Artificial intelligence for family medicine research in Canada: current state and future directions: Report of the CFPC AI Working Group.

OBJECTIVE: To understand the current landscape of artificial intelligence (AI) for family medicine (FM) research in Canada, identify how the College of Family Physicians of Canada (CFPC) could support near-term positive progress in this field, and strengthen the community working in this field. COMPOSITION OF THE COMMITTEE: Members of a scientific planning committee provided guidance alongside members of a CFPC staff advisory committee, led by the CFPC-AMS TechForward Fellow and including CFPC, FM, and AI leaders. METHODS: This initiative included 2 projects. First, an environmental scan of published and gray literature on AI for FM produced between 2018 and 2022 was completed. Second, an invitational round table held in April 2022 brought together AI and FM experts and leaders to discuss priorities and to create a strategy for the future. REPORT: The environmental scan identified research related to 5 major domains of application in FM (preventive care and risk profiling, physician decision support, operational efficiencies, patient self-management, and population health). Although there had been little testing or evaluation of AI-based tools in practice settings, progress since previous reviews has been made in engaging stakeholders to identify key considerations about AI for FM and opportunities in the field. The round-table discussions further emphasized barriers to and facilitators of high-quality research; they also indicated that while there is immense potential for AI to benefit FM practice, the current research trajectory needs to change, and greater support is needed to achieve these expected benefits and to avoid harm. CONCLUSION: Ten candidate action items that the CFPC could adopt to support near-term positive progress in the field were identified, some of which an AI working group has begun pursuing. Candidate action items are roughly divided into avenues where the CFPC is well-suited to take a leadership role in tackling priority issues in AI for FM research and specific activities or initiatives the CFPC could complete. Strong FM leadership is needed to advance AI research that will contribute to positive transformation in FM.

Primary Care Research: Looking Back and Moving Forward With Reflections on NAPCRG's First 50 Years.

NAPCRG celebrated 50 years of leadership and service at its 2022 meeting. A varied team of primary care investigators, clinicians, learners, patients, and community members reflected on the organization's past, present, and future. Started in 1972 by a small group of general practice researchers in the United States, Canada, and the United Kingdom, NAPCRG has evolved into an international, interprofessional, interdisciplinary, and intergenerational group devoted to improving health and health care through primary care research. NAPCRG provides a nurturing home to researchers and teams working in partnership with individuals, families, and communities. The organization builds upon enduring values to create partnerships, advance research methods, and nurture a community of contributors. NAPCRG has made foundational contributions, including identifying the need for primary care research to inform primary care practice, practice-based research networks, qualitative and mixed-methods research, community-based participatory research, patient safety, practice transformation, and partnerships with patients and communities. Landmark documents have helped define classification systems for primary care, responsible research with communities, the central role of primary care in health care systems, opportunities to revitalize generalist practice, and shared strategies to build the future of family medicine. The future of health and health care depends upon strengthening primary care and primary care research with stronger support, infrastructure, training, and workforce. New technologies offer opportunities to advance research, enhance care, and improve outcomes. Stronger partnerships can empower primary care research with patients and communities and increase commitments to diversity and quality care for all. NAPCRG offers a home for all partners in this work.

Perceptions of Artificial Intelligence Use in Primary Care: A Qualitative Study with Providers and Staff of Ontario Community Health Centres.

PURPOSE: To understand staff and health care providers' views on potential use of artificial intelligence (AI)-driven tools to help care for patients within a primary care setting. METHODS: We conducted a qualitative descriptive study using individual semistructured interviews. As part of province-wide Learning Health Organization, Community Health Centres (CHCs) are a community-governed, team-based delivery model providing primary care for people who experience marginalization in Ontario, Canada. CHC health care providers and staff were invited to participate. Interviews were audio-recorded and transcribed verbatim. We performed a thematic analysis using a team approach. RESULTS: We interviewed 27 participants across 6 CHCs. Participants lacked in-depth knowledge about AI. Trust was essential to acceptance of AI; people need to be receptive to using AI and feel confident that the information is accurate. We identified internal influences of AI acceptance, including ease of use and complementing clinical judgment rather than replacing it. External influences included privacy, liability, and financial considerations. Participants felt AI could improve patient care and help prevent burnout for providers; however, there were concerns about the impact on the patient-provider relationship. CONCLUSIONS: The information gained in this study can be used for future research, development, and integration of AI technology.

An introduction to machine learning for classification and prediction.

Classification and prediction tasks are common in health research. With the increasing availability of vast health data repositories (e.g. electronic medical record databases) and advances in computing power, traditional statistical approaches are being augmented or replaced with machine learning (ML) approaches to classify and predict health outcomes. ML describes the automated process of identifying ("learning") patterns in data to perform tasks. Developing an ML model includes selecting between many ML models (e.g. decision trees, support vector machines, neural networks); model specifications such as hyperparameter tuning; and evaluation of model performance. This process is conducted repeatedly to find the model and corresponding specifications that optimize some measure of model performance. ML models can make more accurate classifications and predictions than their statistical counterparts and confer greater flexibility when modelling unstructured data or interactions between covariates; however, many ML models require larger sample sizes to achieve good classification or predictive performance and have been criticized as "black box" for their poor transparency and interpretability. ML holds potential in family medicine for risk profiling of patients' disease risk and clinical decision support to present additional information at times of uncertainty or high demand. In the future, ML approaches are positioned to become commonplace in family medicine. As such, it is important to understand the objectives that can be addressed using ML approaches and the associated techniques and limitations. This article provides a brief introduction into the use of ML approaches for classification and prediction tasks in family medicine.

Competencies for the Use of Artificial Intelligence in Primary Care.

The artificial intelligence (AI) revolution has arrived for the health care sector and is finally penetrating the far-reaching but perpetually underfinanced primary care platform. While AI has the potential to facilitate the achievement of the Quintuple Aim (better patient outcomes, population health, and health equity at lower costs while preserving clinician well-being), inattention to primary care training in the use of AI-based tools risks the opposite effects, imposing harm and exacerbating inequalities. The impact of AI-based tools on these aims will depend heavily on the decisions and skills of primary care clinicians; therefore, appropriate medical education and training will be crucial to maximize potential benefits and minimize harms. To facilitate this training, we propose 6 domains of competency for the effective deployment of AI-based tools in primary care: (1) foundational knowledge (what is this tool?), (2) critical appraisal (should I use this tool?), (3) medical decision making (when should I use this tool?), (4) technical use (how do I use this tool?), (5) patient communication (how should I communicate with patients regarding the use of this tool?), and (6) awareness of unintended consequences (what are the "side effects" of this tool?). Integrating these competencies will not be straightforward because of the breadth of knowledge already incorporated into family medicine training and the constantly changing technological landscape. Nonetheless, even incremental increases in AI-relevant training may be beneficial, and the sooner these challenges are tackled, the sooner the primary care workforce and those served by it will begin to reap the benefits.

Is primary health care ready for artificial intelligence? What do primary health care stakeholders say?

BACKGROUND: Effective deployment of AI tools in primary health care requires the engagement of practitioners in the development and testing of these tools, and a match between the resulting AI tools and clinical/system needs in primary health care. To set the stage for these developments, we must gain a more in-depth understanding of the views of practitioners and decision-makers about the use of AI in primary health care. The objective of this study was to identify key issues regarding the use of AI tools in primary health care by exploring the views of primary health care and digital health stakeholders. METHODS: This study utilized a descriptive qualitative approach, including thematic data analysis. Fourteen in-depth interviews were conducted with primary health care and digital health stakeholders in Ontario. NVivo software was utilized in the coding of the interviews. RESULTS: Five main interconnected themes emerged: (1) Mismatch Between Envisioned Uses and Current Reality-denoting the importance of potential applications of AI in primary health care practice, with a recognition of the current reality characterized by a lack of available tools; (2) Mechanics of AI Don't Matter: Just Another Tool in the Toolbox- reflecting an interest in what value AI tools could bring to practice, rather than concern with the mechanics of the AI tools themselves; (3) AI in Practice: A Double-Edged Sword-the possible benefits of AI use in primary health care contrasted with fundamental concern about the possible threats posed by AI in terms of clinical skills and capacity, mistakes, and loss of control; (4) The Non-Starters: A Guarded Stance Regarding AI Adoption in Primary Health Care-broader concerns centred on the ethical, legal, and social implications of AI use in primary health care; and (5) Necessary Elements: Facilitators of AI in Primary Health Care-elements required to support the uptake of AI tools, including co-creation, availability and use of high quality data, and the need for evaluation. CONCLUSION: The use of AI in primary health care may have a positive impact, but many factors need to be considered regarding its implementation. This study may help to inform the development and deployment of AI tools in primary health care.

Connecting artificial intelligence and primary care challenges: findings from a multi stakeholder collaborative consultation.

Despite widespread advancements in and envisioned uses for artificial intelligence (AI), few examples of successfully implemented AI innovations exist in primary care (PC) settings. OBJECTIVES: To identify priority areas for AI and PC in Ontario, Canada. METHODS: A collaborative consultation event engaged multiple stakeholders in a nominal group technique process to generate, discuss and rank ideas for how AI can support Ontario PC. RESULTS: The consultation process produced nine ranked priorities: (1) preventative care and risk profiling, (2) patient self-management of condition(s), (3) management and synthesis of information, (4) improved communication between PC and AI stakeholders, (5) data sharing and interoperability, (6-tie) clinical decision support, (6-tie) administrative staff support, (8) practitioner clerical and routine task support and (9) increased mental healthcare capacity and support. Themes emerging from small group discussions about barriers, implementation issues and resources needed to support the priorities included: equity and the digital divide; system capacity and culture; data availability and quality; legal and ethical issues; user-centred design; patient-centredness; and proper evaluation of AI-driven tool implementation. DISCUSSION: Findings provide guidance for future work on AI and PC. There are immediate opportunities to use existing resources to develop and test AI for priority areas at the patient, provider and system level. For larger scale, sustainable innovations, there is a need for longer-term projects that lay foundations around data and interdisciplinary work. CONCLUSION: Study findings can be used to inform future research and development of AI for PC, and to guide resource planning and allocation.

Describing a complex primary health care population to support future decision support initiatives.

Introduction: Developing decision support tools using data from a health care organization, to support care within that organization, is a promising paradigm to improve care delivery and population health. Descriptive epidemiology may be a valuable supplement to stakeholder input towards selection of potential initiatives and to inform methodological decisions throughout tool development. We additionally propose that to properly characterize complex populations in large-scale descriptive studies, both simple statistical and machine learning techniques can be useful. Objective: To describe sociodemographic, clinical, and health care use characteristics of primary care clients served by the Alliance for Healthier Communities, which provides team-based primary health care through Community Health Centres (CHCs) across Ontario, Canada. Methods: We used electronic health record data from adult ongoing primary care clients served by CHCs in 2009-2019. We performed traditional table-based summaries for each characteristic; and applied three unsupervised learning techniques to explore patterns of common condition co-occurrence, care provider teams, and care frequency. Results: There were 221,047 eligible clients. Sociodemographics: We described 13 characteristics, stratified by CHC type and client multimorbidity status. Clinical characteristics: Eleven-year prevalence of 24 investigated conditions ranged from 1% (Hepatitis C) to 63% (chronic musculoskeletal problem) with non-uniform risk across the care history; multimorbidity was common (81%) with variable co-occurrence patterns. Health care use characteristics: Most care was provided by physician and nursing providers, with heterogeneous combinations of other provider types. A subset of clients had many issues addressed within single-visits and there was within- and between-client variability in care frequency. In addition to substantive findings, we discuss methodological considerations for future decision support initiatives. Conclusions: We demonstrated the use of methods from statistics and machine learning, applied with an epidemiological lens, to provide an overview of a complex primary care population and lay a foundation for stakeholder engagement and decision support tool development.

A Mobile App to Identify Lifestyle Indicators Related to Undergraduate Mental Health (Smart Healthy Campus): Observational App-Based Ecological Momentary Assessment.

BACKGROUND: Undergraduate studies are challenging, and mental health issues can frequently occur in undergraduate students, straining campus resources that are already in demand for somatic problems. Cost-effective measures with ubiquitous devices, such as smartphones, offer the potential to deliver targeted interventions to monitor and affect lifestyle, which may result in improvements to student mental health. However, the avenues by which this can be done are not particularly well understood, especially in the Canadian context. OBJECTIVE: The aim of this study is to deploy an initial version of the Smart Healthy Campus app at Western University, Canada, and to analyze corresponding data for associations between psychosocial factors (measured by a questionnaire) and behaviors associated with lifestyle (measured by smartphone sensors). METHODS: This preliminary study was conducted as an observational app-based ecological momentary assessment. Undergraduate students were recruited over email, and sampling using a custom 7-item questionnaire occurred on a weekly basis. RESULTS: First, the 7-item Smart Healthy Campus questionnaire, derived from fully validated questionnaires-such as the Brief Resilience Scale; General Anxiety Disorder-7; and Depression, Anxiety, and Stress Scale-21-was shown to significantly correlate with the mental health domains of these validated questionnaires, illustrating that it is a viable tool for a momentary assessment of an overview of undergraduate mental health. Second, data collected through the app were analyzed. There were 312 weekly responses and 813 sensor samples from 139 participants from March 2019 to March 2020; data collection concluded when COVID-19 was declared a pandemic. Demographic information was not collected in this preliminary study because of technical limitations. Approximately 69.8% (97/139) of participants only completed one survey, possibly because of the absence of any incentive. Given the limited amount of data, analysis was not conducted with respect to time, so all data were analyzed as a single collection. On the basis of mean rank, students showing more positive mental health through higher questionnaire scores tended to spend more time completing questionnaires, showed more signs of physical activity based on pedometers, and had their devices running less and plugged in charging less when sampled. In addition, based on mean rank, students on campus tended to report more positive mental health through higher questionnaire scores compared with those who were sampled off campus. Some data from students found in or near residences were also briefly examined. CONCLUSIONS: Given these limited data, participants tended to report a more positive overview of mental health when on campus and when showing signs of higher levels of physical activity. These early findings suggest that device sensors related to physical activity and location are useful for monitoring undergraduate students and designing interventions. However, much more sensor data are needed going forward, especially given the sweeping changes in undergraduate studies due to COVID-19.

Development of a prognostic prediction model to estimate the risk of multiple chronic diseases: constructing a copula-based model using Canadian primary care electronic medical record data.

INTRODUCTION: The ability to estimate risk of multimorbidity will provide valuable information to patients and primary care practitioners in their preventative efforts. Current methods for prognostic prediction modelling are insufficient for the estimation of risk for multiple outcomes, as they do not properly capture the dependence that exists between outcomes. OBJECTIVES: We developed a multivariate prognostic prediction model for the 5-year risk of diabetes, hypertension, and osteoarthritis that quantifies and accounts for the dependence between each disease using a copula-based model. METHODS: We used data from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN) from 2009 onwards, a collection of electronic medical records submitted by participating primary care practitioners across Canada. We identified patients 18 years and older without all three outcome diseases and observed any incident diabetes, osteoarthritis, or hypertension within 5-years, resulting in a large retrospective cohort for model development and internal validation (n=425,228). First, we quantified the dependence between outcomes using unadjusted and adjusted Ø coefficients. We then estimated a copula-based model to quantify the non-linear dependence between outcomes that can be used to derive risk estimates for each outcome, accounting for the observed dependence. Copula-based models are defined by univariate models for each outcome and a dependence function, specified by the parameter θ. Logistic regression was used for the univariate models and the Frank copula was selected as the dependence function. RESULTS: All outcome pairs demonstrated statistically significant dependence that was reduced after adjusting for covariates. The copula-based model yielded statistically significant θ parameters in agreement with the adjusted and unadjusted Ø coefficients. Our copula-based model can effectively be used to estimate trivariate probabilities. DISCUSSION: Quantitative estimates of multimorbidity risk inform discussions between patients and their primary care practitioners around prevention in an effort to reduce the incidence of multimorbidity.

Primary Care Informatics Response to Covid-19 Pandemic: Adaptation, Progress, and Lessons from Four Countries with High ICT Development.

OBJECTIVE: Internationally, primary care practice had to transform in response to the COVID pandemic. Informatics issues included access, privacy, and security, as well as patient concerns of equity, safety, quality, and trust. This paper describes progress and lessons learned. METHODS: IMIA Primary Care Informatics Working Group members from Australia, Canada, United Kingdom and United States developed a standardised template for collection of information. The template guided a rapid literature review. We also included experiential learning from primary care and public health perspectives. RESULTS: All countries responded rapidly. Common themes included rapid reductions then transformation to virtual visits, pausing of non-COVID related informatics projects, all against a background of non-standardized digital development and disparate territory or state regulations and guidance. Common barriers in these four and in less-resourced countries included disparities in internet access and availability including bandwidth limitations when internet access was available, initial lack of coding standards, and fears of primary care clinicians that patients were delaying care despite the availability of televisits. CONCLUSIONS: Primary care clinicians were able to respond to the COVID crisis through telehealth and electronic record enabled change. However, the lack of coordinated national strategies and regulation, assurance of financial viability, and working in silos remained limitations. The potential for primary care informatics to transform current practice was highlighted. More research is needed to confirm preliminary observations and trends noted.

Cognition and motor function: The gait and cognition pooled index.

BACKGROUND: There is a need for outcome measures with improved responsiveness to changes in pre-dementia populations. Both cognitive and motor function play important roles in neurodegeneration; motor function decline is detectable at early stages of cognitive decline. This proof of principle study used a Pooled Index approach to evaluate improved responsiveness of the predominant outcome measure (ADAS-Cog: Alzheimer's Disease Assessment Scale-Cognitive Subscale) when assessment of motor function is added. METHODS: Candidate Pooled Index variables were selected based on theoretical importance and pairwise correlation coefficients. Kruskal-Wallis and Mann-Whitney U tests assessed baseline discrimination. Standardized response means assessed responsiveness to longitudinal change. RESULTS: Final selected variables for the Pooled Index include gait velocity, dual-task cost of gait velocity, and an ADAS-Cog-Proxy (statistical approximation of the ADAS-Cog using similar cognitive tests). The Pooled Index and ADAS-Cog-Proxy scores had similar ability to discriminate between pre-dementia syndromes. The Pooled Index demonstrated trends of similar or greater responsiveness to longitudinal decline than ADAS-Cog-Proxy scores. CONCLUSION: Adding motor function assessments to the ADAS-Cog may improve responsiveness in pre-dementia populations.

Artificial Intelligence and Primary Care Research: A Scoping Review.

PURPOSE: Rapid increases in technology and data motivate the application of artificial intelligence (AI) to primary care, but no comprehensive review exists to guide these efforts. Our objective was to assess the nature and extent of the body of research on AI for primary care. METHODS: We performed a scoping review, searching 11 published or gray literature databases with terms pertaining to AI (eg, machine learning, bayes* network) and primary care (eg, general pract*, nurse). We performed title and abstract and then full-text screening using Covidence. Studies had to involve research, include both AI and primary care, and be published in Eng-lish. We extracted data and summarized studies by 7 attributes: purpose(s); author appointment(s); primary care function(s); intended end user(s); health condition(s); geographic location of data source; and AI subfield(s). RESULTS: Of 5,515 unique documents, 405 met eligibility criteria. The body of research focused on developing or modifying AI methods (66.7%) to support physician diagnostic or treatment recommendations (36.5% and 13.8%), for chronic conditions, using data from higher-income countries. Few studies (14.1%) had even a single author with a primary care appointment. The predominant AI subfields were supervised machine learning (40.0%) and expert systems (22.2%). CONCLUSIONS: Research on AI for primary care is at an early stage of maturity. For the field to progress, more interdisciplinary research teams with end-user engagement and evaluation studies are needed.

The Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog): Modifications and Responsiveness in Pre-Dementia Populations. A Narrative Review.

The Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) was developed in the 1980s to assess the level of cognitive dysfunction in Alzheimer's disease. Advancements in the research field have shifted focus toward pre-dementia populations, and use of the ADAS-Cog has extended into these pre-dementia studies despite concerns about its ability to detect important changes at these milder stages of disease progression. If the ADAS-Cog cannot detect important changes, our understanding of pre-dementia disease progression may be compromised and trials may incorrectly conclude that a novel treatment approach is not beneficial. The purpose of this review was to assess the performance of the ADAS-Cog in pre-dementia populations, and to review all modifications that have been made to the ADAS-Cog to improve its measurement performance in dementia or pre-dementia populations. The contents of this review are based on bibliographic searches of electronic databases to locate all studies using the ADAS-Cog in pre-dementia samples or subsamples, and to locate all modified versions. Citations from relevant articles were also consulted. Overall, our results suggest the original ADAS-Cog is not an optimal outcome measure for pre-dementia studies; however, given the prominence of the ADAS-Cog, care must be taken when considering the use of alternative outcome measures. Thirty-one modified versions of the ADAS-Cog were found. Modification approaches that appear most beneficial include altering scoring methodology or adding tests of memory, executive function, and/or daily functioning. Although modifications improve the performance of the ADAS-Cog, this is at the cost of introducing heterogeneity that may limit between-study comparison.