analysed in detail (Figure 1). The studies were published between 1986 and 2021, with a median of 43 citations (IQR 53) per article (Table 2). The total number of citations was 9538, and the number of citations per year ranged from 10 to 151.57, with median and mean citations per year being 14 (IQR 22) and 23.25 (SD 25.92), respectively.
A rapid increase was seen starting from 1998 and it peaked in 2015 after which it declined with minor fluctuations (Figure 2). The number of publications in each 5-year phase was the highest in the 2015-20 period (Figure 3).
Table-1: Primary key words used in the search strategy.
Artificial Intelligence Machine Learning Computational Intelligence Computer Reasoning AI (Artificial Intelligence) Robotics Machine Intelligence Neural Networks Expert System Knowledge Engineering Intelligent Retrieval Deep learning Natural Language Processing Data Mining Fuzzy Medical Medicine Surgical Surgery Healthcare
Figure-1: Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.
Table 2: Top 100 articles, their citations, and citations per year.
Table-3: Top 5 journals in terms of highest number of cited articles.
The journal ‘Artificial Intelligence in Medicine’ accounted for the highest number 9(9%) of articles. The impact factor of these journals ranged from 87.241 to 0.941. The impact factor of the 5 journals which contributed most of the articles to our bibliometric analysis ranged from 2.021 to 15.36 (Table 3).
The 100 top-cited articles originated from 37 different countries, with the highest number of articles originating from the United States 36(36%), followed by the United Kingdom 13(13%) and Germany 12(12%). Of the total, 2(2%) studies had unidentified origin (Figure 4).
The institutions affiliated with the selected articles were Harvard Medical School and its affiliated institutes, Brigham and Women’s Hospital and Massachusetts General Hospital, which accounted for 11(11%) articles, followed by Univerza v Ljubljani 4(4%), University of California, Irvine, Yale School of Medicine, University of California, San Francisco 3(3%) publications each.
The articles were funded by some organisations, with the National Institute of Health (NIH) funding of most of the articles 8(8%), followed by other organisations (Figure 5). The biggest part of the contribution was from the field of Medicine, followed by Computer Science (Table 4).
Figure-2: Total citations of the articles in the top-100 list every year.
Table-4: Publication distribution by research domains.
Figure-3: Number of Publications in each 5-year period.
Discussion
The 100 most cited articles (Table 2) showed the extent of research and the evolving trends in the field of AI in medicine. The articles were published between 1986 and 2021, but most of the articles (n=65) were published between 2015 and 2020, which signifies a marked increase in research on the subject in recent years that is accompanied by its widespread use in the medical field and the development of the field of AI in recent times.
Compared to bibliometric studies in other fields of medical practice, like General Surgery, Orthopaedics and Neurosurgery that have published their highly influential articles before the 1980s, there is a great rise in research trends in the field of AI in medicine in recent years. However, the study of some other fields, like cardiovascular magnetic resonance and diabetes mellitus have their most impactful research done in recent years, like that of AI.
The phenomenon of obliteration-by-incorporation suggests that some classic and highly impactful articles may not be cited as frequently as before because the information extracted from them has become so widespread in use in the field of medicine that researchers do not specifically cite them in their studies.
Moreover, the trend of bibliometrics towards citing recently published papers that show the rapidly changing medical practices with the passing of time may lead to the exclusion of some important research papers. The absence of articles published before 1986 in the current bibliometric analysis highlights that old articles are not relevant anymore as research trends have changed with time. The number of citations of an article reflects the overall impact of an article, but the current influence of an article cannot be perceived by just looking at this number alone.
Figure-4: Number of articles originating from each country.
Figure-5: Top-10 funding sources of the articles.
To have an idea of the impact of an article in a particular year, the element of ‘citations per year’ (Table 2) was used. The article, ‘Machine learning in medicine’, had the most number of citations and citations per year, which, in part, was due to its old publication year (2015) and it was also funded by institutes, like the NIH and the National Heart, Lung and Blood Institute. Moreover, it incorporated many research domains, like Medicine and Computer Science, and was published in a high impact factor journal. All these factors contributed to the high citation score.
The highest number of articles (n=9) were published in the journal, ‘Artificial Intelligence in Medicine’. Other articles were published in a broad variety of journals encompassing various fields, like Medicine, Biochemistry, Computer Science, Engineering, Social Sciences and Nursing. The top most cited articles were published in high impact factor journals with high citation index which showed the growing interest of research in the field of AI in medicine.
In the current bibliometric analysis of the 100 most cited articles on AI in medicine, the greatest number of articles on the list (n=36) originated from the US, which is probably attributed to the funding support provided by the NIH and various other platforms that promote research. Several European countries, like the United Kingdom and Germany followed, but there were only a few articles contributed from Asian countries, which can be attributed to the lack of research programmes, lack of motivation to do research, unwillingness to spend funds on research purposes, and an almost non-existent development of research institutes in these countries.
The current review has its limitations. It is reported that Scopus is likely to miss old citations prior to the 1980s, which could have resulted in the exclusion of the articles from the past . Also, self-citations could lead to bias in the bibliometric analysis, which was expected as many of the authors had collaborated on many studies. Besides, some highly impactful articles published recently may not have made to the list of top 100 most-cited articles as citations number builds up over some period of time, and the articles published long ago have a greater number of citations which is because of the increased number of citations over time. Despite the limitations, the current review provided a thorough understanding of the most frequently cited and impactful articles relating to AI usage in medicine.
Conclusion
The review identified the current trends in the field of AI in medicine using studies indexed with the Scopus database. These studies were carefully selected from medical literature and will serve the researchers in the use of this emerging topic to transform the field of medicine. The analysis showed the growth of healthcare-related AI publications in the last decade because of their adoption in the field of medicine to transform healthcare, and also helped in the identification of countries and institutions that have made the most contribution. This will help future researchers in identifying the emerging patterns, and the need to explore the domain further.
Method & References
See the original publication
