Artificial Intelligence in Healthcare: 2024 Year in Review Dataset

BackgroundResearch related to Artificial Intelligence (AI) in healthcare applications is evolving. It is essential to incorporate collaborative learning from published research to comprehend the challenges and accessibility of opportunities when integrating AI in healthcare systems. To investigate the role of AI, a qualitative and quantitative year in review study was conducted, encompassing the evaluation of literature published in 2024 to gain insight into the recent advancements of the field.MethodsTo find research articles about integrating new AI technologies into healthcare systems, a PubMed search using the terms “2024”, “artificial intelligence”, and “large language models” was conducted. The search was restricted to human subject research and used a deep-learning-based approach to assess the reliability of publications as of December 31, 2024 on January 1, 2025. In addition, for each publication, each mature article was manually annotated for the AI model type (e.g., LLM, DL, ML), healthcare specialty, and the data type used (image, text, tabular, or audio).Additionally,qualitative and quantitative analyses were performed to illuminate statistics and trends of combined published articles.ResultsOur PubMed search yielded 28,180 total articles; 1,693 were initially labeled mature, after which 1,551 articles were analyzed after exclusions. Similar to the prior years, we excluded systematic reviews in the final analysis and were excluded in this year's dataset.The most prevalent specialties within our PubMed search originated from imaging (407), head and neck (127), and General (122). Analysis of AI model types showed that the Large Language Model (LLM) was the most popular utilized in 479 publications, followed by AI General (448), and DL (372). Qualitative data was obtained on the data types, and it was revealed that the image data was predominant and used in 57.0% of the mature sources, followed by text (33.1%), followed by tabular (7.59%). The utilization of Large Language Models (LLMs) is the highest in publications associated with education at 18.6%, followed by General at 13.6%. These results indicate that LLMs are frequently applied in educational contexts and administrative tasks amongst the healthcare specialties for research.ConclusionHealthcare specialties, including imaging, head and neck, and general medicine, have taken over the realm of AI in healthcare. Other specialties that distinctive types of AI and LLMs could likely drive in the future include education, pathology, as well as surgery. It is essential to use a collaborative approach to investigate the multimodal models of AI in healthcare applications to provide a thorough encapsulation of AI in healthcare.Data Files DescriptionOne data file is provided, which illustrates the annotations of the mature sources used in our review. The first file is named Annotated_OnlyMature_Unique_2024_YIR_All_Publications - Annotated_OnlyMature_Unique_2024_YIR_All_Publications and includes ‘Title’, ‘DOI’, ‘Abstract’, ‘Author Address’, ‘Specialty’, ‘Model’, and 'Data Type’. The ‘Specialty’, ‘Model’, and ‘Data Type’ were predominantly analyzed by the BrainXAI research team to produce our meta-analysis of the mature sources of AI. This year we have excluded systematic reviews from the dataset compared to the 2023 year in review dataset, but can be provided on request.