Artificial intelligence models represented in machine learning algorithms are promising tools for risk assessment used to guide clinical and other health care decisions. Machine learning algorithms, however, may house biases that propagate stereotypes, inequities, and discrimination that contribute to socioeconomic health care disparities. The biases include those related to some sociodemographic characteristics such as race, ethnicity, gender, age, insurance, and socioeconomic status from the use of erroneous electronic health record data. Additionally, there is concern that training data and algorithmic biases in large language models pose potential drawbacks. These biases affect the lives and livelihoods of a significant percentage of the population in the United States and globally. The social and economic consequences of the associated backlash cannot be underestimated. Here, we outline some of the sociodemographic, training data, and algorithmic biases that undermine sound health care risk assessment and medical decision-making that should be addressed in the health care system. We present a perspective and overview of these biases by gender, race, ethnicity, age, historically marginalized communities, algorithmic bias, biased evaluations, implicit bias, selection/sampling bias, socioeconomic status biases, biased data distributions, cultural biases and insurance status bias, conformation bias, information bias and anchoring biases and make recommendations to improve large language model training data, including de-biasing techniques such as counterfactual role-reversed sentences during knowledge distillation, fine-tuning, prefix attachment at training time, the use of toxicity classifiers, retrieval augmented generation and algorithmic modification to mitigate the biases moving forward.

The novel coronavirus disease 2019 (COVID-19) pandemic has changed the medical education platform for students in the United States of America (USA). In that light, medical schools had to rapidly rearrange the dynamics of their educational curricula from the traditional platforms, to incorporate telemedicine. The telemedicine platform is supported in many specialties, allowing students various options to continue their education without interruption during the COVID-19 pandemic, and beyond. Telemedicine platforms are projected to grow exponentially due to the COVID-19 pandemic, allowing a segue for medical schools to modify their curricula by incorporating telemedicine programs. These distant-, e-learning (tele-education) programs align with the recommendations and guidelines for practicing social distancing. In this article, we surveyed fourth-year medical students to better understand their views on multiple aspects of e-learning, and its impact on their medical education during the COVID-19 pandemic. We assessed the medical students’ experiences, satisfaction, insight and knowledge with e-learning, tele-education, telehealth, and their related modalities during COVID-19. We provide an organized overview and analysis of the main factors that influence medical education during the COVID-19 pandemic, while bringing forth the main challenges, limitations, and emerging approaches in the field of telemedicine and its application as it relates to medical education and e-learning across medical specialties. We outline the main themes and ideas that the medical students voiced, as to how their medical education is being impacted by the COVID-19 pandemic and how they will incorporate telemedicine and tele-education in their future career. A cross-sectional, mixed-method survey was developed and distributed via Google Surveys to 181 University at Buffalo, Jacobs School of Medicine and Biomedical Sciences, United States of America, 4th year medical students, in December 2020. Results were compiled and analyzed after a 6-day open period for responses to be submitted. The survey instrument consisted of questions that inquire about the students’ perspectives as it relates to their rapid switch from their traditional method of learning to the on-line version of medical education during the COVID-19 pandemic. A total of 65 students responded to the survey, of which 63 completed the survey. More than half of the students (n = 63, 57%) indicated that both their specialty of interest, and (n = 21, 33%) their sub-internships were impacted by the temporary lockdown, due to the COVID-19 pandemic. Students also indicated that the top three specialties that were affected included surgery, internal medicine and obstetrics and gynecology. When the students were asked if they were satisfied with the use of aquifer for their health care e-learning, only 35% of the students were satisfied. The students expressed that the school’s administration team did a good job in developing the new tele-education curriculum for those in clinical training. In addition, responses indicated that students were open to case-based video learning and readings, when combined with the abbreviated clinical exposure during the make-up “clinical immersions periods” allowed for adequate learning. Overall, the survey responses show that more than half, approximately 54% of the medical students utilized telemedicine platforms during their clerkships that were impacted by COVID-19. The 4th-year medical students did not find tele-education and e-learning to be as effective as traditional medical education that combines in-person didactic classroom instructions and in-person face-to-face in hospital clerkships. Students felt that the telemedicine program that was rapidly set up due to the COVID-19 ‘lockdown’ was fragmented, since it was not a formal integration of a telemedicine E-learning program. Students would have preferred more ‘real’ cases to follow, instead of the ready-made, aquifer type of cases. Telemedicine has significant potential to address many of the challenges facing the medical education environment today. We believe now that people have become comfortable with this method of teaching, that even after the pandemic ends, we will continue to see tele-education used as a platform for medical education.

Ebola virus disease (EVD) is extremely virulent with an estimated mortality rate of up to 90%. However, the state-of-the-art treatment for EVD is limited to quarantine and supportive care. The 2014 Ebola epidemic in West Africa, the largest in history, is believed to have caused more than 11,000 fatalities. The countries worst affected are also among the poorest in the world. Given the complexities, time, and resources required for a novel drug development, finding efficient drug discovery pathways is going to be crucial in the fight against future outbreaks. We have developed a Computational Analysis of Novel Drug Opportunities (CANDO) platform based on the hypothesis that drugs function by interacting with multiple protein targets to create a molecular interaction signature that can be exploited for rapid therapeutic repurposing and discovery. We used the CANDO platform to identify and rank FDA-approved drug candidates that bind and inhibit all proteins encoded by the genomes of five different Ebola virus strains. Top ranking drug candidates for EVD treatment generated by CANDO were compared to in vitro screening studies against Ebola virus-like particles (VLPs) by Kouznetsova et al. and genetically engineered Ebola virus and cell viability studies by Johansen et al. to identify drug overlaps between the in virtuale and in vitro studies as putative treatments for future EVD outbreaks. Our results indicate that integrating computational docking predictions on a proteomic scale with results from in vitro screening studies may be used to select and prioritize compounds for further in vivo and clinical testing. This approach will significantly reduce the lead time, risk, cost, and resources required to determine efficacious therapies against future EVD outbreaks.

Background: Nonvalvular atrial fibrillation (NVAF) affects almost 6 million Americans and is a major contributor to stroke but is significantly undiagnosed and undertreated despite explicit guidelines for oral anticoagulation. Objective: The aim of this study is to investigate whether the use of semisupervised natural language processing (NLP) of electronic health record’s (EHR) free-text information combined with structured EHR data improves NVAF discovery and treatment and perhaps offers a method to prevent thousands of deaths and save billions of dollars. Methods: We abstracted 96,681 participants from the University of Buffalo faculty practice’s EHR. NLP was used to index the notes and compare the ability to identify NVAF, congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke or transient ischemic attack, vascular disease, age 65 to 74 years, sex category (CHA2DS2-VASc), and Hypertension, Abnormal liver/renal function, Stroke history, Bleeding history or predisposition, Labile INR, Elderly, Drug/alcohol usage (HAS-BLED) scores using unstructured data (International Classification of Diseases codes) versus structured and unstructured data from clinical notes. In addition, we analyzed data from 63,296,120 participants in the Optum and Truven databases to determine the NVAF frequency, rates of CHA2DS2‑VASc ≥2, and no contraindications to oral anticoagulants, rates of stroke and death in the untreated population, and first year’s costs after stroke. Results: The structured-plus-unstructured method would have identified 3,976,056 additional true NVAF cases (P<.001) and improved sensitivity for CHA2DS2-VASc and HAS-BLED scores compared with the structured data alone (P=.002 and P<.001, respectively), causing a 32.1% improvement. For the United States, this method would prevent an estimated 176,537 strokes, save 10,575 lives, and save >US $13.5 billion. Conclusions: Artificial intelligence–informed bio-surveillance combining NLP of free-text information with structured EHR data improves data completeness, prevents thousands of strokes, and saves lives and funds. This method is applicable to many disorders with profound public health consequences.

Importance: Large language models (LLMs) are being implemented in health care. Enhanced accuracy and methods to maintain accuracy over time are needed to maximize LLM benefits.Objective: To evaluate whether LLM performance on the US Medical Licensing Examination (USMLE) can be improved by including formally represented semantic clinical knowledge.Design, Setting, and Participants: This comparative effectiveness research study was conducted between June 2024 and February 2025 at the Department of Biomedical Informatics, Jacobs School of Medicine and Biomedical Sciences, University at Buffalo, Buffalo, New York, using sample questions from the USMLE Steps 1, 2, and 3.Intervention: Semantic clinical artificial intelligence (SCAI) was developed to insert formally represented semantic clinical knowledge into LLMs using retrieval augmented generation (RAG).Main Outcomes and Measures: The SCAI method was evaluated by comparing the performance of 3 Llama LLMs (13B, 70B, and 405B; Meta) with and without SCAI RAG on text-based questions from the USMLE Steps 1, 2, and 3. LLM accuracy for answering questions was determined by comparing the LLM output with the USMLE answer key.Results: The LLMs were tested on 87 questions in the USMLE Step 1, 103 in Step 2, and 123 in Step 3. The 13B LLM enhanced by SCAI RAG was associated with significantly improved performance on Steps 1 and 3 but only met the 60% passing threshold on Step 3 (74 questions correct [60.2%]). The 70B and 405B LLMs passed all the USMLE steps with and without SCAI RAG. The SCAI RAG 70B model scored 80 questions (92.0%) correctly on Step 1, 82 (79.6%) on Step 2, and 112 (91.1%) on Step 3. The SCAI RAG 405B model scored 79 (90.8%) correctly on Step 1, 87 (84.5%) on Step 2, and 117 (95.1%) on Step 3. Significant improvements associated with SCAI RAG were found for the 13B model on Steps 1 and 3, the 70B model on Step 2, and the 405B parameter model on Step 3. The 70B model was significantly better than the 13B model, and the 405B model was not significantly better than the 70B model.Conclusions and Relevance: In this comparative effectiveness research study, SCAI RAG was associated with significantly improved scores on the USMLE Steps 1, 2, and 3. The 13B model passed Step 3 with RAG, and the 70B and 405B models passed and scored well on Steps 1, 2, and 3 with or without augmentation. New forms of reasoning by LLMs, like semantic reasoning, have potential to improve the accuracy of LLM performance on important medical questions. Improving LLM performance in health care with targeted, up-to-date clinical knowledge is an important step in LLM implementation and acceptance.