Artificial Intelligence Research Receives Similar Online Attention but Increased Citation Rates Compared With Control Articles

Abstract

Purpose To compare the social media attention and citation rates between artificial intelligence (AI) and non‐AI research in the orthopaedic surgery literature. Methods Research articles using AI for the purpose of statistical analysis or evaluating the use of AI for specific clinical use cases were included. Articles selected from 6 high‐impact orthopaedic surgery journals between January 2018 and December 2023 were selected by month of publication in a 1:4 ratio within the same journal issue. The primary outcomes were the Altmetric Attention Score (AAS), an aggregate score comprising the degree of social media attention received on multiple online platforms (with higher scores indicating more attention), and cumulative citation rates. Multivariate linear regression analyses were conducted to assess the relationship (1) between article type and the AAS and (2) between article type and the citation rate. A subanalysis comparing the AAS and citation rates between articles using AI for the purpose of statistical analysis versus evaluating the use of AI for a specific clinical use case was performed. Results A total of 540 articles (110 AI‐related articles [20.0%] and 440 control articles [80.0%]) were included. The median AAS was 2.0 (interquartile range [IQR], 1‐6) for AI‐related articles and 1.0 (IQR, 0‐5) for control articles ( P = .006). However, when controlling for covariates in the multivariate regression, article type (AI vs control) was not significantly associated with the AAS (β = –1.1, P = .602). The median number of citations was significantly greater among AI‐related articles (median, 18 [IQR, 9‐35]) compared with control articles (median, 9 [IQR, 4‐16]) ( P < .001). Multivariate linear regression showed an additional average increase of 13.4 citations (95% confidence interval, 9.4‐17.5; P < .001) per article for AI articles compared with control articles. Conclusions AI research published in 6 orthopaedic surgery journals received approximately 13 additional citations on average compared with control articles published during a similar period. AI research did not receive greater social media attention on average after controlling for confounding variables. Clinical Relevance AI research in orthopaedic surgery did not receive greater online attention but was cited more frequently relative to control articles. Given the increasing academic impact of these articles, orthopaedic surgeons should become familiarized with AI research as clinical applications and AI solutions derived from such research become increasingly relevant and implemented within clinical practice workflows.

Ähnliche Arbeiten