Large language models (LLMs) have rapidly emerged as transformative tools across multiple domains, including healthcare. The ability of LLMs to process vast amounts of data and generate human-like responses has facilitated their integration into patient care, particularly in enhancing communication, improving patient satisfaction, and streamlining administrative processes. Despite this potential, there are concerns regarding their accuracy, reliability, and ethical use in clinical settings. This scoping review aims to investigate and map the current literature on the use of LLMs in improving provider-patient experience and interaction efficiency. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews guidelines, we conducted a systematic search of Ovid MEDLINE, PubMed, and Google Scholar databases to identify relevant articles published between January 2015 and June 2024. Of the 3568 articles initially screened, 47 satisfied the inclusion criteria. These articles spanned 13 countries and encompassed diverse healthcare settings. Thematic areas of LLM utilization included improving communication between patients and healthcare providers, resolving patient inquiries, enhancing patient education, and increasing operational efficiency. Although numerous studies have yielded positive outcomes, significant challenges related to data accuracy, hallucinations, bias, and ethical concerns remain. LLMs can considerably improve patient experience in healthcare, particularly in areas of communication, education, and administrative efficiency. However, concerns regarding accuracy, ethical implications, and the need for rigorous safeguards to prevent misinformation impede their widespread adoption. Future research should focus on developing context-specific LLMs tailored to healthcare environments and addressing the identified limitations to optimize their implementation in clinical practice.

Artificial intelligence (AI) is rapidly transforming orthopedic surgery, particularly in total joint arthroplasty (TJA), offering new possibilities for improving patient outcomes. Thus, this narrative review examines the current applications and future directions of machine learning (ML) in hip, knee, and shoulder arthroplasty, focusing on predictive models for clinical outcomes, complications, and patient-reported outcome measures (PROMs). Preoperatively, ML algorithms have shown promise in identifying implants, predicting implant sizes, and assessing implant positioning on radiographs. In outcome prediction, ML models have been developed to predict PROMs, readmissions, length of stay, and healthcare costs associated with TJA. By analyzing large datasets to generate personalized predictions for patients, these models represent a novel approach to assist clinicians in individualized patient decision-making. Furthermore, AI has shown promise in predicting specific post-operative complications, such as dislocations, implant loosening, and prolonged opioid use, highlighting its value in improving surgical planning and patient management. Looking ahead, AI holds the potential to revolutionize orthopedic surgery by equipping clinicians with valuable tools to enhance decision-making and improve patient outcomes. However, the current efforts are shadowed by the challenges of transparency and validation of AI models. As AI continues to find utility in orthopedic clinics and operating rooms, efforts to enhance transparency and validate models will be crucial in realizing its full potential in orthopedic surgery.

Designing highly targeted, selective drugs with desirable absorption, distribution, metabolism, excretion, and pharmacokinetic (PK) profiles; single-digit nanomolar efficacy; and a wider therapeutic index are challenging. In the traditional drug discovery process, researchers screen thousands of chemical compounds during pre-clinical development, progressing through hit identification, lead optimization, and candidate selection to shortlist - potential clinical candidates with favorable PK profiles, high tolerability, and manageable toxicity. The selected candidate must demonstrate sufficient efficacy in treating the target disease in humans. Despite these efforts, the success rate of the pre-clinical candidate to sail through Phase I, Phase II, and Phase III in clinical trials remains exceedingly low. Supported by powerful data-driven tools, artificial intelligence (AI) has transformed this traditional drug discovery process by enabling the analysis of large quantities of omics, phenotypic, and expression data to identify the biological mechanisms of pathological conditions and in turn identify druggable proteins or genes. The generative AI-powered toolbox creates novel compounds from scratch, assists scientists in optimizing druggability attributes, and bridges the differences between animal and human physiology and anatomy to predict potential toxicity in humans with high accuracy. This review discusses the bottlenecks in the traditional drug discovery approach, the impact of AI and machine learning (ML) in drug discovery, and potential challenges associated with AI/ML adoption.

With the rapid and accelerating advancement of generative artificial intelligence (AI), research is lagging on how to ensure that the health workforce becomes and stays AI-literate. This paper describes a way forward specifically toward establishing an AI-augmented curriculum within global health workforce education. By global health workforce education, we refer to the academic staff or faculty and students in medicine, nursing, global public health, and other health science fields. AI, unlike other technological advancements, is constantly changing. Therefore, the adoption of specific tools for health workforce education has to be analyzed in the context of the educational setting for shaping a sustainable and equitable AI-augmented global health workforce curriculum. This necessitates an integration of AI algorithmic literacy within academic curricula. In this paper, we propose the algorithmic literacy framework (ALF) for global health workforce education to address individual and organizational readiness. At the individual level, ALF examines one’s knowledge of and skills needed to implement AI within the context of their respective health education expertise. At the organizational level, ALF examines readiness across five areas: infrastructure and support systems, institutional support, Information and Communications Technology technical expertise, student engagement, faculty engagement, and analytics technical expertise. ALF offers universities and health workforce training institutions a way of organizing their approach to algorithmic/AI literacy readiness that embraces their organization’s values and, at the same time, urging them to act.

Accurate mortality prediction in intensive care units (ICUs) is essential for optimizing patient treatment, nursing care, and resource allocation. Traditional models, such as Acute Physiology and Chronic Health Evaluation and Simplified Acute Physiology Score, have been very important in clinical practice, but they frequently have issues with prediction accuracy and adaptability, especially when dealing with complex and evolving patient data. These issues can be resolved, and the accuracy of mortality prediction increased due to recent developments in machine learning, especially deep learning. The present study introduces a new deep learning ensemble model that achieves a significant improvement over existing methods. Using stacked ensemble learning, our approach combines the advantages of one Random Forests model and two CatBoost models. We achieved a notable performance in mortality prediction by carefully training and optimizing this ensemble using the electronic ICU Collaborative Research Database. Our model boasts an accuracy of 94.19%, precision of 94.097%, recall of 94.29%, and F1-score of 94.191%, demonstrating a substantial improvement over conventional approaches. The prediction of ICU mortality has been significantly improved using ensemble learning, which helps medical and nursing staff to better treat patients individually, allocate resources efficiently, and enhance patient outcomes. This approach gives healthcare experts the ability to make data-driven decisions, leading to more effective and efficient care within the ICU.

Automatic liver tumor segmentation is an essential part of computer-aided diagnosis systems. Despite the significant progress made by fully convolutional neural networks (FCNs) in recent years, existing methods fail to effectively segment multiscale tumors and accurately delineate tumor boundaries. This indicates that the single-connection network structure overly relies on the high-level semantic information of the image but fails to fully utilize the spatial information of the target across different levels. To solve these problems, we introduce a novel end-to-end segmentation network, named intra- and inter-block densely connected FCN (I2-DenseFCN). This network comprising intra-block and inter-block dense connections in the encoder and decoder, respectively, is developed to effectively handle the segmentation of multiscale tumors. We also designed a hybrid loss function to achieve more accurate tumor boundary delineation by adaptively adjusting the optimization of the cross-entropy and Lovász Softmax loss. In addition, we are the first to propose an intensity-adaptive image pre-processing method that effectively mitigates the differences between real-world medical scenarios and model inference for liver tumor segmentation tasks. Our training data, derived from the Liver Tumor Segmentation Challenge (LiTS) public dataset, have been expanded to 25,755 samples of liver and their tumors through data augmentation techniques. Experimental evaluations on the LiTS and 3DIRCADb databases demonstrated the superiority of the proposed I2-DenseFCN over classical methods. We believe this research will contribute to the field of tumor segmentation using AI-based medical image processing methods, which is a promising application in computer-aided diagnosis systems.

Most modern models designed for ultrasound (US) image segmentation are characterized by high computational and memory requirements, limiting their practical utility in point-of-care US settings. Consequently, researchers have devised innovative approaches to compress these large models, enabling the training of smaller networks capable of achieving comparable generalization performance. Among these strategies, knowledge distillation (KD) has emerged as particularly suitable for scenarios involving small datasets or where significant efficiency improvements are desired. While previous KD-based methods have focused on extracting comprehensive information from diverse levels of teacher representation, they often overlook the identification of the most effective representation level. Additionally, many existing techniques propose intricate strategies that present implementation challenges. To address this gap, our study concentrates on selecting optimal teacher representations from various levels. Through an exhaustive analysis of KD pathways, loss functions, and the impact of augmentation, we offer valuable insights into the mechanisms underlying knowledge transfer from the teacher to the student networks. Our proposed methodology significantly enhances student performance, elevating the Dice similarity score from 73% to 80%, while the teacher model achieves 81%. Notably, our student model achieves this improvement with only 0.82 million parameters, compared to the teacher model’s 96 million parameters.

In this study, we aim to examine the efficacy of deep learning methods in predicting the 1-year risk of developing cirrhosis in patients with chronic hepatitis C (CHC), as defined by transient elastography (TE), in comparison with conventional models, as well as to assess whether semi-supervised learning can improve performance relative to supervised learning when the labels are limited. We used the electronic health records of the 169,317 valid patients in the Veterans Health Administration system from 2000 to 2016. Predictor variables contained baseline characteristics, such as age, gender, race, hepatitis C virus genotype, and 26 liver-related longitudinal variables such as sustained virologic response and laboratory data. The response variable, developing cirrhosis, is defined as liver stiffness >12.5 kPa on TE within a 1-year window. Using baseline and longitudinal variables, we fitted four prediction models, including logistic regression (LR), random forest (RF), supervised recurrent neural network (RNN), and semi-supervised RNN (semi-RNN) and evaluated their performances. Both RNN (area under the receiver operating characteristic curve [AuROC] 0.744) and semi-RNN (AuROC 0.785) accurately predicted the risk of cirrhosis within 1 year and significantly outperformed RF (AuROC 0.731) and LR (AuROC 0.724). By enabling early identification of high-risk patients, these models hold promise for targeted interventions in clinical CHC treatment.

Breast cancer is one of the most prevalent cancers affecting women globally. Early diagnosis is crucial for effective treatment and improved survival rates. Imaging techniques such as mammography and ultrasound are widely used conventional diagnostic methods. However, these methods have limitations, including low sensitivity and specificity, especially in patients with dense breast tissue. For instance, mammograms miss approximately 20% of breast cancer cases, leading to false negatives and delayed treatment that can have fatal consequences. To address these challenges, artificial intelligence (AI)-based diagnostic tools have been developed to assist healthcare professionals in accurately detecting breast cancer. These tools work in conjunction with human radiologists to improve diagnostic outcomes. In addition, biomarkers present a promising non-invasive, more convenient alternative for the early detection of breast cancer, potentially overcoming the limitations of traditional screening methods. Various biomarkers, such as circulating tumor cells, cell-free tumor nucleic acids, and microRNAs, have shown promise in early breast cancer diagnosis. A systematic literature review is needed to consolidate ongoing efforts in molecular biology and biomedical sciences aimed at achieving early breast cancer diagnosis. One of the limitations of previously published research is the heterogeneity of methodologies, which can compromise the credibility of comparisons due to potential inaccuracies in the original data. Hence, future studies should prioritize using consistent datasets and developing robust techniques to manage missing values, outliers, and class imbalances to improve the reliability of breast cancer detection models. This literature review seeks to bridge the knowledge gap by reporting recent high-performing AI models and effective biomarkers that can serve as diagnostic tools in clinical practice.

Improvements in the artificial intelligence (AI) health system have been effective in reducing the risks associated with transgender medical tourism and travel. The ability to track medical travel from the place of origin to the final treatment destination is dependent on the development of AI. This project aims to improve the AI health system to promote travel and medical tourism, utilizing quantitative research methodologies, including survey-based research and partial least squares structural equation modeling. The participants included 381 people from medical professionals, tourism experts, transgenders, and technology enthusiasts interested in AI and health. The findings show that the AI health system has significantly improved medical travel and tourism; key factors such as medical tourism, AI systems, medical travel and risk factors, attitude, behavioral intention, and medical destination image have all contributed to better healthcare experiences for transgender individuals. Specialized care should be provided to transgender individuals traveling for surgeries and medical treatments, emphasizing their unique needs. Subsequent investigations might focus on the broader function of AI, particularly in terms of ensuring the dignity and respect of the tourism site. This study proposes further integration of AI into healthcare systems to maximize the benefits of safe medical travel and secure tourist locations.