https://orcid.org/0009-0001-3929-2560
Figures
Abstract
Introduction
Large language models’ (LLMs’) rapid evolution and intersection with diverse groups and institutions require up-to-date policies, practices, and behaviors to ensure safe and effective implementation. Because physicians play a central role in care provision and face the dual mandate of embracing innovations and safeguarding patient welfare, understanding physicians’ views on LLMs can elucidate the complex interplay of technical, ethical, and professional considerations influencing LLM adoption.
Methods and analysis
This qualitative study aims to employ a descriptive qualitative design to explore how primary care physicians perceive the adoption of LLMs in the context of their clinical practice. We plan to use semi-structured interviews with purposively sampled primary care physicians from British Columbia, Canada. The data collection will draw on the technology adoption behavior framework, a novel model that integrates the most advanced theories of technological uptake. We expect to use thematic analysis drawing on deductive and inductive approaches to describe physicians’ perceptions. The multidisciplinary research team will prepare and conduct reflexive memos and discussions to ensure nuanced interpretations.
Dissemination
We aim to disseminate the findings through peer-reviewed journals, professional organizations, and policymaker briefings to support the development of policies, practices, and behaviors that support safe and effective LLM integration into health care.
Strengths and limitations
- The study may provide timely input into relevant policies, practices, and behaviors for policymakers, health professionals, and patients around the use of large language models in health care services.
- This study uses the technology adoption behavior framework to guide the design of data collection and analysis.
- The semi-structured interview may reveal the interviewees’ internal views but limits the variety of insights gleaned.
Citation: Pokharel BB, Hsu M, Hedden L, Nimmon L, Bloom DE, Tsuei SH (2026) Understanding primary care physicians’ perceptions of large language model adoption in clinical practice: A qualitative research protocol leveraging the technology adoption behavior framework. PLoS One 21(8): e0355491. https://doi.org/10.1371/journal.pone.0355491
Editor: Noah Hammarlund, University of Florida, UNITED STATES OF AMERICA
Received: July 18, 2025; Accepted: July 22, 2026; Published: August 7, 2026
Copyright: © 2026 Pokharel et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: No datasets were generated or analysed during the current study. Deidentified data may be made available upon study completion.
Funding: Dr. Sian Tsuei received the Research Trainee Award (RT-2023-3307) from Michael Smith Health Research BC (https://healthresearchbc.ca/). The funder did not play any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Dr. Sian Tsuei serves as a member on the Artificial Intelligence Advisory Council for the College of Canadian Family Physicians. This does not alter our adherence to PLOS ONE’s policies on sharing data and materials.
1 Introduction
Developing appropriate policies, practices, and behaviors for new technologies is crucial for protecting and improving health care system performance. However, identifying the types and magnitude of relevant consequences can be challenging in the early stages of technological development and implementation. For example, electronic medical records were meant to provide high-quality documentation but were also associated with disrupted clinical workflow and increased provider burnout [1,2].
Designing policies that mitigate the harms and maximize the benefits from modern large language models (LLMs) can be similarly challenging. LLMs are a class of artificial intelligence systems trained on vast text datasets to understand and generate human-like language. In healthcare, their ability to process unstructured clinical information positions them as emerging tools for documentation, decision support, patient communication, and administrative workflows. Their wide-ranging capacity can help support clerical tasks, such as summarizing patient visits and drafting letters to patients [3]. They can also support diagnostic and therapeutic tasks, such as taking a patient’s history, interpreting certain investigations, recommending potential diagnoses, and suggesting investigations and therapies. In some cases, LLMs’ diagnostic performance may at times rival that of some human specialists [4–7]. These studies suggest promising performance in many contexts; however, this research is preliminary and subject to important limitations given the novelty of the field. Additionally, LLMs can also perpetuate biases, generate false information, and potentially displace humans from their jobs [8,9]. LLMs’ rapid development and intersection with multiple spheres of behavioral norms further complicate their assessment [10–13].
Studying how physicians consider the diverse benefits and risks can be particularly enlightening. First, their roles cross both leadership and frontline service provision. Second, they have to weigh the need to introduce advanced technologies for patient benefit against their potential harms [14,15]. Lastly, physicians using LLMs can improve their service quality and efficiency, but excessive reliance on LLMs can undermine physicians’ skills and long-term importance in the health care service market.
To reveal deeper insights into the complex multidimensional dynamics [16], this study thus aims to qualitatively examine how physicians perceive the uptake of LLMs. Overall, this study provides two contributions. First, this informs literature on how professions absorb new technologies amid evolving practical, legal, ethical, and existential considerations. Few technologies have had such wide implications. Second, clarifying how physicians balance the various considerations can grant policymakers the necessary insight to consider whether and how to adjust policies for optimal health system performance. Although a small but growing body of work has begun to examine clinicians’ perceptions of LLMs [17–19], prior studies are largely atheoretical or lightly theory informed and rarely focus on primary care. To our knowledge, no study in primary care has used an explicit framework to guide both data collection and analysis.
We focus on family physicians in Canada for two reasons. First, strong primary care can improve the quality, access, efficiency, and mortality [20–22]. Family physicians constitute the largest share of the physician workforce in this country [23], often gatekeeping patients’ access to other health care services. However, the estimated number of patients without a family doctor rose from 4.5 million in 2019 to about 6.5 million in 2023 [24], driving up public discontent and missed opportunities to improve health [25,26]. Understanding the motivations and reservations facing this group of health care service providers may thus be particularly helpful in unlocking the relevant support for downstream health benefits. Second, many family medicine diagnoses rely on verbal information and cognitive considerations, which coincide with LLMs’ support capacity. This may be particularly helpful in illustrating the physicians’ intricate affinity and tensions with such a technology.
2 Methodology
2.1 Overall study design
We draw on the qualitative descriptive approach to capture the complexity of how physicians perceive the use of LLM tools [27,28]. This approach offers a “summary of events in the everyday terms of those events” with flexible sampling, data collection, analysis, and representation [29]. We will employ semi-structured interviews, which provides an in-depth understanding of participant perspectives [30]. We will use purposive sampling to ensure representation from those who use and avoid LLMs. The data collection and analysis will leverage the technological adoption behavior (TAB) framework. This is an expanded theoretical framework on technological uptake we developed to unite multiple evidence-based theoretical streams (Fig 1). The final analysis will conduct thematic analysis with deductive and inductive coding.
2.2 Theoretical framework
We followed the guidance from Varpio et al. in creating “a logically developed and connected set of concepts and premises—developed from one or more theories—that …scaffold a study” [31]. Fig 1 shows the TAB framework, and Appendix A (S1 Appendix) describes the historical evolution of the various frameworks underlying its development. The TAB framework builds on two established traditions in research on technology uptake: one emphasizing the role of user perceptions and the other emphasizing the fit between technology, tasks, and users. To make these links transparent, we organize this section into three parts. Section 2.2.1 outlines the importance of perceptions, drawing from models such as the Technology Acceptance Model (TAM) [32]. Section 2.2.2 describes the technology–environment fit tradition, including both early formulations [33,34] and more recent refinements [35]. Section 2.2.3 explains how TAB integrates these perspectives into a unified framework. Although some foundational contributions are older, they remain the basis for contemporary refinements, and we complement them with recent extensions [35] to ensure the framework reflects current scholarship.
2.2.1 Importance of perception.
The former camp emphasizes that the motivation to incorporate a tool into users’ work stems from their perceptions. The two key perceptions are the perceived ease of use (i.e., the degree of expected difficulty in taking on a new technology) and perceived usefulness (i.e., the degree a technology is expected to improve users’ actions). Multiple individual, task, and social characteristics in turn affect these two factors.
The users’ perceived capacity to operate the technology effectively (i.e., self-efficacy and anxiety related to the technology) primarily influence the perceived ease of use. If the users perceive the technology to be playful, then learning to work with the technology becomes a spontaneous exploration. Furthermore, supportive resources for using the technology may also lower the perceived challenge [32].
In contrast, both technical and social considerations influence perceived usefulness. The technical considerations largely capture whether the technology can support the tasks’ needs. Higher technology’s relevance and expected output quality for the task may lead to higher apparent usefulness. The social considerations suggest that as more people use a technology, the normative pressure to use it also increases, potentially translating to a better image for the users [32].
2.2.2 Importance of technology’s fit with the environment.
The latter camp emphasizes that research on adoption should instead focus on the technologies’ fit with the environment. The most advanced form includes the technology’s fit with the task and the user [33–35]. The task characteristics include the difficulty, routineness, complexity, and variety of tasks. For instance, as the tasks’ difficulty, complexity, and variety increase, the technologies may need to be more sophisticated and adaptable to be useful. This may then improve the chances of incorporating the technology into the workflow. The user’s characteristics includes the person’s education, training, experience, and motivation. For example, less educated users may require an adjustment of the technology’s user interface for better uptake.
2.2.3 Unification of the two camps.
The early scholars who studied technology-environment fit recognized that fit alone was insufficient; the fit needs to be recognized as useful to affect technological adoption [34]. However, the theoretical expansion of this vein has subsequently focused on the interdependence of the technology, user, and tasks. To capture the potential factors affecting LLM uptake more comprehensively, we combine the two veins of theoretical development together into the TAB framework.
The TAB framework follows the former camp’s tradition by suggesting that both perceived ease of use and perceived usefulness drive technological adoption. It also incorporates the latter camp’s emphasis on the technology’s fit with the user and task. We follow Goodhue and Thompson’s conceptualization to integrate the two camps by suggesting that the technology-environment fit feeds into the perceived usefulness. We add in additional individual and social considerations from other empirical literature.
2.3 Study setting, sample, and recruitment
We focus on the Canadian province of British Columbia. Despite the government’s encouragement for health care providers in this region to use innovative health care technologies, physicians’ uptake of digital technologies may be delayed and variable [36–38].
We will purposively sample practicing physicians with varying levels of LLM usage, ranging from frequent, occasional, and nearly no use (i.e., the physician has used LLMs 50 + , 5–50, and <5 times in their lifetime, respectively, in the context of health care services). The sampling aims to reach sufficiency, achieved when there is enough evidence and depth of conceptual understanding to meaningfully answer the research question [39]. We will also ensure variability in terms of the years of practice, which may affect participants’ perceptions of digital technologies. The diverse perspectives may help reveal broadly applicable insights and support analytical generalizability [40]. We will exclude family physicians who work exclusively in non-clinical roles (e.g., academic, research, or administrative). We aim to recruit participants from existing professional networks and local physician organizations such as Doctors of British Columbia, BC Family Doctors, and British Columbia College of Family Physicians. Recruitment will occur through organizational listserv emails, posted invitations on member websites, and outreach at educational events. In addition, physicians known to the research team may be approached via personal contacts; snowball sampling will be used when participants recommend colleagues.
2.4 Semi-structured interview and question guide
The semi-structured approach of the interview allows for depth and flexibility as we explore the central phenomenon of interest. We designed the interview guide to start with open-ended questions around key constructs of the TAB framework (Table 1). The current version is designed to be more comprehensive, and the first two interviews will serve as pilots to ensure smooth interview flow. The guide includes prompts to allow further exploration of key predefined concepts in the TAB framework. Because different experiences and perspectives in the application of LLMs may drive varying perceptions of LLM usage, interviewers will briefly clarify the contexts under which the participants have interacted with LLMs. The interviewers will do so at the outset of each interview. The questions will ask the participant to clarify the examples of LLMs used or considered, intended purpose of such LLMs (e.g., prevention, diagnostics, or therapy), target user groups, potential hardware required, and relevant settings of use. Additionally, we will prompt for disconfirming evidence to enrich the data collection process [41], actively probing for accounts that contradict or challenge emerging patterns or assumptions.
One primary researcher will conduct the interviews using in-person, telephone, or video conferencing approaches. All the interviews will be recorded with a physical recording device or video conferencing software and subsequently transcribed verbatim and deidentified.
2.5 Data analysis
The diverse background of our team members helps ensure that the data interpretation captures diverse perspectives and interpretations. They include a medical trainee, a practicing family physician, two medical educators, an economist, and three health system researchers. The team will write reflexive memos and engage in reflexive dialogues on how their identity may have influenced their data interpretation. The senior author will be attuned to all team members perspectives, carefully accounting for all contributions [42,43].
We aim to use both deductive and inductive coding approaches to flexibly understand and capture the content from the participants. Table 2 shows the series of deductive codes based on the TAB framework. The inductive coding refines the existing codes and / or develop new ones based on the empirically collected data [44]. We will rely on this approach when the data extend beyond the existing codes developed based on the TAB framework. Additionally, we will write analytical memos throughout the coding process to document how the empirical findings may extend the TAB framework and refine relevant codes appropriately. Primary coding will be conducted by a new member of the research team after the person is trained in the TAB framework and qualitative coding. The person will be in charge of developing additional codes, and every code addition, removal, and refinement will be discussed with the senior author. The adjustments will be discussed with the rest of the team every month until the codebook is finalized. The team will aim to balance the TAB framework with the team’s diverse perspectives to focus analysis [42].
We will iteratively discuss within the group to distill the meaning of the findings, and as the relationships and dynamics of the codes emerge, this will lead us towards broader themes [45]. We aim to generate clear and concise definitions for each theme and review and refine the identified themes to ensure accurate representation. We will use NVivo to conduct all analysis.
2.6 Patient and public engagement
Although the team did not consult patients or the public in developing the study, the work is designed to inform the ongoing need for meaningful policies that can protect the safety and quality of care when introducing AI tools [46].
2.7 Ethical considerations
We have obtained approval from the harmonized ethics review board (UBC Behavioral Research Ethics Board; H25-01658). All potential participants will first receive recruitment materials approved by the review board, and those who agree to participate will be given informed consent forms also approved by the board. At the start of the interview, we will inform the participants of the study’s aims, the participants’ rights (including the right to withdraw at any point without repercussions), sharing of deidentified data, and potential resources for post-interview debriefing. After the transcriptions, the original audio files will be destroyed and the transcripts will be deidentified. The team will work only with the deidentified version of the data.
2.8. Study status
We have only completed the study design, and we have not yet begun participant recruitment or data collection. We expect participant recruitment and data collection to be completed within six months (anticipated timeline from August 1, 2025, to January 31, 2026). We expect that once the data collection begins, we can complete the study and produce results within a year.
Supporting information
S1 Appendix. Review of technology uptake frameworks leading up to the technological adoption behavior framework.
https://doi.org/10.1371/journal.pone.0355491.s001
(DOCX)
S1 Fig. Illustration highlighting key differences between (A) the theory of reasoned action (TRA) and (B) the theory of planned behavior (TPB).
TRA assumes that the intention to complete a behavior is sufficient for actual behavior. TPB adds perceived behavioral control as another variable contributing to behavioral intention and as a variable that independently contributes to actual behavior.
https://doi.org/10.1371/journal.pone.0355491.s002
(TIF)
S2 Fig. The technology acceptance model (TAM) combined with TPB.
The contribution of TAM is colored in blue. Both these variables—perceived usefulness and perceived ease of use—specifically refer to features of a technology that ultimately contribute to users’ uptake of the technology.
https://doi.org/10.1371/journal.pone.0355491.s003
(TIF)
S3 Fig. Simplified TAM3 model that combines TAM with TRA/TPB to develop a technological use behavior prediction model.
*External variables include perceived ease of use, subjective norm, image, and result demonstrability. **External variables include computer self-efficacy, computer anxiety, and computer playfulness, and perceptions of external control.
https://doi.org/10.1371/journal.pone.0355491.s004
(TIF)
S4 Fig. Goodhue and Thompson’s task technology fit framework of technological adoption relative to TAM3.
The interplay between task/technology and individual/technology is key for appropriate task technology fit, which is a predictor of technological performance. User evaluation is sufficient in measuring technological performance. This figure also demonstrates that technological performance is an external variable that influences the perceived usefulness variable in TAM, ultimately influencing technological use behavior.
https://doi.org/10.1371/journal.pone.0355491.s005
(TIF)
S5 Fig. Fit among individuals, task, and technology (FITT) framework of technological adoption.
FITT emphasizes the connection between task and individual as another key benchmark for appropriate technological fit while the TTF framework does not.
https://doi.org/10.1371/journal.pone.0355491.s006
(TIF)
S6 Fig. Ruyobeza et al.’s fit among individuals, skills, tasks, and technology framework (2023).
https://doi.org/10.1371/journal.pone.0355491.s007
(TIF)
Acknowledgments
The development of the TAB framework and protocol benefited from discussions with Cypress Knudson.
References
- 1. Howe JL, Adams KT, Hettinger AZ, Ratwani RM. Electronic Health Record Usability Issues and Potential Contribution to Patient Harm. JAMA. 2018;319(12):1276–8. pmid:29584833
- 2. Li C, Parpia C, Sriharan A, Keefe DT. Electronic medical record-related burnout in healthcare providers: a scoping review of outcomes and interventions. BMJ Open. 2022;12(8):e060865. pmid:35985785
- 3. Denecke K, May R, LLMHealthGroup, Rivera Romero O. Potential of Large Language Models in Health Care: Delphi Study. J Med Internet Res. 2024;26:e52399. pmid:38739445
- 4. Adams LC, Truhn D, Busch F, Kader A, Niehues SM, Makowski MR, et al. Leveraging GPT-4 for Post Hoc Transformation of Free-text Radiology Reports into Structured Reporting: A Multilingual Feasibility Study. Radiology. 2023;307(4):e230725. pmid:37014240
- 5. Hirosawa T, Harada Y, Mizuta K, Sakamoto T, Tokumasu K, Shimizu T. Evaluating ChatGPT-4’s Accuracy in Identifying Final Diagnoses Within Differential Diagnoses Compared With Those of Physicians: Experimental Study for Diagnostic Cases. JMIR Form Res. 2024;8:e59267. pmid:38924784
- 6. Savage T, Nayak A, Gallo R, Rangan E, Chen JH. Diagnostic reasoning prompts reveal the potential for large language model interpretability in medicine. NPJ Digit Med. 2024;7(1):20. pmid:38267608
- 7. Shah A, Wahood S, Guermazi D, Brem CE, Saliba E. Skin and Syntax: Large Language Models in Dermatopathology. Dermatopathology (Basel). 2024;11(1):101–11. pmid:38390851
- 8. Doraiswamy PM, Blease C, Bodner K. Artificial intelligence and the future of psychiatry: Insights from a global physician survey. Artificial Intelligence in Medicine. 2020;102:101753.
- 9. Mu Y, He D. The Potential Applications and Challenges of ChatGPT in the Medical Field. Int J Gen Med. 2024;17:817–26. pmid:38476626
- 10. Blumenthal D, Patel B. The Regulation of Clinical Artificial Intelligence. NEJM AI. 2024;1(8).
- 11. UK Government. The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023. 2023. https://go.nature.com/49BFvBG
- 12. Wang Y, Li N, Chen L, Wu M, Meng S, Dai Z, et al. Guidelines, Consensus Statements, and Standards for the Use of Artificial Intelligence in Medicine: Systematic Review. J Med Internet Res. 2023;25:e46089. pmid:37991819
- 13. Yang D. AI in Health Care: 7 Principles of Responsible Use. 2024. https://about.kaiserpermanente.org/news/ai-in-health-care-7-principles-of-responsible-use
- 14.
Robertson GB, Picard EI. Legal Liability of Doctors and Hospitals in Canada. 5th ed. Toronto, Ontario: Thomson Reuters Canada Limited. 2017.
- 15. Rowland SP, Fitzgerald JE, Lungren M, Lee EH, Harned Z, McGregor AH. Digital health technology-specific risks for medical malpractice liability. NPJ Digit Med. 2022;5(1):157. pmid:36261469
- 16. Greenhalgh T, Annandale E, Ashcroft R, Barlow J, Black N, Bleakley A. An open letter to The BMJ editors on qualitative research. BMJ. 2016;2016:i563.
- 17. Mirzaei T, Amini L, Esmaeilzadeh P. Clinician voices on ethics of LLM integration in healthcare: a thematic analysis of ethical concerns and implications. BMC Med Inform Decis Mak. 2024;24(1):250. pmid:39252056
- 18. Spotnitz M, Idnay B, Gordon ER, Shyu R, Zhang G, Liu C, et al. A Survey of Clinicians’ Views of the Utility of Large Language Models. Appl Clin Inform. 2024;15(2):306–12. pmid:38442909
- 19. Sumner J, Wang Y, Tan SY, Chew EHH, Wenjun Yip A. Perspectives and Experiences With Large Language Models in Health Care: Survey Study. J Med Internet Res. 2025;27:e67383. pmid:40310666
- 20. Shi L. The impact of primary care: a focused review. Scientifica (Cairo). 2012;2012:432892. pmid:24278694
- 21. Starfield B. Primary care: an increasingly important contributor to effectiveness, equity, and efficiency of health services. SESPAS report 2012. Gac Sanit. 2012;26 Suppl 1:20–6. pmid:22265645
- 22. Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q. 2005;83(3):457–502. pmid:16202000
- 23. Lemire F, Slade S. Scope of work and the future of family practice. Can Fam Physician. 2022;68(8):626. pmid:35961728
- 24. Duong D, Vogel L. National survey highlights worsening primary care access. CMAJ. 2023;195(16):E592–3. pmid:37094873
- 25. Glauser W. Part-time doctors - reducing hours to reduce burnout. CMAJ. 2018;190(35):E1055–6. pmid:30181155
- 26. Rudoler D, Peterson S, Stock D, Taylor C, Wilton D, Blackie D, et al. Changes over time in patient visits and continuity of care among graduating cohorts of family physicians in 4 Canadian provinces. CMAJ. 2022;194(48):E1639–46. pmid:36511867
- 27. Colorafi KJ, Evans B. Qualitative Descriptive Methods in Health Science Research. HERD. 2016;9(4):16–25. pmid:26791375
- 28. Hunter DJ, McCallum J, Howes D. Defining exploratory-descriptive qualitative (EDQ) research and considering its application to healthcare. GSTF Journal of Nursing and Health Care. 2019;4. http://dl6.globalstf.org/index.php/jnhc/article/view/1975
- 29. Sandelowski M. Whatever happened to qualitative description?. Res Nurs Health. 2000;23(4):334–40.
- 30. Adams WC. Conducting Semi‐Structured Interviews. Handbook of Practical Program Evaluation. Wiley. 2015. 492–505.
- 31. Varpio L, Paradis E, Uijtdehaage S, Young M. The Distinctions Between Theory, Theoretical Framework, and Conceptual Framework. Acad Med. 2020;95(7):989–94. pmid:31725464
- 32. Venkatesh V, Bala H. Technology Acceptance Model 3 and a Research Agenda on Interventions. Decision Sciences. 2008;39(2):273–315.
- 33. Ammenwerth E, Iller C, Mahler C. IT-adoption and the interaction of task, technology and individuals: a fit framework and a case study. BMC Med Inform Decis Mak. 2006;6:3. pmid:16401336
- 34. Goodhue DL, Thompson RL. Task-Technology Fit and Individual Performance. MIS Quarterly. 1995;19(2):213–36.
- 35. Ruyobeza BB, Grobbelaar SSS, Botha A. From FITT to FISTT: The task-skills fit before the introduction of assistive, digital health technologies. Heliyon. 2023;9(6):e16885. pmid:37360076
- 36.
Chaudhury RA. Adoption and usage of electronic medical records in Canadian family practice: Are small practices at a disadvantage?. McMaster University. 2012. https://macsphere.mcmaster.ca/bitstream/11375/12614/1/fulltext.pdf
- 37. Lai J, Lau F, Shaw N. A study of information technology use and implementation of electronic medical record systems in BC medical practices. British Columbia Medical Journal. 2009;51:114–21.
- 38. Terekhova E, Tabassi H, Gabriel P, Jafari S. Telemedicine in primary care: Who are the current users in British Columbia? BCMJ. 2017;59:264–8.
- 39. Vasileiou K, Barnett J, Thorpe S, Young T. Characterising and justifying sample size sufficiency in interview-based studies: systematic analysis of qualitative health research over a 15-year period. BMC Med Res Methodol. 2018;18(1):148. pmid:30463515
- 40. Smith B. Generalizability in qualitative research: misunderstandings, opportunities and recommendations for the sport and exercise sciences. Qualitative Research in Sport, Exercise and Health. 2017;10(1):137–49.
- 41.
Erickson F. Qualitative Methods in Research on Teaching. In: Wittrock M. Handbook of Research on Teaching. New York: MacMillan. 1986. 119–61.
- 42. Jain NR, Nimmon L, Bulk LY. How to … bring a JEDI (justice, equity, diversity and inclusion) lens to your research. Clin Teach. 2024;21(1):e13660. pmid:37874114
- 43. Olmos-Vega FM, Stalmeijer RE, Varpio L, Kahlke R. A practical guide to reflexivity in qualitative research: AMEE Guide No. 149. Medical Teacher. 2023;45:241–51.
- 44.
Saldaña J. The coding manual for qualitative researchers. 3 ed. London, England: SAGE. 2016.
- 45. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3(2):77–101.
- 46. Tsuei SHT, Pokharel BB. Artificial intelligence in healthcare: An opportunity for health system analysis and normative discussion. UBCMJ. 2023;15:6–8.