Innovation update: Exploring AI scribes

Documentation is a critical part of care. But it’s also time-consuming.

Dr. Mark Makowsky

For Dr. Mark Makowsky, Associate Professor in the Faculty of Pharmacy and Pharmaceutical Sciences at the University of Alberta, that challenge sparked a research project that explores the use of artificial intelligence (AI) scribes for clinical documentation. Funded by the Nashat Family Community Pharmacy Fund, part of CFP’s Innovation Fund, the project aims to determine whether AI-generated notes can accurately capture pharmacist consultations and ultimately improve both workflow and patient care.

“Pharmacists in Alberta are conducting increasingly complex consultations in pharmacist-led clinics and community practice settings,” says Makowsky. “Documentation is essential, but it can take attention away from the patient when clinicians are trying to document and provide care at the same time.”

As a first step, Makowsky and Mike Kimmins, a third-year pharmacy student, conducted a pilot study using two simulated consultations involving common upper respiratory tract infections. After each role-play session, both the student and Makowsky independently created documentation notes. A commercially available AI scribe also generated a documentation note from the recorded consultation.

Three independent pharmacist reviewers then evaluated the notes using the Physician Documentation Quality Index (PDQI-9), a validated tool designed to assess documentation quality.

The early findings were encouraging, says Makowsky. The AI-generated notes achieved overall quality scores comparable to those produced by an experienced pharmacist, although results varied somewhat between cases and reviewers. The work also helped identify important considerations for future research, including the need for multiple reviewers and access to consultation transcripts when assessing note accuracy.

“I was surprised that the AI scribe performed as well as it did,” says Makowsky. “But we’ve only looked at two cases. The real test will be evaluating these tools across a much larger number of patients in real-world practice.”

The pilot study also highlighted potential limitations. While the AI-generated notes scored well overall, reviewers found they were sometimes less thorough and less effective at synthesizing clinical information than the notes produced by an experienced clinician. These findings reinforce the importance of pharmacist oversight and review of any AI-generated documentation.

Makowsky is now working to launch a larger Alberta-based study in 2027 involving community pharmacists and real patient consultations. The research will examine not only documentation quality, but broader questions surrounding workflow, patient experience and clinician satisfaction.

Ultimately, he hopes the research will help determine if AI scribes reduce the time pharmacists spend documenting and allow pharmacists to focus more fully on patients during consultations. It’s also important to ask patients how they feel about AI being used in their care.

“Many clinicians are documenting while patients are speaking,” says Makowsky. “If AI can take on some of that burden, it may allow pharmacists to be more present, maintain better eye contact and focus more on the patient rather than the computer screen.”

The work comes at a time when pharmacist scope of practice continues to expand and AI tools are rapidly entering healthcare settings. While research involving physician use of AI scribes is growing, pharmacy-specific evidence remains limited.

“We’re really at the beginning of understanding how these tools can be used in pharmacy practice,” he says. “Our goal is to help ensure they’re implemented appropriately, safely and in ways that benefit both pharmacists and patients.”

Makowsky is currently seeking graduate students interested in pharmacy practice research to help advance the next phase of the project.