Innovations in MedEd

Integrating Pedagogy, Reflection, and AI with Self-Directed Learning

By Loren Deutsch

The transition to medical school presents numerous firsts, including preparation for the first exam. During the transition from undergraduate studies to medical school, students question their study strategies as they navigate new protocols for learning detail-rich content while cultivating clinical decision-making skills and an effective approach for test preparation. It is not uncommon to rely on outdated strategies, and for some, it is the first time they have had to establish a protocol for learning. As M1s prepare for their first exams, it is a good time to consider how to cultivate a self-directed approach to medical learning and whether to upgrade personal study skills to learn the course content within a study workflow that optimizes medical knowledge, clinical reasoning, physician identity, and confidence.

As educators, we have first-hand knowledge about lifelong learning and recognize that medical education and the allied health sciences are academically organized around defined learning objectives that include course curricula, clinical skills development, and associated licensing and board certifying exams. Though this article references strategies about the United States Medical Licensing Examination (USMLE) for Step 1, in our experience at LAS, these strategies are equally germane to Step 2 and Step 3, as well as the National Board of Osteopathic Medical Examiners COMLEX Exams, the National Council of State Boards of Nursing (NCSBN) Exams, the Physician Assistant National Certifying Exam, and those in GME among others.

Consider this a sampling of fresh insights into the medical education landscape with actionable coaching strategies and evidence-based approaches to resources that you can use to support yourself or your students as they build content and process knowledge for lifelong learning. The examples are intended to address the executive functioning skills, test stressors, burnout, and other challenges that pertain to self-directed learning and test preparation in medical education.

Optimize learning 

In 2024, the USMLE updated the content categories for Step 1 and recategorized the foundational science content. Revisions included Human Development, a new category for normal age-related findings, and care of the well patient. Updates to Physician Tasks were also made, with revisions to four competencies, including medical knowledge, patient care, communication and interpersonal skills, and practice-based learning and improvement (USMLE.org). Below are examples of updated organ systems and their distribution on Step 1. This information provides students with an organizing function for their Step 1 preparation as well as a framework for M1 and M2 learning goals.

Additional Step 1 updates include increases in the distribution of questions from all categories by organ system. Specifics are available on the USMLE.org website and offer a framework for students prioritizing their study content and workflow from highest to lowest yield topics. The content outline at USMLE.org provides a useful foundation for students looking to organize and prioritize their approach to learning and test preparation, including their first licensing exam. However, the outline is not designed as a study guide in itself. It is important to review your medical school’s curriculum along with the content specifications for Step 1 before designing daily or weekly learning goals.

ChatGPT for case data, local knowledge base, and practice questions

ChatGPT is increasingly popular among medical students, in part because they can use the large language model to quickly retrieve information and personalize learning experiences. It is also helpful for prompt engineering and text vectorization

In the Journal of Medical Education and Curricular Development (Zhang et. al. 2024), 740 students were surveyed, and 131 responded. Among the respondents, students described an increasingly widespread interest in ChatGPT among medical students in the US. The article explains that among the students who already use ChatGPT, 43.7% use it weekly, several times per week, or daily, and 45.3% prefer ChatGPT over other medical resources, including a professor or attending.

Given the speed at which ChatGPT can retrieve and generate information, it is no surprise that students, faculty, and independent contractors are working as quickly as possible in text mining to develop or generate case data for a local knowledge base in medical education.

It is also no surprise that ChatGPT has significant potential in developing practice questions for high-stakes medical exams. According to an article in the Journal of Medical Internet Research (Gilson et. al. 2023), ChatGPT was able to achieve the equivalent of a passing score for a third-year medical student when assessed using the (free) NBME Step 1 assessment. Not too long ago, I remember how difficult it was trying to find (or write) additional practice questions for students preparing for Step 1. Apart from the inherent costs involved in purchasing digital Q-Bank subscriptions, it also took weeks, months, and even years to create novel practice questions or assessments when students needed more practice.

Define Goals that Support Clinical Confidence

Custom Study Plans That Work – No more cramming! LAS Coaches guide students in implementing a daily workflow to study and learn with measurable goals. LAS designs individualized plans based on personal needs, classroom content, self-reflective strategies, and active recall with spaced repetition. Using structured practice blocks and self-assessments to deepen knowledge, enhance process skills, and track progress, LAS designs individualized plans that align with personal needs, goals, and learning preferences, supplementing with AI-generated content. LAS Coaches collaborate and advise students and faculty on integrative approaches for utilizing classroom content and AI content such as natural language processing (NLP), rule-based systems, Test Prep Bots, ChatLabs, Chatbots, and other AI-generated content to supplement (but not replace) traditional learning strategies.

Exam Coaching That Builds Knowledge – Timed Q-banks provide data-driven progress tracking. When combined with self-reflective goals and strategies, students learn how to learn in sustainable and measurable ways, with insights about confidence and test readiness. Combining goals for content and process promotes test readiness, high achievement outcomes, and clinical confidence – all while deepening medical knowledge.

Coaching and Consulting Beyond Content – LAS Coaches provide structure, support, and accountability to all students and faculty and data shows that the LAS Coaching Model is highly effective, particularly when transitioning to medical school, residency, or fellowship, navigating challenges associated with executive functioning, anxiety, depression, burnout, and ADHD, accommodations requests/appeals, and core competencies that include interpersonal communication and communication breakdowns with faculty and patients, professionalism, clinical skills development, and patient safety.

Self-reflective practices – With innovations to medical education and an ever-changing AI landscape, students must examine their protocols for learning and achievement, and faculty and independent contractors must be willing to do the same. Integrating AI with traditional resources offers a realistic starting point while recognizing the importance of human connection and relationship-based models for learning and mentorship. There are limitations to using AI for learning and studying in medical education. Though students may readily gravitate toward AI resources for study tools because they are time-efficient and effective, it is vitally important that they also learn to cultivate strong interpersonal communication skills, clinical decision-making, and a professional identity that includes learning how to learn in self-directed and autonomous ways. Especially because of a quickly changing landscape in medical education and healthcare, students need to learn and practice within a relationship-based model so that they can optimize their medical knowledge and lifelong learning skills.

Hyperlinks

Loren Academic Services (LAS) 

The American Academy of Physician Associates 

The National Board of Osteopathic Medical Examiners

The National Council of State Boards of Nursing

The United States Medical Licensing Examination