Advancing patient care through human-centred and responsible AI

Fraser Street Medical supports the careful use of artificial intelligence to improve patient care, reduce unnecessary administrative work, strengthen clinic operations and support continuous quality improvement.

AI is intended to assist—not replace—physicians, nurse practitioners, nurses, pharmacists, medical office assistants and other healthcare professionals. Clinical judgment, professional accountability and the relationship between the patient and their healthcare team must remain central.

The purpose of this strategy is to provide a practical framework for evaluating, introducing and monitoring AI tools across Fraser Street Medical.

Our objectives

Fraser Street Medical will use AI where it can:

• Improve the quality, safety and consistency of patient care.

• Reduce administrative burden for clinicians and staff.

• Improve documentation and completion of clinical work.

• Support preventive care, recalls and chronic disease management.

• Improve patient access and communication.

• Strengthen clinic workflows and team-based care.

• Identify gaps, delays and opportunities for improvement.

• Allow clinicians and staff to spend more time on work requiring human judgment, empathy and direct patient interaction.

AI should only be used when it provides a meaningful benefit and can be implemented safely.

Guiding principles

1. Patient first

Every use of AI must support safe, respectful and patient-centred care.

• AI should improve—not weaken—the patient experience.
• Patients must continue to feel heard and understood.
• Technology should not interfere with the therapeutic relationship.
• AI-generated information must never be accepted without appropriate human review.

• Urgent or high-risk concerns must continue to be assessed by a qualified healthcare professional.

2. Human oversight and accountability

A human team member remains responsible for every clinical or operational decision.

• AI may suggest, organize, summarize or draft information.
• AI must not independently diagnose, prescribe, triage or communicate significant clinical decisions.
• The responsible clinician must review and approve clinical documentation.
• Staff must confirm that AI-generated work is complete, accurate and assigned to the correct patient.
• AI does not replace the requirement to sign, complete and take ownership of one’s work.

3. Privacy, confidentiality and security

Patient and staff information must be protected at all times.

• Identifiable patient information may only be entered into tools that have been approved for that specific clinical purpose.
• Staff must not enter patient names, personal health numbers, dates of birth, contact information, clinical notes or other identifiable information into public or unapproved AI systems.
• De-identification must be used whenever identifiable information is not necessary.
• AI tools must be reviewed for privacy, security, data storage, data retention and access controls.
• Relevant privacy impact, contractual and regulatory requirements must be addressed before implementation.

• Provincial, professional, health-authority and clinic privacy requirements take precedence over convenience.

4. Accuracy and critical review

AI can produce incomplete, inaccurate or misleading information.

• All AI-generated content must be checked against the patient chart and other reliable sources.

• Clinical recommendations must be supported by current evidence and professional judgment.
• Staff must confirm names, dates, medications, doses, investigations, referrals and follow-up instructions.
• AI should not be used as the sole source for medical, legal, financial or employment-related decisions.
• Errors or concerning patterns must be reported and reviewed.

5. Equity and accessibility

AI should improve access to care without creating new barriers.

• AI tools should be evaluated for bias and unequal performance across patient populations.
• Language, disability, digital literacy, cultural background and access to technology must be considered.
• Patients must have a reasonable alternative when they cannot or do not wish to use an AI-enabled process.
• AI must not be used to disadvantage patients with complex needs or those requiring additional support.

6. Transparency and consent

Patients should understand how technology is being used in their care when the use is meaningful or material.

• Patients should be informed when an AI scribe or similar tool is used during a visit.

• Appropriate consent must be obtained and documented according to clinic policy.

• Patients may decline optional AI-supported processes without compromising their access to appropriate care.

• Staff should be able to explain the purpose of an AI tool in clear language.

7. Continuous improvement

AI tools must be evaluated after implementation.

• Tools should be assessed for accuracy, efficiency, patient experience and staff workload.
• The clinic should measure whether the technology actually solves the intended problem.
• Workflows should be adjusted based on staff and patient feedback.

• Tools that do not provide sufficient benefit or create unacceptable risks should be modified or discontinued.

Areas of application

Clinical documentation

Approved AI scribes may support:

• Drafting visit notes.

• Organizing histories, examinations and plans.

• Preparing patient instructions.

• Drafting referral letters and consultation requests.

• Summarizing lengthy chart information.

The treating clinician must review, correct and sign all documentation. AI-generated notes must accurately reflect what occurred during the encounter.

Preventive care and recalls

AI may help identify patients who may be due for:

• Cancer screening.
• Diabetes monitoring.
• Kidney and cardiovascular risk assessment.

• Immunizations.

• Chronic disease follow-up.
• Laboratory testing.
• Medication monitoring.

Before contacting a patient, staff must review the chart, recent investigations, outstanding requisitions, specialist recommendations and prior procedures to avoid inappropriate or duplicate recalls.

Chronic disease management

AI may assist the healthcare team with:

• Identifying care gaps.
• Organizing diabetes, chronic kidney disease, heart failure, hypertension, dyslipidemia, asthma, COPD and obesity-management workflows.
• Preparing standardized patient education.
• Tracking whether investigations and follow-up plans have been completed.
• Supporting team-based care between physicians, nurse practitioners, nurses, pharmacists and medical office assistants.

Clinical decisions remain the responsibility of the treating healthcare professional.

Patient communication

AI may assist with drafting:

• Appointment instructions.

• Screening reminders.
• Educational messages.
• Follow-up letters.

• Administrative responses.

• Frequently asked questions.

Messages must be reviewed before being sent. Staff must confirm that the communication is accurate, appropriate for the individual patient and written in understandable language.

Clinic administration

AI may support:

• Meeting summaries and action items.
• Policy and procedure development.
• Staff training materials.
• Workflow mapping.
• Project planning.
• Website and patient education content.
• Quality-improvement reports.
• Drafting non-clinical emails and documents.

Confidential information must only be used in approved systems.

Telephone and reception support

AI may eventually assist with:

• Organizing telephone queues.
• Categorizing administrative requests.
• Identifying common call reasons.
• Preparing scripts and booking guidance.
• Supporting virtual reception workflows.
• Identifying calls that require escalation.

AI must not independently provide emergency advice, make a diagnosis or override Fraser Street Medical’s booking and triage protocols.

Responsibilities of all Fraser Street Medical team members

Every team member using AI must:

• Use only approved tools for patient-related work.
• Follow clinic privacy and confidentiality policies.

• Review AI-generated content before using, saving or sending it.

• Correct errors rather than assuming another person will do so.

• Confirm that work is placed in the correct patient chart.

• Complete, sign and close their work appropriately.

• Escalate clinical questions to a qualified healthcare professional.

• Report privacy concerns, unsafe outputs and repeated inaccuracies.

• Participate in required training.

• Use AI only within their role and scope of practice.

Fraser Street Medical position statement

Fraser Street Medical supports responsible AI innovation when it improves patient care, strengthens the healthcare team and reduces unnecessary administrative burden.

AI will be implemented deliberately, with appropriate privacy safeguards, human oversight, patient transparency and ongoing evaluation.

Technology may assist our work, but responsibility for safe, compassionate and high-quality care will always remain with the people providing that care.