Although reviews have been carried out on the CLE in primary care for nursing students¹⁹ and on the decentralised and rural training platforms,^20,21 this review focuses on a conducive CLE for medical students. A scoping review is the best tool for a broad overview of the relevant concepts.

What is understood about the factors that contribute to a conducive CLE for medical students in primary care?

This overarching question will be addressed through the following sub-questions:

This scoping review was conducted in line with the Joanna Briggs Institute (JBI) methodology for scoping reviews and the framework of Arksey and O’Malley, with enhancements by Peters et al.²⁴ Reporting followed the PRISMA-ScR checklist.²⁵ The review protocol was registered on the Open Science Framework (https://doi.org/10.17605/OSF.IO/WV4MA) on 19 June 2025. This registration was completed retrospectively, after the initial database searches (09 March 2025–29 March 2025). The retrospective registration reflects the sequencing of project logistics rather than deviations from the protocol established before conducting the review.

A three-step search strategy was employed, as recommended by the JBI. Firstly, a pilot search was conducted in PubMed Scopus, and EBSCOhost on 09 March 2025 to refine keywords and Boolean operators. Secondly, a comprehensive search was undertaken across PubMed, Scopus, Web of Science and EBSCOhost on 29 March 2025. No date or language restrictions were applied. The complete search strategy is provided in Online Appendix 1. Finally, the reference lists of key articles were screened manually for additional studies.^4,25

Grey literature was excluded to ensure a reproducible search strategy across indexed databases. This may have resulted in the omission of relevant policy documents or implementation strategies and may introduce some publication bias.

Eligibility was defined using the Population-Concept-Context (PCC) framework (Table 1). The population of interest was undergraduate medical students including Bachelor of Medicine and Bachelor of Surgery (MBChB, MBBS, MBBCh, BMBS), Doctor of Medicine (MD), and Doctor of Osteopathic Medicine (DO) programmes. The concept was the CLE, defined as patient-facing settings where learning occurs through clinical interaction. The context was primary care, including clinics, community health centres, district hospitals and general practices. Studies conducted in district hospitals were included only when the educational activities described occurred within generalist or family medicine services exclusions applied to postgraduate students, other health professions, homeopathy, simulation-based training, lectures, outreach programmes and hospital-based learning environments. Interprofessional studies were included if data about medical students could be extracted. Only primary research articles were considered, and thus data synthesis and opinion articles were excluded.

After the search, all identified citations were imported into the Covidence software package, and duplicates were removed. The first stage involved screening titles and abstracts against the inclusion criteria by two reviewers (MV and OE). During the second stage the same two reviewers, blinded to each other’s decisions, assessed the full texts of potentially eligible studies. Reasons for exclusion at the full-text stage were documented. Any disagreements between the reviewers at any stage were resolved through discussion and consensus. The search and selection process results are presented using a PRISMA-ScR flow diagram in the third stage²⁵ (Figure 1).

A data extraction tool was developed by the researchers (Online Appendix 2). This tool systematically charts key attributes, including the title, author, year of publication, journal, country of publication, type of study, study purpose, sample size, presence of an evaluation, evaluation tools used, enablers, obstacles, strategies for improvement, and the advantages of primary care as a CLE. Two reviewers conducted the data extraction, which was subsequently reviewed and discussed with the research team to ensure accuracy and consistency. Data charting and initial coding were conducted by a single researcher using a piloted extraction form, based on the Aksey and O’Malley framework.²⁴ A hybrid thematic approach was applied: preliminary codes were informed by the Gruppen framework (Figure 2), while additional codes were generated inductively as new concepts emerged from the data. Related codes were compared and grouped into broader thematic categories based on the principles of reflexive thematic analysis.²⁶ This process involved repeated review and refinement of the coding structure with the research team. The coding structure and thematic synthesis were discussed with the research team to minimise individual interpretive bias. Disagreements regarding coding or thematic allocation were resolved through discussion until consensus was reached as recommended by health professions education research methodology.²⁷

Ethical clearance to conduct this study was obtained from the University of Pretoria Faculty of Health Sciences Research Ethics Committee (No. 546/2024).

The selected articles were geographically diverse. Eleven studies were conducted in the United Kingdom^{3,28,29,30,31,32,33,34,35,36,37}: three each in the Netherlands,^38,39,40 Sweden,^41,42,43 Australia^44,45,46 and South Africa^14,47,48; two each in New Zealand,^49,50 Indonesia^51,52,53 and the United States [US]^53,54; one each in Turkey,⁵⁵ Canada,⁵⁶ Ireland⁵⁷ and Denmark⁵⁸; and a multicountry survey was conducted in 10 African countries.⁵⁹ More than half of the studies were conducted in Europe (56%), and only one-fifth (20%) were conducted in lower and middle-income countries: South Africa, Turkey, Indonesia, and the multicountry African study. Only 6% of the studies were conducted in the United States. Regarding study design, 13 were qualitative, 10 were cross-sectional, and 8 were mixed-methods, one each was a workshop, nominal group technique and a development of a psychometric tool. Among the 34 included studies, most investigated the experiences or perspectives of medical students (n = 28). Only one study focused on patients,⁴⁵ two examined both patients and students,^29,41 two explored faculty perspectives,^14,47 and one investigated clinical teachers.⁵⁰ The enablers and obstacles of the CLE are reported in a table based on a modification of the framework of Gruppen et al. The original framework comprised four interlinked components: personal, social, organisational and space (Figure 2). In the modification, a fifth category, the ‘mentor or teacher’, was added, which intersects with the other categories. The ‘spaces’ category was also clarified to include the clinical interaction with patients in the physical space. Recurrent themes were identified, consolidated, and coded, yielding 31 distinct enablers and obstacles (Table 2). Frequency refers to the number of included studies that identified a specific theme, counted once per study.

The category most frequently reported was mentor or teacher (n = 18), with common enablers including skilled and enthusiastic mentors who model professionalism, facilitation of student autonomy, and providing constructive feedback. Obstacles in this category included a lack of institutional support, authoritarian teaching styles, and poor mentor-university communication. Spaces (physical, virtual, and clinical) (n = 14) featured enablers such as a manageable patient load, variety of clinical conditions encountered, dedicated student space, and engaged patients. Obstacles included long travel distances, inadequate consultation space, and poor internet connectivity. The social domain (n = 13) was characterised by enablers such as a warm, welcoming atmosphere, positive relationships with staff, and integration into the clinical team. Obstacles included a lack of social opportunities and hierarchical cultures. In the organisational category (n = 10), enablers included longitudinal rotations, institutional and clinical learning goal alignment, and embedded feedback processes. Common obstacles included overly theoretical courses, insufficient staff development, misalignment between taught and modelled practice, and poor communication with clinical sites. Finally, the personal category (n = 9) included enablers such as motivation, independence and decision-making autonomy, while obstacles encompassed a lack of agency, fatigue, and insecurity in clinical contexts. The enablers and obstacles are summarised thematically in Figure 3 and Figure 4.

Only six studies explicitly described strategies to enhance the CLE.^{14,38,45,48,49,55} Reported strategies included implementing structured faculty development programmes⁵⁴; providing students with training to improve their sense of agency⁴⁸; involving non-faculty stakeholders in curriculum design⁵⁰; strengthening alignment between universities, clinical sites, and departments of health¹⁴; and fostering collaboration with patients in the planning and delivery of student training.⁴⁵

This was the domain most often mentioned in this review, and it intersected with all the other domains. The mentor’s influence extended beyond immediate supervision to shaping the culture of the learning environment and supporting students’ professional growth. The importance of the clinical mentor is a universal theme across different types of CLEs, from primary care to academic hospitals.^8,21 It seems the difference is that in an academic teaching hospital, mentorship is provided through multiple teachers. In contrast, there is more continuity and attention from a smaller group of mentors in primary care.^8,35,38 An enthusiastic and skilled mentor can facilitate learning even in suboptimal conditions.

The concept of appropriate student autonomy or freedom also cuts across multiple domains. Autonomy is closely related to agency, defined as students’ ability to make decisions and take ownership of learning.⁵⁶ Autonomy appears to be equally important in the hospital and primary care environments. Autonomy was facilitated by adequate space for independent consultations, graded responsibility and mentor support.^38,42 However, obstacles such as controlling teaching styles or a lack of student agency hindered the benefits of autonomy.^50,55 Supporting both autonomy and agency requires explicit preparation of students and staff.

Positive relationships across all domains of the Gruppen framework were central to shaping the CLE, fostering belonging, engagement, and professional growth. From peers and mentors to patients and institutions, the quality of these interactions strongly influenced students’ well-being, motivation, and confidence. The importance of relationships in the CLE is echoed in a scoping review by Hepburn et al.²¹ on rural and remote clinical placements. It appears that in primary care and rural and remote placements, the relationships with staff, peers, and patients are more critical than in hospital-based CLEs.⁸ Despite their importance, little curricular attention is devoted to teamwork and relationship-building skills within CLEs.⁸

The challenge of the availability of appropriate space and ICT infrastructure for students in the CLE is not unique to primary care, as this is also shared with academic hospitals.⁶⁰ A unique challenge to the primary care platform is the cost and time of transport to the CLE.^4,15,19 These logistical obstacles directly affect equity of access, with students from less resourced backgrounds being disproportionately disadvantaged.⁶¹ Extended travel time can also lead to fatigue and loss of clinical time.

The strategies identified in this review primarily targeted organisational domains, reflecting the institutional perspective of most researchers.⁶² These included structured faculty development, fostering student agency, aligning universities and clinical sites, and engaging patients as partners. However, limited attention was given to systemic obstacles such as transport and ICT infrastructure.

The complexity and wide-ranging findings of this review suggest that designing and sustaining a conducive CLE in primary care can be understood as a wicked problem.^16,17,18 The 10 characteristics of wicked problems are mapped against the results of the scoping review in Table 3.

Although none of the studies referred to a conducive CLE directly as a wicked problem, all 10 characteristics of wicked problems can be mapped from the results. In the CLE, mentorship, autonomy, relationships, organisational processes, and logistics are tightly interwoven. Interventions that address one aspect, such as faculty development, may not succeed if other obstacles, such as transport or space constraints, remain unresolved. Similarly, strategies effective in one context (e.g. longitudinal placements in the UK) may be impractical in low-resource settings. Given the complex and context-specific nature of the CLE, the strengths and limitations of this study should be considered in the application of these findings.

This study followed a systematic and rigorous approach that included pre-registration to improve transparency. Major biomedical, educational, and transdisciplinary databases were searched, leading to a significant number of studies (n = 627). Multidisciplinary authorship further strengthened the study, with the lead author’s primary care experience providing contextual insight. The exclusion of grey literature was a limitation that potentially omitted implementation-focused evidence and introduced publication bias. The role of the lead author as coordinator for medical students during their primary care rotation may have introduced interpretive bias, although mitigated through team-based discussions.

The following implications cannot be viewed or addressed in isolation but should be seen holistically and addressed concurrently (Figure 5).

Training institutions should formally acknowledge the central role of mentors and/or clinical teachers and invest in sustained faculty development that integrates clinical expertise with educational skills.¹⁴ This support should extend to mentors and/or clinical teachers not directly affiliated with universities, ensuring their inclusion in curriculum planning and assessment.⁵⁰ They should be supported to pursue further professional development through short courses, diplomas or postgraduate courses, with protected time allocated for their teaching activities.^36,46,50 Mentors should be developed to facilitate these processes in order to shift the focus from what students are taught, to how students learn.^14,48 Mentors should be equipped with the principles of adult learning, and shift the focus away from what students are taught to how students learn.^14,48

Curricula should incorporate a structured approach to autonomy and agency, gradually preparing students to assume responsibility for patient care while equipping mentors to facilitate this process.^3,48 Training should explicitly distinguish autonomy and agency, highlighting students’ active role in shaping their own learning.^39,42 Students should be guided to actively set personal learning goals and to reflect on the learning process.

Relationship-building and teamwork must be recognised as core student competencies. Explicit teaching of interprofessional collaboration, conflict resolution, and patient partnership should be integrated into primary care placements.^3,34,40 Patients should be engaged as active contributors to student learning and their feedback on student interactions should be sought and valued.^37,41

Finally, logistical obstacles must be addressed systematically, particularly transportation, space, and Information and Communication Technology (ICT) infrastructure.^3,14,37 Universities and health systems should collaborate to ensure equitable access, including subsidised travel, adequate consultation space, and reliable digital connectivity.^14,42 This should include investment by universities, such as refurbishment of consultation rooms and internet access for students.^34,40 For cooperation and investment to be sustainable, long-term service level agreements need to be developed between stakeholders.¹⁴