Transforming Medical Education: The Integration of Artificial Intelligence
Introduction
Artificial intelligence (AI) is rapidly changing various sectors, and medical education is no exception. This article explores the current applications of AI in medical education, its potential to revolutionize learning and assessment, and the challenges that need to be addressed for its successful integration.
The Rise of AI in Medical Education
Since 2018, there has been a notable increase in publications related to AI in medical education. These studies primarily originate from Anglo-North American countries and focus on undergraduate medical education, with a smaller percentage addressing graduate medical education and continuing professional development. The applications of AI span across various basic science disciplines and clinical specialties.
AI's Impact on Medical Education
Curriculum Development and Analysis
AI can minimize the time it takes to review various curricula, solve multidimensional problems, improve classification accuracy, and indicate a relationship between the parameters in curriculum assessment. For example, AI can assess the "effectiveness of curriculum" and the "overall happiness" of medical students with the course curriculum, which is critical in preparing future doctors.
Personalized Learning
AI can assist students in receiving adaptive and tailored instructional content enhanced by student feedback, allowing students to detect their knowledge gaps and respond appropriately. AI offers interactive learning experiences and aids in literature review and paper drafting. By analyzing vast amounts of data and utilizing machine learning algorithms, AI can deliver adaptive instructional content that is suitable for each student's knowledge gaps and learning speed.
Assessment
AI can be used to assess students’ clinical skills. One way is for students to interact with AI-generated standardized patients on a computer, with the AI tool responding to student questions and comments. The tool evaluates the visit according to standards on which it has been trained, such as the widely adopted Objective Structured Clinical Examination assessment tool. AI can also analyze student quiz results, so professors can see “students who are consistently getting questions wrong in a certain area,” such as dermatology or diabetes care.
Read also: Machine Learning Curriculum
Research
AI aids in literature reviews and paper outlines but must adhere to scientific standards. In research, AI aids in literature reviews and paper outlines but must adhere to scientific standards.
Specific AI Applications in Medical Education
Virtual Patients
Several schools are starting to use AI to assess students’ clinical skills. One way is for students to interact with AI-generated standardized patients on a computer, with the AI tool responding to student questions and comments. The tool evaluates the visit according to standards on which it has been trained, such as the widely adopted Objective Structured Clinical Examination assessment tool. For example, UT Health San Antonio fed an AI program scores of simulated cases covering a wide range of conditions, symptoms, histories, and demographics, Rodriquez says. Plus, staff provided this important instruction: “Be honest with your responses. Don’t offer any more than what they ask, and only answer if you know the correct answer. At UMN Medical School, students also conduct clinical visits on a computer with AI-generated standardized patients, Violato says. What are faculty learning from these interactions? One example: “Students who get them right tend to ask questions throughout the entire encounter, whereas students who get them incorrect tend to ask questions at the end.
AI-Powered Tutoring Systems
Teachers can use a virtual inquiry system like “DxR Clinician” as a beneficial analysis tool to assist them in understanding their students' behaviour and altering their courses based on the assessment results. Students can gain skills to solve clinical problems quickly. They can learn a lot about critical disease diagnosis by interacting with the examples. Simultaneously, the system can detect errors made by students during the case study, do deep learning and analysis, and assist students in resolving these issues. “Intelligent Tutor Systems”, similar to DxR Clinician, can also follow the learner's "psychological processes" in solving problems to diagnose incorrect notions. It also evaluates the learners' level of understanding.
Generating Study Questions
UC College of Medicine created an AI tool to analyze content from a specific course, then generate USMLE-style study questions and answers (along with explanations for the correct answers) based on that content. The school conducted a pilot test last academic year in a course about the blood system. The course director determined that 85% of the questions and answers met USMLE criteria.
AI in Recommendation Letters
Generative AI (genAI) offers faculty in academic medicine a practical way to streamline the time-intensive task of writing letters of recommendation (LORs). This article introduces an ethical, structured framework for GenAI to draft personalized, bias-aware LORs. It explores selecting appropriate genAI models, creating effective prompts, and safeguarding privacy while maintaining transparency. Faculty can enhance LOR quality by moving away from repetitive shortcuts like copying from CVs.
Read also: Artificial Intelligence Education
Clinical Skills Assessment
Several schools are starting to use AI to assess students’ clinical skills. One way is for students to interact with AI-generated standardized patients on a computer, with the AI tool responding to student questions and comments. The tool evaluates the visit according to standards on which it has been trained, such as the widely adopted Objective Structured Clinical Examination assessment tool.
Challenges and Considerations
Ethical Concerns
As AI spreads, worries about algorithmic bias, security, and data privacy surface. It is crucial to build strong ethical frameworks to guarantee that AI algorithms are transparent, just, and unbiased in their decision-making processes. The AAMC framework for integrating AI into medical education emphasizes maintaining a human-centered approach, ensuring ethical use, providing equitable access, fostering continuous education, and developing curricula through interdisciplinary collaboration. It highlights protecting data privacy and constant evaluation of AI tools.
The Need for AI Literacy
The paper explores the transformative role of AI in medical education, stressing the need for AI literacy and core competencies among healthcare professionals. It addresses the integration of AI in four key areas: selection into medical programs, learning, assessment, and research.
Infrastructure and Technical Support
For AI to be used in medical education, many infrastructures and technological support are required. Educational institutions and healthcare organizations must make investments in cutting-edge computing hardware, data storage, and secure networks if they want to fully benefit from AI.
Over-Reliance and Human Oversight
AI has limitations, including outdated information and the risk of over-reliance, highlighting the necessity of human oversight. "The human is the final arbiter of the ‘ground truth’” in assessing AI-generated work with and for students. “The human part of the equation is such an important component.
Read also: Benefits and Challenges of AI in Higher Ed
Bias in AI Training Data
The problem of bias in AI training data is also well documented. A recent study from Brigham and Women’s shows that including more detail in AI-training datasets can reduce observed disparities, and ongoing research by a Mass General pediatrician is training AI to recognize bias in faculty evaluations of students.
Accuracy of AI Content Detection Tools
The article emphasizes the inconsistencies in AI content detection tools, suggesting a more holistic approach involving manual review for academic integrity. It advocates leveraging AI to enhance learning and assessment, including personalized feedback and complex tasks. Medical educators should integrate AI-enhanced methods, regularly evaluate AI tools, and foster academic honesty.
Integrating AI into the Curriculum
The article advocates for integrating AI education into medical school curricula, emphasizing enhancing medical competency. Key findings include identifying 82 essential AI competencies, suggesting methods for integration, and underscoring the importance of AI literacy and application skills. Recommendations include incorporating AI into biostatistics, case-based learning, clinical rotations, and involving guest lecturers. The article emphasizes defining AI competencies essential for medical graduates, advocating curriculum integration to address evolving healthcare technologies. Key findings suggest incorporating AI principles, ethical considerations, and effective health service applications into medical education. It encourages curriculum developers to create AI-focused content, revise existing courses, and involve in further research to ensure comprehensive training.
Key Competencies
The study emphasizes defining AI competencies essential for medical graduates, advocating curriculum integration to address evolving healthcare technologies. Key findings suggest incorporating AI principles, ethical considerations, and effective health service applications into medical education. It encourages curriculum developers to create AI-focused content, revise existing courses, and involve in further research to ensure comprehensive training.
Methods for Integration
Suggestions include incorporating discretionary learning outcomes through continuing education, vertically integrating content into existing curricula, and gradually introducing material. Utilizing an integrated curriculum approach is recommended.
The Student Perspective
Healthcare students hold a positive attitude towards AI in medicine; however, their knowledge and skills in this area are lacking due to insufficient exposure and inadequate AI-specific training in the curriculum. The review suggests incorporating face-to-face instruction, detailed training manuals, and interdisciplinary courses in digital health to bridge the educational gap.
Institutional Initiatives
Harvard Medical School (HMS)
HMS is getting a jump on this shift by building generative AI (also called genAI) into the curriculum today. Among the changes incorporated this fall is a one-month introductory course on AI in health care for all incoming students on the Health Sciences and Technology (HST) track. A PhD track that starts this semester, AI in Medicine (AIM), is taking AI-integrated education even further. HMS has also opened a third avenue for medical students and faculty who are interested in the technology: the Dean’s Innovation Awards for the Use of Artificial Intelligence in Education, Research, and Administration, which were announced last year and offer grants of up to $100,000 for each project selected
American Medical Association (AMA)
At its annual conference in 2018, the American Medical Association (AMA) adopted its first policy on augmented intelligence. It supported studies that highlighted how AI should be undertaken in medical education.
Duke Institute for Health Innovation
Medical students at Duke Institute for Health Innovation work with data experts to build care-enhanced technologies for doctors.
Stanford University Centre for AI in Medicine and Imaging
Similarly, Stanford University Centre for AI in Medicine and Imaging engages graduate and postgraduate students in using machine learning to solve healthcare problems.
University of Florida
The residents of the radiology department of the University of Florida collaborated with a technology firm to develop computer-aided detection for mammography.
Carle Illinois College of Medicine
Carle Illinois College of Medicine offers a course taught by a scientist, clinical scientist, and engineer to learn about new technologies. In addition, the Sharon Lund Medical Intelligence and Innovation Institute offers a course on the latest healthcare technology, which is open to medical students.
UC College of Medicine
UC College of Medicine created an AI tool to analyze content from a specific course, then generate USMLE-style study questions and answers (along with explanations for the correct answers) based on that content. The school conducted a pilot test last academic year in a course about the blood system. The course director determined that 85% of the questions and answers met USMLE criteria.
The Future of Medical Education with AI
AI provides a promising future for medical education with personalized learning experiences, enhanced diagnostic capabilities, and easier curriculum creation/implementation. It is critical to solve the problems of ethics, training, and infrastructure in order to fully realize the potential of AI in medical education.
tags: #artificial #intelligence #in #medical #education #applications
