Presentation: AI as a tool for
Enhancing Clinical Education in
Speech Language Pathology.
Proposal: Developing a Supervised Training Research AI Database for Speech-Language Pathology Graduate StudentsImagine this: You are a new SLP graduate student, eager to dive into your first clinical rotation. You are handed a stack of patient files, each one a treasure trove of medical history, assessment results, and therapy notes. Hours slip away as you analyze these documents, trying to piece together a clear picture of your client’s needs. Next comes the daunting task of crafting an effective treatment plan. You need to consult countless research articles to find evidence-based approaches with positive patient outcomes. But with so much information at your fingertips, it is hard to know where to start. You would probably do what most of our students do: turn to AI. However, AI struggles to answer very specific questions and give the most relevant information for this very specific case.
Now, imagine having a super-smart AI assistant named Vinny. Vinny isn't just any AI; it's been specifically trained on speech-language pathology using Retrieval-Augmented Generation (RAG) and a supervised training model for a curated dataset. RAG is like giving Vinny a powerful search engine that can instantly access a huge library of research papers and clinical guidelines. When you ask Vinny a question about speech-language pathology, it first uses the search engine to find the most relevant information from the library. Supervised training teaches Vinny how to use this new tool effectively. We show Vinny many examples of questions and the best answers, and we give it feedback on its performance. Over time, Vinny gets better and better at using the search engine to find the right information and provide the most helpful responses. Then, Vinny uses its own intelligence to understand and process this information, and finally, it gives students a clear and helpful answer. This makes Vinny much more accurate and informative, especially when it comes to complex topics like speech-language pathology. It can provide evidence-based answers, suggest treatment plans, and even help you understand difficult research papers.
By combining RAG and supervised training, we can create an AI system that is a valuable tool for SLP graduate students. This system can help students learn more effectively, make better decisions, and ultimately provide better care to their patients. Students can ask Vinny anything, from specific treatment techniques for a particular disorder to the latest research on a complex clinical case. Vinny can instantly access a vast library of information, process it, and provide you with clear, concise answers giving students the information needed to discuss this case with their clinical supervisor, citing the evidence on the different treatment approaches they’ve learned about, and why they’ve selected a specific method for the patient.
But Vinny doesn’t just help students. By acting as an initial competent learning partner, Vinny can save clinical educators time by explaining complex medical histories and intervention strategies to students. This time can now be better spent enhancing students’ clinical problem-solving skills and developing their clinical skills, something AI can’t provide. Vinny also helps clinical educators save time by completing an initial review of student documentation and suggesting clinical feedback and data-based suggestions that clinical educators can include in their final documentation edits for students. Vinny can even go as far as to track students’ clinical writing and skill progress over the quarter to ensure students receive individualized recommendations for skill development and help clinical educators have more consistent and accurate grading of students' clinical skills.
Project Proposal:
The increasing complexity of clinical cases and the focus on evidence-based practice (EBP) in speech-language pathology education require innovative strategies to improve clinical decision-making efficiency. The integration of artificial intelligence (AI) into healthcare, particularly speech-language pathology (SLP), has the potential to revolutionize clinical education and practice by enhancing clinical operations, improving decision-making, and optimizing student learning. AI solutions can help students and clinical educators make informed decisions, leading to better client outcomes and giving students valuable real-world experience in a structured setting.
This proposal outlines the development of a Retrieval-Augmented Generation (RAG) architecture to augment a Large Language Model (LLM) using a supervised training method for a research-based database designed to support SLP graduate students in their clinical education and decision-making processes. By providing evidence-based research, clinical guidelines, and expert insights, this AI tool aims to enhance student learning, improve patient care, and streamline faculty workload.
Database Design and Functionality
The AI database will be designed to provide a comprehensive and user-friendly platform for SLP graduate students. This research database will use a supervised training method and allow current clients to opt in to include their deidentified PII for graduate student training as well as curated dummy profiles created by faculty to provide more data to better train the model. This database will be similar to or integrated with current medical AI databases like Elicit, Consensus, Research Rabit, Nabla, Augmedix, and Fathom.
Key features and functionalities will include:
1. Evidence-Based Research:
* A curated collection of peer-reviewed articles, systematic reviews, and meta-analyses relevant to various SLP disorders and treatment approaches.
* Advanced search capabilities to quickly locate specific information based on keywords, diagnoses, or treatment goals.
* AI-powered summarization of complex research articles to facilitate understanding and application.
2. Clinical Guidelines:
* Access to current clinical practice guidelines and standards of care from reputable organizations such as ASHA.
* Customizable templates for developing treatment plans, progress notes, and discharge summaries.
* AI-assisted decision support tools to aid in diagnosis and treatment planning based on patient characteristics and evidence-based practices.
3. Expert Insights:
* A knowledge base of expert opinions and case studies from experienced SLP clinicians.
* A forum for students to ask questions and receive feedback from peers and faculty.
* AI-powered chatbots to provide immediate answers to common questions and troubleshoot technical issues.
4. Skill Development:
* Interactive tutorials and simulations to practice clinical skills, such as assessment, intervention, and counseling.
* Personalized learning plans based on individual student needs and progress.
* AI-powered assessment of student performance on clinical tasks and provision of targeted feedback.
Benefits for Graduate Students
* Enhanced Clinical Decision-Making: Access to evidence-based research and expert insights will empower students to make informed decisions about diagnosis and treatment planning.
* Improved Problem-Solving Skills: AI-powered tools can act as a competent partner for students to assist them in breaking down complex clinical problems and developing effective problem-solving strategies.
* Increased Efficiency: By automating routine tasks, such as literature reviews and documentation, students can focus on higher-level clinical reasoning and patient care.
* Lifelong Learning: The AI database will provide a valuable resource for continued professional development throughout their careers.
Benefits for Faculty Clinical Educators
* Streamlined Workflow: AI can automate tasks such as analyzing lesson plans, summarizing student-faculty meetings, and providing follow-up directions, freeing up faculty time for more meaningful interactions with students.
* Enhanced Student Support: AI-powered tools can provide timely feedback and support to students, allowing faculty to focus on more complex issues and personalized guidance.
* Improved Student Assessment: AI can analyze student performance data to identify strengths and weaknesses, inform targeted interventions, and track progress over time.
* Data-Driven Decision Making: By collecting and analyzing data on student performance and clinical outcomes, faculty can make evidence-based decisions about curriculum development and program evaluation.
Implementation and Evaluation
To successfully implement the AI database, we propose the following steps:
1. Database Development: A team of experts, including SLP clinicians, AI engineers, and educational technologists, will collaborate to design and develop the database.
2. Pilot Testing: A small group of graduate students will be selected to pilot test the database and provide feedback on its usability and effectiveness.
3. Training and Support: Faculty and students will receive comprehensive training on how to use the AI database and its features.
4. Evaluation: The impact of the AI database on student learning, clinical decision-making, and faculty workload will be evaluated through quantitative and qualitative methods.
By leveraging the power of AI, this database has the potential to transform the way SLP graduate students learn and practice. It will equip them with the knowledge, skills, and tools needed to provide high-quality care to their patients. Additionally, it will streamline faculty workload and enhance the overall quality of clinical education programs.
