Reimagining geriatric care in the age of artificial intelligence

What began as a simple ChatGPT search ultimately led Dr. Nihal Haque to expand his geriatric medicine expertise into the world of artificial intelligence research.

Dr. Haque is a geriatrician at North York General Hospital (NYGH) and an Adjunct Assistant Professor at the University of Toronto. He was also recently named an AMS-Fitzgerald Fellow in AI and Human-Centred Leadership.

His interest in the care of older adults began during his internal medicine residency when he became fascinated by how medical advances have successfully prolonged lifespan, and by the emerging challenge of improving quality of life as patients age.

“I also really enjoy this age group of patients as they are filled with rich life experiences, which I always learn from myself as I am involved in their care,” he says. “These experiences encouraged me to pursue a career in geriatric medicine.”

When ChatGPT first became publicly available, Dr. Haque began experimenting with the tool by asking it questions related to geriatrics. That curiosity led him to publish a commentary in the Journal of the American Geriatrics Society (JAGS) reflecting on his experience using the program. In the process, however, he also discovered its limitations – most notably, its tendency to fabricate references.

“AI will change every aspect of our lives, including medical care delivered to our patients,” he says. “For this reason, I wanted to explore AI in medicine, and more specifically, geriatrics.”

With AI technology poised to expand rapidly over the coming years, Dr. Haque hopes clinicians will feel equipped to critically appraise these tools and thoughtfully select those that meaningfully improve patient care.

In early 2023, NYGH established an AI working group, led by Vice President of Strategy & Digital Health, Duska Kennedy, and Chief Medical Information Officer, Dr. Phil Shin, to explore how AI could inform the hospital’s 10-year strategic plan. Representatives from across the organization collaborated to identify three strategic priority projects.

Dr. Haque proposed a project aimed at expediting the production of discharge summaries using large language models (LLMs) and the group successfully secured funding from Canada Health Infoway to support the research.

Additionally, Dr. Haque has been collaborating with NYGH’s AI Research Chair, Dr. Ervin Sejdic, to develop an AI-simulated video chat system designed to reduce agitation, restraint use, and antipsychotic prescribing in older adults hospitalized with delirium. The system provides calming, personalized interactions through simulated video chat that engages patients in comforting dialogue and supports re-orientation. The team is currently evaluating feasibility, patient experience, and its potential impact on agitation management in delirium.

Dr. Haque has also partnered with Dr. Michael Liut, Assistant Professor in the Teaching Stream (Computer Science) at the University of Toronto Mississauga, to develop a chatbot called MedBot. Designed as a simulated patient, MedBot helps medical learners practice interviewing skills and uncover causes of falls in older adults. Initially developed as a text-based platform, the system has evolved to support voice-to-voice interactions while delivering feedback assessed using an instructor’s rubric. It is easily accessible to University of Toronto learners through integration with university platforms.

By offering a safe environment to practice history-taking, clinical reasoning, and management planning, the chatbot supports learner development without risk to patients. Early pilot work suggests it improves learner engagement and interest in the topics – an encouraging finding for Dr. Haque and his collaborator.

Beyond his research, Dr. Haque serves as a faculty advisor on the education committee of the Temerty Centre for AI Research and Education in Medicine (T-CAIREM). In this role, he developed one of Canada’s first accredited Continuing Medical Education online courses to support practising physicians as they navigate the evolving AI landscape in medicine. The course covers core AI concepts, as well as their implications for physicians’ ability to deliver care in the coming years.

The course launch was highly successful, prompting T-CAIREM to offer a second iteration within the same year due to demand. Additional sessions are now planned for 2026 and beyond.

“AI will change how we practise medicine and will augment us as we endeavour to deliver better patient care,” he says. “As such, it will be important to always centre the patient’s perspective in any discussions on AI in healthcare.”