ClinicGPT — Large Language Model for Institutional Dental Clinics
Sep 5, 2023
Western University, Schulich School of Medicine & Dentistry
Designing and validating a domain-specific language model that helps dental school clinics generate high-quality patient communication and documentation while protecting institutional data.
What I did
Before Rebrief, I wanted to understand how large language models could safely support real dental clinics. At Schulich Dentistry, I:
Designed ClinicGPT, a custom LLM workflow tuned on de-identified institutional documents and guidelines.
Focused on practical clinic tasks: drafting referral letters, post-operative instructions, and patient education materials aligned with school policies.
Built an interface that allowed staff and students to review, edit, and approve outputs rather than treating them as final.
Led the evaluation study, measuring readability, accuracy, and clinician satisfaction compared to existing templates and manual drafting.
Results
ClinicGPT produced drafts that reduced the time to create patient-facing documents while maintaining high ratings for clarity and professionalism.
The work was published in the Journal of Dental Education, one of the first descriptions of a tailored LLM workflow for academic dental clinics.
The project provided a blueprint for responsible LLM deployment in healthcare, directly informing later work on Rebrief’s AI pipelines and safety controls.
