Skip to main content

Main menu

  • HOME
  • ARTICLES
    • Current Issue
    • Abstracts In Press
    • Archives
    • Special Issue Archive
    • Subject Collections
  • INFO FOR
    • Authors
    • Reviewers
    • Call For Papers
    • Subscribers
    • Advertisers
  • SUBMIT
    • Manuscript
    • Peer Review
  • ABOUT
    • The JABFM
    • The Editing Fellowship
    • Editorial Board
    • Indexing
    • Editors' Blog
  • CLASSIFIEDS
  • Other Publications
    • abfm

User menu

Search

  • Advanced search
American Board of Family Medicine
  • Other Publications
    • abfm
American Board of Family Medicine

American Board of Family Medicine

Advanced Search

  • HOME
  • ARTICLES
    • Current Issue
    • Abstracts In Press
    • Archives
    • Special Issue Archive
    • Subject Collections
  • INFO FOR
    • Authors
    • Reviewers
    • Call For Papers
    • Subscribers
    • Advertisers
  • SUBMIT
    • Manuscript
    • Peer Review
  • ABOUT
    • The JABFM
    • The Editing Fellowship
    • Editorial Board
    • Indexing
    • Editors' Blog
  • CLASSIFIEDS
  • JABFM on Bluesky
  • JABFM On Facebook
  • JABFM On Twitter
  • JABFM On YouTube
Research ArticleOriginal Research

Priorities for Artificial Intelligence Applications in Primary Care: A Canadian Deliberative Dialogue with Patients, Providers, and Health System Leaders

Tara L. Upshaw, Amy Craig-Neil, Jillian Macklin, Carolyn Steele Gray, Timothy C. Y. Chan, Jennifer Gibson and Andrew D. Pinto
The Journal of the American Board of Family Medicine March 2023, jabfm.2022.220171R1; DOI: https://doi.org/10.3122/jabfm.2022.220171R1
Tara L. Upshaw
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
MHSc
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Amy Craig-Neil
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
MSc
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Jillian Macklin
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
MSc
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Carolyn Steele Gray
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
MA, PhD
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Timothy C. Y. Chan
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
PhD
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Jennifer Gibson
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
PhD
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Andrew D. Pinto
From the Upstream Lab, MAP/Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, Unity Health Toronto, Toronto, Ontario, Canada (TLU, CAN, JM, ADP); Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada (TLU); Undergraduate Medical Education, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (TLU, JM); Bridgepoint Collaboratory for Research and Innovation, Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, Ontario, Canada (CSG); Institute of Health Policy, Management and Evaluation, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (CSG, ADP); Department of Mechanical and Industrial Engineering, Faculty of Applied Science and Engineering, University of Toronto, Toronto, Ontario, Canada (TCYC); Joint Centre for Bioethics, University of Toronto, Toronto, Ontario, Canada (JG); Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada (JG); Department of Family and Community Medicine, St. Michael's Hospital, Toronto, Ontario, Canada (ADP); Department of Family and Community Medicine, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (ADP).
MD, MSc
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
  • Article
  • Figures & Data
  • References
  • Info & Metrics
  • PDF
Loading

Article Figures & Data

Figures

  • Tables
  • Figure 1.
    • Download figure
    • Open in new tab
    Figure 1.

    Deliberative dialogue process. Abbreviations: AI, artificial intelligence; PHC, primary health care.

  • Appendix 2 Figure 1.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 1.

    Session topics for Future Perspectives on AI in Canadian Primary Care dialogue series.

  • Appendix 2 Figure 2.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 2.

    Artificial intelligence, machine learning, and Big Data (adapted from Jillian Macklin).

  • Appendix 2 Figure 3.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 3.

    Passport photograph of Alan Turing, the father of artificial intelligence, at age 16, by unknown author/CC BY 4.0.

  • Appendix 2 Figure 4.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 4.

    Layers of a supervised deep learning algorithm, by unknown author.

  • Appendix 2 Figure 5.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 5.

    Example photographs of cats and dogs that can be used to train a deep learning image classification algorithm, by unknown author.

  • Appendix 2 Figure 6.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 6.

    Facebook CEO Mark Zuckerberg demonstrates Facebook's facial filtering technology by Mark Zuckerberg.

  • Appendix 2 Figure 7.
    • Download figure
    • Open in new tab
    Appendix 2 Figure 7.

    Example images from a test dataset used to assess the performance of a deep learning algorithm by Esteva et al.12

Tables

  • Figures
    • View popup
    Table 1.

    Patient, Provider and Health System Leader Demographics

    PatientsProvidersHealth System Leader
    Province n (%)
    British Columbia1 (4.5%)1 (4.8%)1 (20%)
    Alberta4 (18%)2 (9.5%)
    Saskatchewan--1 (20%)
    Manitoba-5 (24%)
    New Brunswick1 (4.5%)-
    Nova Scotia3 (14%)1 (4.8%)*
    Ontario12 (55%)11 (52%)3 (60%)
    Quebec1 (4.5%)1 (4.8%)
    Age in years Range, mean (SD)23 to 73, 40 (16)28 to 64, 42 (8.7)
    Gender n (%)
    Women†12 (55%)9 (43%)
    Men9 (41%)12 (57%)
    Nonbinary1 (5%)-
    Race or ethnicity n (%)
    Black2 (9%)1 (4.8%)
    East/Southeast Asian1 (5%)2 (9.5%)
    South Asian6 (27%)5 (24%)
    White13 (59%)10 (48%)
    Mixed-3 (14%)
    Self-rated artificial intelligence knowledge‡2.7 (1.1)3.3 (1.2)
    Mean (SD)
    Provider type n (%)
    Chiropractor-1 (4.8%)-
    Clerical staff-1 (4.8%)-
    Family physician-14 (67%)-
    Family medicine resident-2 (9.5%)-
    Nurse practitioner-2 (9.5%)-
    Social worker-1 (4.8%)-
    Years in practice Mean (SD)§-12 (10)-
    Practice size‖
    <250 patients-6 (30%)-
    250 to 750 patients-6 (30%)-
    >750 to 1250 patients-5 (25%)-
    Full-time equivalent clinical hours per week Mean (SD)¶-0.67 (0.30%)-
    • Notes: Due to rounding, some totals may not perfectly sum to 100.

    • ↵* Canadian province in which Denmark-licensed family physician studied health information technology in primary care settings.

    • ↵† Including one trans woman.

    • ↵‡ Participants rated their knowledge of AI on a 5-point Likert scale (1 was “Not knowledgeable at all,” 5 was “extremely knowledgeable.”

    • ↵§ Calculation includes years in residency.

    • ↵‖ Excluding residents, clerical participant.

    • ↵¶ Calculation excludes clerical participant.

    • Abbreviation: SD, standard deviation.

    • View popup
    Table 2.

    Overview of Themes

    Theme 1Priority applications of AI in primary care
    Main ideaHighest priority applications of AI are to areas where the current state of technology drives provider burnout, challenges patient-centeredness, or limits access to care.
    QuotesAs a patient, I don't want my doctor spending his time facing the computer. I want him facing me…So in terms of looking at all of those admin tasks that are taking away from the patient care, I think AI has the potential to free up that time so that I have more face time with my doctor. — 39-year-old patient from British Columbia
    I think that doctors are often overwhelmed and overworked and if AI can be used to help with that, I'm all for that, so that they can be more efficient and more effective in their work. — 34-year-old patient from Alberta
    I was struck by the overlap in interests between providers and patients…What stood out was creating more time to be able to focus on actual patient and provider interaction… just having more of that time not taken up by all these other nonclinical issues. That…could represent a very safe, low-risk place to start and to sort of build upon AI within the primary care setting. — 42-year-old family physician from Manitoba
    My sense is that [triage in primary care] is absolutely abysmal…Like, there is no function, currently, of sophisticated pre-visit triage…In [my] clinic, people wait 20 minutes and hang up. They don't even call because it's so hard to get in. That's our triage system. If [there was] a way to…symptom check or…prescreen a little bit, they might be more likely to [come in]. — 48-year-old family physician from Ontario
    Theme 2Impact of AI on primary care provider roles
    Main ideaAI is not a substitute for provider expertise. It should be applied in ways that supplement core clinical skills and enhance patient-centered care.
    QuotesWhen I'm struggling [to manage my diabetes], sometimes…I'm just tired of being a diabetic. It's not because…I don't know how to take care of myself…It's only when a trust relationship has been built up with the doctor that he can begin to say, “Okay, I know you know how to take care of it. You don't seem to be taking care of it right now. What's going on for you?” I think you have to rely…on the trust relationship between the doctor and the patient to recognize specifically what's going on. It may be more subtle than the things that AI might pick up. — 73-year-old patient from Alberta
    In my case, I can tell you for sure that an AI would say, “Oh, she needs this prescription.” Meanwhile, that could kill me…There are nuances here that I don't think an AI could know…And look at it this way: in finance, we have controls in place…So, I would want the doctor to review that first. I'm all for them not having to do as much typing. I think it would save time, but there would have to be that review. — 55-year-old patient from Ontario
    Electronic health records have advantages for sure…But…the one thing I miss so much [is] that I can no longer do a genogram. There's nothing. They were never designed [for EMRs]. So, my family histories are so different than how I was taught with that picture. I did it with patients and we could really understand their family history from so many different angles. I used to love that part of care…We just have to be careful [with AI so] that we don't lose…strengths of [the] older model. — 57-year-old family physician from Ontario
    How does AI…consider the triad of, you know, what is the evidence? What is my experience that I've had after 35 years in practice? And what are the patient preferences?…I'm not sure how AI could pick up understanding my clinical experience. I don't know how AI can pick up what a patient's preferences [are] either. So, I think [of AI] as a tool…within our evidence-based medicine model. — 64-year-old chiropractor from Ontario
    Theme 3Considerations for provider training in AI
    Main ideaFormative and continuing education of primary care and other health professionals should cultivate basic AI literacy, algorithm critical appraisal skills, and safe, effective use within clinical reasoning processes and workflows.
    QuotesI would like to trust that my provider has a good perspective in AI so that he doesn't just sort of follow it slavishly but considers it as part of his care for me. — 73-year-old patient from Alberta
    I don't mean to sound like an alarmist about the dependency part. I'm not being reactionary. It's just, I have concerns about that. So, I wonder how [the use of AI] would be monitored, how doctors would be trained, how we would ensure that patients are getting still this benefit of experience and knowledge and not just this dependency. — 55-year-old patient from Ontario
    I have a 15-year-old who's learning how to drive. We have one car that has sensors and all sorts of safety features, and another car that doesn't….I feel like he needs to understand how to operate a vehicle at its base level before he can really make use of [safety features]…Maybe we need to emphasize the diagnostic reasoning, the history and physical pieces and the test ordering first, and introduce AI to that senior clinical learner…rather than right off the bat, so that they've got those building blocks behind them. — 44-year-old family physician from Alberta
    I think that part of what can help physicians [manage] our medical legal liability and risk is education and training around what these technologies are and what their purpose is…It's going to take a lot of awareness building among physicians to stay on top of how we can practice safely with the best interests of our patients in mind as these technologies become more widespread. — 42-year-old family physician from Manitoba
    • AI, artificial intelligence.

    • View popup
    Table 3.

    Desired Functions and Benefits of Priority Applications of Artificial Intelligence in Primary Care

    Application AreaDesired FunctionsDesired Benefits
    Higher priority
    Documentation
    • Automate charting

    • Manage prescriptions (generation, refilling, forwarding)

    • Manage referrals

    • Mitigate provider burnout

    • Liberate time and cognitive freedom for:

      • ∘ Direct, face-to-face patient interaction

      • ∘ Discussions during a visit for patient goals, preferences, and circumstances

      • ∘ Managing medical and social complexity

      • ∘ Coordinating access to care for patients who face high barriers

    Practice operations
    • Collect and verify patient information

    • Optimize staff and learner scheduling†

    • Predict surges in visits to direct resource planning†,§

    Triage
    • Set an agenda before a visit by distilling patient concerns and taking partial histories

    • Assign concerns to management by virtual or in-person care modalities

    • Consider quality of life and functional impacts in determination of acuity*

    • Help patients decide if they can safely self-manage a concern, and if not, what health services they require

    • Prioritized access to synchronous modality of care delivery for those at risk of decompensation, medical need

    • Greater convenience for patients who receive virtual care, especially those with mobility challenges or living rurally*

    • Optimized provision of and access to care across delivery modalities

    • Decreased wait times*

    Lower priority
    Clinical decision support
    • Synthesize administrative, clinical (eg, patient history, past treatments), biometric data, and other “sources of truth” (eg, evidence, guidelines) to guide diagnosis, treatment, and care planning of rare diseases and common conditions†

    • Present guidance in real-time during a consult†,§

    • Assist with management of medical complexity

    • Provide a “second opinion”; help providers “think outside the box”; transfer specialist knowledge into primary care

    • Improve speed and accuracy of diagnosis

    • Reduce trial-and-error treatments, uninformative or burdensome diagnostic tests, and multiple specialist referrals

    • Reduce rates of medical error

    Proactive and preventative care
    • Identify patients who are at high risk of decompensation or loss to follow-up, particularly those who experience structural vulnerability†,§

    • Recall patients with abnormal test results or those due for routine screening†,§

    • Automate early referral to interprofessional care†

    • Optimize contact with the provider to improve

      • ∘ Care access

      • ∘ Health equity

      • ∘ Patient safety

      • ∘ Team-based care†

    • Notes: Views held exclusively by one participant group are indicated by the symbols below. The absence of a symbol reflects a view shared by patients and provides.

    • ↵* Patient only view.

    • ↵† Provider only view.

    • ↵§ System leader support.

    • View popup
    Appendix 1 Table 1.

    Participant Sampling Frames and Recruitment Strategies

    Participant GroupSampling FrameEligibility CriteriaRecruitment
    PHC patientsPurposive maximum variation (age, gender, race and ethnicity, education level, income, province of residence)
    • English-speaking

    • Online advertisements (Kijiji, social media channels) in 10 provinces

    • Aged 18 or older

    • Patient advisor distribution list*

    • Visited PHC provider at least once in last year (virtual or in-person)

    PHC providersPurposive maximum variation (gender, race and ethnicity, provider type, country of health professions education, years in practice, practice size, and province of practice)
    • Any provider or administrative support person working in a PHC setting

    • Emailed invitation from Dr. Andrew Pinto (AP) to PHC colleagues interested in digital health or health informatics

    • Worked at least one clinical day per week

    • Emailed invitation from AP to practice colleagues†

    • Online advertisements (social media channels, PHC research network news channels)

    System leadersCritical caseInvolved in digital health, health informatics, or PHC governanceEmailed invitation from AP send to individuals within study team members' professional networks and those in relevant roles listed on health ministry public directories in 5 provinces‡
    • ↵* Patient advisory network in British Columbia, patient-family advisory council of downtown Toronto hospital.

    • ↵†Downtown Toronto academic family health team.

    • ↵‡British Columbia, Alberta, Saskatchewan, Manitoba, Ontario.

    • Abbreviation: PHC, primary health care.

    • View popup
    Appendix 1 Table 2.

    Patient, Provider, and Health System Leader Demographics

    PatientsProvidersHealth System Leader
    Province
    British Columbia1 (4.5%)1 (4.8%)1 (20%)
    Alberta4 (18%)2 (9.5%)
    Saskatchewan--1 (20%)
    Manitoba-5 (24%)
    New Brunswick1 (4.5%)-
    Nova Scotia3 (14%)1 (4.8%)*
    Ontario12 (55%)11 (52%)3 (60%)
    Quebec1 (4.5%)1 (4.8%)
    Age in years Range, mean (SD)23 to 73, 40 (16%)28 to 64, 42 (8.7%)
    Gender n (%)
    Female†12 (55%)9 (43%)
    Male9 (41%)12 (57%)
    Nonbinary1 (5%)-
    Race or ethnicity n (%)
    Black2 (9%)1 (4.8%)
    East/Southeast Asian1 (5%)2 (9.5%)
    South Asian6 (27%)5 (24%)
    White13 (59%)10 (48%)
    Mixed-3 (14%)
    Self-rated artificial intelligence knowledge‡ Mean (SD)2.7 (1.1)3.3 (1.2)
    Provider type
    Chiropractor-1 (4.8%)-
    Clerical staff-1 (4.8%)-
    Family physician-14 (67%)-
    Family medicine resident-2 (9.5%)-
    Nurse practitioner-2 (9.5%)-
    Social worker-1 (4.8%)-
    Years in practice Mean (SD)§-12 (10)-
    Practice size‖
    <250 patients-6 (30%)-
    250 to 750 patients-6 (30%)-
    750 to 1250 patients-5 (25%)-
    Full-time equivalent clinical hours per week Mean (SD)¶-0.67 (0.30%)-
    • Notes: Due to rounding, some totals may not perfectly sum to 100.

    • ↵* Canadian province in which Denmark-licensed family physician studied health information technology in primary care settings.

    • ↵†Including one trans woman.

    • ↵‡Participants rated their knowledge of AI on 5 five-point Likert scale (1 was “Not knowledgeable at all,” 5 was “extremely knowledgeable.”

    • ↵§ Calculation includes years in residency.

    • ↵‖ Excluding residents, clerical participant.

    • ↵¶ Calculation excludes clerical participant.

    • Abbreviation: SD, standard deviation.

    • View popup
    Appendix 2 Table 1.

    Potential Uses of AI in Primary Care

    CategoryDefinitionExample
    Self-care, illness prevention, and wellnessTools that support people in living healthier livesA machine learning algorithm analyzes vital sign data from a patient's smartwatch in real time, documenting trends in their primary care electronic medical record (EMR). They receive personalized reminders to exercise, eat well, and get enough sleep. Their physician is alerted when trends show a decline in heart health.
    Triage and early diagnosisTools that help triage patients and identify the need for additional health resourcesA machine learning-based symptom checker informs a patient with a gradual development of severe foot pain to book an appointment with their primary care provider as early as possible.
    DiagnosticsTools that assist providers with point-of-care diagnosisA primary care provider uploads a cell phone photo taken of a patient's retina to an app that uses deep learning to predict the risk of complications from diabetes. He refers the patient to an ophthalmologist.19
    Clinical decision supportTools that structure relevant information to help physicians determine treatment course or need for referral to specialist or acute care servicesAn EMR-integrated machine learning algorithm predicts which patients are at high risk for becoming infected with HIV within a 3-year timeframe. Risk profiles can help primary care providers who would most benefit from pre-exposure prophylaxis medications.20
    Care deliveryTools that support direct interactions between patients and providersA natural language processing tool automatically converts the conversation between a patient and provider into chart notes, orders laboratory tests, and writes referrals to specialists during a clinic visit. This tool can also reach out to patients in advance of the appointment to gather necessary information.21
    Chronic care managementTools that help patients and providers manage chronic diseases like diabetes or heart diseaseA patient with diabetes has a blood glucose monitor that syncs with an AI-based app on their phone. The algorithm learns the patient's dietary and insulin delivery schedule over time. It begins to send helpful reminders to eat, check blood glucose, and inject insulin. The app is integrated with the patient's primary care EMR. It notifies the provider when the patient's insulin needs appear to change significantly.
    Population health managementTools that analyze large datasets to identify trends in population health to inform shifts in clinical programs and intervention targetingA deep learning algorithm analyzes a clinic's raw EMR data. It identifies the patients at the highest risk for hospital admission within the next 30 days. Providers in the clinic schedule appointments with these patients to discuss their health and preventative interventions.14
    Health care operationsTools that decrease time spent on routine administrative tasks that occur in the background of patient careA classical machine learning algorithm learns that times and days of the week where appointments are in highest demand, and helps clinic clerical staff optimize the staffing schedule
Previous
Back to top

In this issue

The Journal of the American Board of Family     Medicine: 38 (1)
The Journal of the American Board of Family Medicine
Vol. 38, Issue 1
January-February 2025
  • Table of Contents
  • Table of Contents (PDF)
  • Cover (PDF)
  • Index by author
  • Back Matter (PDF)
  • Front Matter (PDF)
Print
Download PDF
Article Alerts
Sign In to Email Alerts with your Email Address
Email Article

Thank you for your interest in spreading the word on American Board of Family Medicine.

NOTE: We only request your email address so that the person you are recommending the page to knows that you wanted them to see it, and that it is not junk mail. We do not capture any email address.

Enter multiple addresses on separate lines or separate them with commas.
Priorities for Artificial Intelligence Applications in Primary Care: A Canadian Deliberative Dialogue with Patients, Providers, and Health System Leaders
(Your Name) has sent you a message from American Board of Family Medicine
(Your Name) thought you would like to see the American Board of Family Medicine web site.
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
4 + 0 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.
Citation Tools
Priorities for Artificial Intelligence Applications in Primary Care: A Canadian Deliberative Dialogue with Patients, Providers, and Health System Leaders
Tara L. Upshaw, Amy Craig-Neil, Jillian Macklin, Carolyn Steele Gray, Timothy C. Y. Chan, Jennifer Gibson, Andrew D. Pinto
The Journal of the American Board of Family Medicine Mar 2023, jabfm.2022.220171R1; DOI: 10.3122/jabfm.2022.220171R1

Citation Manager Formats

  • BibTeX
  • Bookends
  • EasyBib
  • EndNote (tagged)
  • EndNote 8 (xml)
  • Medlars
  • Mendeley
  • Papers
  • RefWorks Tagged
  • Ref Manager
  • RIS
  • Zotero
Share
Priorities for Artificial Intelligence Applications in Primary Care: A Canadian Deliberative Dialogue with Patients, Providers, and Health System Leaders
Tara L. Upshaw, Amy Craig-Neil, Jillian Macklin, Carolyn Steele Gray, Timothy C. Y. Chan, Jennifer Gibson, Andrew D. Pinto
The Journal of the American Board of Family Medicine Mar 2023, jabfm.2022.220171R1; DOI: 10.3122/jabfm.2022.220171R1
Twitter logo Facebook logo Mendeley logo
  • Tweet Widget
  • Facebook Like
  • Google Plus One

Jump to section

  • Article
    • Abstract
    • Introduction
    • Methods
    • Results
    • Discussion
    • Acknowledgments
    • Appendix 1. Deliberative Dialogue Process and Guides
    • Appendix 2. Participant Informational Module
    • Notes
    • References
  • Figures & Data
  • References
  • Info & Metrics
  • PDF

Related Articles

  • No related articles found.
  • PubMed
  • Google Scholar

Cited By...

  • Artificial intelligence in primary care practice: Qualitative study to understand perspectives on using AI to derive patient social data
  • Developing an AI Tool to Derive Social Determinants of Health for Primary Care Patients: Qualitative Findings From a Codesign Workshop
  • Research Representing the Changing Landscape of Family Medicine
  • Google Scholar

More in this TOC Section

  • Identifying and Addressing Social Determinants of Health with an Electronic Health Record
  • Integrating Adverse Childhood Experiences and Social Risks Screening in Adult Primary Care
  • A Pilot Comparison of Clinical Data Collection Methods Using Paper, Electronic Health Record Prompt, and a Smartphone Application
Show more Original Research

Similar Articles

Keywords

  • Artificial Intelligence
  • Canada
  • Family Medicine
  • Qualitative Research

Navigate

  • Home
  • Current Issue
  • Past Issues

Authors & Reviewers

  • Info For Authors
  • Info For Reviewers
  • Submit A Manuscript/Review

Other Services

  • Get Email Alerts
  • Classifieds
  • Reprints and Permissions

Other Resources

  • Forms
  • Contact Us
  • ABFM News

© 2025 American Board of Family Medicine

Powered by HighWire