Table 3

Final list of priority areas for AI and PC identified and ranked in the multi-stakeholder collaborative consultation day

RankAI and PHC priorityExtended descriptions from small group discussions
1Preventative care and risk profilingOverarching goal: Support decisions in cases of uncertainty around screening and/or potential diagnoses, and to free up time during clinical consults.
Example specific tasks/outcomes:
  • Screening reminders for patients at high risk of negative outcomes; reminders would be more personalised than general guidelines.

  • Facilitate earlier diagnosis when potentially beneficial and mitigate unnecessary testing otherwise.

2Patient self-management of condition(s)Overarching goal: Support patient self-care or self-management of condition(s) with the possibility of sharing information between patients and providers.
Example specific tasks/outcomes:
  • Vaccines, including COVID-19, or medication reminders.

  • Health coaching and other resources to support goal achievement, including feedback on progress between clinical appointments.

  • Scheduling and appointment reminders.

  • Education about conditions and expectations.

3Management and synthesis of information sourcesOverall goal: Use, combine, and/or synthesise information from multiple sources to expand the scope of practice and improve equity and care access.
Example specific tasks/outcomes:
  • Identify relevant information sources/content for different (1) users, (2) questions, and (3) tasks.

  • Manage the overwhelming amount of information from multiple sources.

4Improved communication between PC and AI stakeholdersOverall goal: Support communication between AI practitioners, PC practitioners, and patients to mitigate misunderstandings and poor application of techniques.
Example specific tasks:
  • Establish a shared vocabulary/lexicon.

  • Include PC, AI, and patient stakeholders on projects.

5Data sharing and interoperability between providersOverarching goal: Improve data sharing and interoperability between providers and jurisdictions.
Example specific tasks:
  • Establish data standards to enable interoperability.

  • Establish data linkages between provinces and health systems.

  • Highlight the potential for individuals to contribute through data sharing, similar to concepts used for organ donation and the greater good.

6A (tie)Clinical decision supportOverarching goal: Support care decisions during times of uncertainty and/or high demand by providing suggestions or support to clinicians.
Example specific tasks:
  • Individualised recommendations for interventions at the individual or group level.

  • Standardise and/or summarise information from EMRs so that both patients and clinicians have access and can track relevant data.

  • Support patient triage decisions during times of high demand (eg, COVID-19 pandemic recovery phase) and as clinicians adjust to different modes of delivery post-pandemic (eg, continuation of virtual visit options).

6B (tie)Administrative staff supportOverarching goal: Support administrative staff and patient appointment preparation.
Example specific tasks:
  • Scheduling.

  • Patient triage and deciding appointment modality.

  • Chat bot that interacts with patient to provide appointment reminders and gather logistical questions about appointments, for example, preferred language and transportation needs. Gathered information can be communicated back to administrative staff to anticipate appointment needs in advance.

8Practitioner clerical and routine task supportOverarching goal: Decrease the burden of routine tasks, such as documentation.
Example specific tasks:
  • Centralised referral system between PC and specialty care.

  • Automatic transcription/documentation to reduce note taking.

  • Identify outstanding requisitions or tests needing follow-up.

  • Group discussions note that referral centralisation may be more important/appealing, but transcription seems more feasible in the short-term.

9Increased mental healthcare capacity and supportOverarching goal: Support and/or increase the scope of mental healthcare from PC settings.
Example specific tasks:
  • Avatars and digital identity (non-live photos) to increase patient comfort in seeking care.

  • Decision support systems to mitigate ‘anchoring bias’ (move beyond initial information fixation to see/care for the whole person).

  • Linking familial patients.

  • Risk prediction and decision support.

  • System level tools to avoid people falling through the cracks.

  • AI, artificial intelligence; EMR, electronic medical record; PC, primary care.