How an adaptable AI-enabled VR Sim Lab and Bloom’s Framework can become every healthcare program’s favorite training tool
Author: Elbert Waller, Clinical Education Specialist, Lumeto
Writing from my background in simulation program management, I am excited to see how the recent tech wave of AI and XR can be leveraged for healthcare training. So, with the opportunity to play around in this tech pool, my first thought was to see how AI can become an effective aide to healthcare educators and how learners can have truly immersive experiential learning at their own pace, tailored to their competency level.
In this pursuit, I leveraged Lumeto’s InvolveXR platform to create immersive Self-Directed Learning Experiences (SDLE), where an AI-driven checklist, built on Bloom’s taxonomy, guides and assesses learners. This journey builds on my experiences in leading programme development and simulation education, where I have witnessed first-hand the transformative impact of tailored educational approaches. Utilising Bloom’s Taxonomy, these SDLE are educational modules that are not only deeply customisable but also easily integrated into existing curricula at all learning levels. Through this blog, we will delve into how these cutting-edge technologies can be leveraged to enhance understanding, improve skills retention, and ultimately elevate patient care.
GenAI, GPT, and LLM – What’s all the fuss about?
Before I go further, here’s a quick intro into all the newly popular buzzwords, for those interested. Generative AI (GenAI) is a form of Machine Learning, comprising a class of algorithms that can generate text, images, video, etc. as a response to “prompts”. Large Language Model (LLM) is a type of GenAI that’s based on a language model developed with a large amount of textual data, producing human-like textual responses. Hence they’re used widely with natural language processing applications. Generative Pre-trained Transformers (GPT) is a form of LLM, introduced by OpenAI. The latest version, GPT4, has improved alignment, which is its ability to follow user intentions while also making it more truthful and generating less offensive or dangerous output. GPT4 also has significantly reduced latency.
The role of Large Language Model (LLM)-based AI in enhancing Learning Experiences
LLM-based AI in an immersive learning platform brings several interesting capabilities to enhance the entire learning experience and the range of learning experiences. I’ve seen this firsthand on the InvolveXR platform, where it manifests in several ways.
LLM-driven automated checklists as an educator’s aide and learner’s guide
In a self-directed experience, it’s crucial for an effective proxy to the educator, to guide the learners, ensuring that the learning is seamless and the learning objectives are met. For this, LLM can be used to design and automate a guided checklist with clear instructions and explanations. The AI actively listens to the learner and checks off the instructions, giving the learners a sense of task completion and learning. While this is great for an educational mode, this checklist can also be hidden from the learner in the subsequent assessment modes. This way, the checklist can also serve as a tool for formative and summative assessments.
I have been struck by how working with AI patients on Lumeto’s InvolveXR platform evokes an emotional response that mirrors interactions with human patients. The joy of making a breakthrough in patient communication or the sadness felt during a patient’s distress highlights the system’s realism and immersiveness. This emotional engagement makes each simulation a powerful tool for honing empathetic and professional skills as a nurse.
Kimberly Workum RN, BScN, MEd, CHSE, CCNE
Director, Clinical Competence Assessment Centre and Digital Strategies,
University of Manitoba
Conversational AI for realistic patient conversations
Using simple text prompts, simulation educators can easily create a multitude of patients with varied backgrounds – clinical and psychosocial. And this translates to learners in VR going through a realistic patient encounter, practicing, and developing several communication-based skills. With the integration of GPT4, the latency in response is almost zero, delivering a true suspension of belief and realism with fluent conversations. In the video below, you can see a segment of a Learning Experience I developed on the InvolveXR that focuses on the recognition and management of elder abuse. You can observe the realistic patient responses and also note the check-off of the checklist when the learner completes their questionnaire and makes the detection of elder abuse.
Easy customization and creation of Learning Experiences
Using simple text, simulation educators can create and customize a range of learning experiences to meet their training objectives. This could be writing LLM prompts to create a multitude of patients with varied backgrounds – clinical and psychosocial. And they can also create LLM rules for checklist automation. Simply type away. It takes me about 20 minutes to create an SDLE from scratch on the InvolveXR platform, and I’ve just begun doing this three weeks ago!
Localization and language support
With an LLM such as GPT4, quite a lot of different languages and dialects are already supported. This can help in simulating patient encounters in different languages, or with different cultural and geographic references. As a technology that is improving at an exponential rate, this capability is only set to improve and grow.
Adapting an SDLE to suit different learners with Bloom’s Taxonomy
Bloom’s Taxonomy serves as a framework to categorize educational goals and objectives across cognitive levels, from basic knowledge recall to complex analytical and creative tasks. Often used in simulation education, this became my choice of framework to guide the structuring of the SDLE on InvolevXR to progressively challenge and engage students. For example:
- Knowledge and Comprehension: Introductory modules can engage students with foundational concepts and procedures while getting them familiar with the modality.
- Application and Analysis: More advanced modules can challenge healthcare professionals to apply their knowledge to solve complex medical problems within the VR environment, with AI providing real-time analysis and feedback on their decisions.
- Synthesis and Evaluation: In higher-level modules, learners might be tasked with creating sophisticated treatment strategies for managing intricate emergency situations in VR, encouraging them to synthesize information and evaluate outcomes critically.
Implementation in Curricula
Incorporating AI-enhanced SDLEs into existing curricula can be achieved with minimal disruption. Educators can introduce modular components targeting specific learning outcomes and expand their use to broader curricular integrations over time. This approach allows educational institutions to adopt innovative technologies gradually and scale effectively.
VR Simulation Training in Nursing Education
Reflecting on the potential of VR and AI in nursing education, imagine a scenario where nursing students can engage in VR simulations designed to mimic complex clinical environments. These AI-enhanced VR simulations can span a nursing curriculum, ranging from patient interaction to becoming familiar with the ten rights of medication administration. The AI enhanced conversation and feedback on the learning objectives will provide real-time feedback and create adaptive learning paths that mirror real patient care situations. The application of this knowledge in difficult healthcare simulation scenarios might come after these sessions in a safe VR environment. Imagine a high-pressure scenario in an emergency room where nursing students must prioritize patient care, administer medications correctly, and communicate effectively under stress. The simulation would evaluate their performance based on various outcomes and provide tailored feedback to help them improve their decision-making and technical skills. This approach not only reinforces theoretical knowledge but also hones practical skills in a safe, controlled environment, preparing students for real-life challenges they will face in healthcare settings.
The integration of AI-driven conversations and VR in healthcare education using Bloom’s framework presents a progressive approach to training healthcare professionals. It not only personalizes the learning but also makes it more engaging and effective. As educators and institutions look to the future, adopting these technologies will be crucial to maintaining the cutting edge of medical training.
If you would like to get a demo of how this would work in a cutting edge VR environment, use the Calendy link below to book a chat with me and our team at Lumeto.
Further Reading
To delve deeper into this subject, the following publications are recommended:
“Virtual Reality in Medical Education: A Systematic Review” – explores the effectiveness of VR as a teaching tool in medicine.
“The Application of Bloom’s Taxonomy in Higher Education” – discusses how Bloom’s educational framework can be utilized to enhance learning outcomes.
“Innovations in Simulation: Use of AI in Training Healthcare Professionals” – examines the role of AI in medical simulation training.