Leveraging Copilot to Develop Clinical Case Studies

Leveraging Copilot to Develop Clinical Case Studies

Emily Wood teaches in the Department of Speech Language Pathology and is currently is the instructor for several courses spanning the two-year clinical master's program, including SLP1520H: Principles of Clinical Practice; SLP1535H Advanced Principles of Clinical Practice; SLP1527H: Clinical Analysis of Communication and Swallowing disorders; and SLP1521H: Alternative and Augmentative Communication. She has also taught SLP1505Y: Child Language Development and Assessment and SLP1507H: Clinical Laboratory in Speech-Language Pathology. Emily’s research is focused on (i) pedagogical inquiry - specifically examining the impact of integrating experiential learning in early academic coursework for healthcare professionals; and (ii) clinical practice research - with a focus on equitable assessment practices in speech-language pathology.  
Emily Wood
Assistant Professor, Teaching Stream, Department of Speech-Language Pathology, Temerty Faculty of Medicine

Objectives

The primary objective of this assignment was to provide students with a case-based assessment task using an AI-generated clinical case in the context of child language development. The goal was for students to apply their knowledge to analyze the case, distinguish between language differences and disorders, and demonstrate clinical reasoning skills.  

By leveraging generative AI, instructors may efficiently produce case scenarios that reflect real-world complexity, while maintaining accuracy and cultural sensitivity. This approach to assignment design also supports instructors’ skills in prompt design and critical evaluation of AI outputs, ensuring that generated content is pedagogically sound and free from bias.

Process

Emily Wood integrated generative AI into a case-based learning (CBL) framework for a child language development course in Speech-Language Pathology. The process followed seven structured steps:

Step 1: Instructor Identifies Learners (Prompt Development)

  • Target audience: First-year master’s students in Speech-Language Pathology at the University of Toronto
  • Course context: Child language development

Step 2: Instructor Sets Assessment Parameters (Prompt Development)

  • Formative, in-class group activity (3–5 students per group)
  • Duration: 20 minutes
  • Case length: One page with 3–5 open-ended questions

Step 3: Instructor Describes Case Particularities (Prompt Development)

  • Focus: A bilingual or multilingual child (age 4–6) with possible developmental language disorder
  • Context: Public school setting in Toronto; no other health concerns

Step 4: Instructor Defines Skills to be Evaluated (Prompt Development)

  • Discriminate between language difference and disorder in multilingual populations
  • Identify case details suggesting language difference vs. disorder
  • Formulate preliminary clinical conclusions

Step 5: Instructor Integrates and Refines Prompt (Prompt Revision)

  • Combine all parameters into a comprehensive Copilot prompt to generate the case

Step 6: Instructor Review and Revision (Prompt Revision)

  • Validate case for relevance, accuracy, and bias
  • Example cases included profiles of children (e.g., Maya and Amar) with detailed linguistic backgrounds and discussion questions:
    • Which aspects suggest typical bilingual acquisition?
    • Which aspects indicate possible developmental language disorder?
    • What further assessments would clarify the diagnosis?

Step 7: Student Engagement With the AI-Generated Cases

  • Prior to the final exam, students collaboratively work in groups to analyze the AI-generated cases
  • In the summative final exam, students independently respond to open-ended questions, applying clinical reasoning and theoretical knowledge

Future-Focused Skill Development

This approach to assessment design aligns with the University of Calgary’s STRIVE model by emphasizing Transparency, Responsibility, and Validity in the design of AI-enabled assessments. Transparency is demonstrated through clear communication about how generative AI is used to create case studies and the rationale behind its integration. Responsibility is reflected in the instructor’s role in critically reviewing and refining AI-generated cases to ensure accuracy, cultural sensitivity, and ethical standards. Finally, Validity is supported by aligning the AI-generated cases with course learning outcomes and professional competencies, ensuring that the assessment remains authentic and pedagogically sound.

Student Feedback

Professor Wood shares:

Student feedback following the final exam was overwhelmingly positive. Many students expressed that the case study component encouraged them to integrate theoretical content learned over the course of the semester, with their developing clinical reasoning skills in a way that mirrors the scenarios they anticipate facing in their upcoming clinical placements. One student commented, “I find that the [AI-generated] cases at the end of the exam are a true assessment of what we may encounter in practice.”  

In their final course evaluations several students also requested that additional case-based learning opportunities be incorporated throughout the semester to help them familiarize themselves with this method of critical thinking and knowledge integration. For example, one student wrote “…Maybe have more case studies/practice questions that help again integrate knowledge.”  

As an educator, I also found using an AI-generated case study and short answer approach in the exam to be a more effective means of evaluating student knowledge and competence. With this approach, there were multiple rather than a singular pathway to a correct response, which in my experience as a clinician is more reflective of real-world practice. This contrasts with a multiple-choice or traditional short-answer questions where typically only one answer is correct, and no context or reasoning is provided or accounted for. Finally, I found that while grading exam answers, I garnered valuable insights into how students were connecting concepts across the course while also highlighting areas where further instruction might be needed.

Back to Top