Role modelling to support careers in general practice: a realist review protocol

Stakeholder and patient–public input

Integrated knowledge transition principles will be applied¹⁷ to facilitate co-produced research where stakeholders are involved in all parts of the research process, with the aim of increasing the impact of the work. A group of stakeholders and patient and public involvement (PPI) participants has been formed in the preparation stages of this project and will contribute to the generation of knowledge¹⁸ through three meetings during the review process. The PPI group comprises five members of the public who all have experience of using general practice services; the group has been selected to represent a diverse population, including representatives from areas of socioeconomic deprivation. The stakeholder group has been selected to represent those with a stake in the education of the future GP workforce, including those who have actively chosen not to become a GP. The group consists of a GP working in an area of socioeconomic deprivation, a GP trainee, four medical students from different northern medical schools, a senior GP leader within an undergraduate institution, a secondary care representative, a practice manager, and a senior GP undergraduate teacher. Initial meetings with the stakeholder and PPI group have contributed to the initial programme theory (IPT) (Figure 1) and developing the search strategy. Both groups will support with developing context–mechanism–outcome configurations (CMOCs) as the review progresses and with translating findings into recommendations and potential interventions.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Initial programme theory demonstrating potential contexts, mechanisms and outcomes for exploration in this realist review

Setting

The review will explore literature published between 2013 and 2024 in the English language and will include grey literature. This period has been selected as in 2013 the Shape of training review was published, a key report that impacted on the medical education landscape, highlighting the importance of education in ‘ensuring doctors are trained with appropriate skills, competencies and aptitudes to meet changing needs’.¹⁹

While literature describing UK-based undergraduate education will be the main focus, a worldwide search will be undertaken, as there may be areas of relevance to the developing programme theory, particularly in descriptions of interventions to support role modelling.

Participants

The populations to be considered are GPs and undergraduate medical students; the intervention is role modelling; and the outcomes will include, but are not limited to, career choice. The search strategy is to be undertaken in MEDLINE initially and replicated in other databases, and is described in Table 1.

Information sources

The following electronic databases will be searched: MEDLINE, Embase, CINAHL, ERIC (Education Resources Information Center), the Cochrane Library, and APA PsycInfo. Overton electronic database will be used to identify grey literature, and online magazines GP Online and Pulse will be hand-searched. Policy documents will be identified through searching websites including The King’s Fund, General Medical Council, and Medical Schools Council. The AI search engine Consensus (version 2.0) will also be used for identifying further literature of relevance. Reference lists of included articles will be screened for further articles of relevance.

Article selection

Table 2 shows the inclusion and exclusion criteria that have been developed by the research team, with stakeholder and PPI input, based on the IPT.

Data management

Data will be managed using Rayyan software. Search results will be imported to EndNote 21, then exported into Rayyan once all searches are completed. De-duplication will take place, followed by screening of titles and abstracts.

Selection process

Using the inclusion and exclusion criteria outlined in Table 2, articles will be excluded, included, or assigned to a ‘maybe’ category by the lead author. Those in the included and maybe categories will then be screened at full text to make a final decision on inclusion or exclusion. Other authors will participate in article screening, with a 5% sample shared between them for secondary screening at both title and abstract and full-text screening. Rayyan software will support the allocation of this sample and will blind the secondary screeners to the primary screener decision. Where there are discrepancies in decisions, these will be discussed among the whole team. Articles will be selected for inclusion in the realist review following an assessment of each document’s relevance (whether it contains data relating to relevant contexts, mechanisms, or outcomes or the relationships between these) and rigour (whether the methods used to generate each piece of data are credible and trustworthy).¹⁵ A wide range of documents will contribute to the realist review, which may include primary research studies, policy documents, news articles, websites, and opinion pieces.

Data extraction

Data will be extracted into a data-extraction form, which will be refined and piloted before use. Data will be extracted by the lead author, with a further 5% sample reviewed independently by other authors. Extracted data may consist of descriptions of role modelling interventions, outcomes of interventions, or explanations about how and why role modelling interventions have worked in particular contexts.²⁰ Data that support the use of realist logic to answer the review questions will be of particular interest; for example, data on CMOCs, demi-regularities (semi-predictable patterns of outcomes), and programme theories.²⁰ If articles are unavailable online, authors will be contacted directly. The realist review process is iterative and as the programme theory develops, data will be collected to test the emerging programme theory and reveal underlying mechanisms, therefore the data extracted are likely to change as the review progresses.

Data synthesis

Data synthesis will aim to use collected data to further develop and refine the IPT. Through iterative data collection and synthesis, the IPT will be developed into a series of CMOCs, which will then be tested through further data collection to confirm or refute the CMOCs. Included articles will be considered for data that may be interpreted as functioning as a context or outcome, and then interpretations will be made on how these contexts and outcomes link via a mechanism. Judgements will then be made about how CMOCs relate to the IPT. During analysis, codes will be developed through inductive (data from documents), deductive (from the programme theory) and retroductive (interpretations of the data on underlying mechanisms causing contexts to lead to outcomes) reasoning. Codes will be refined as concepts emerge throughout the analysis, and quotes from the literature assigned to the codes as supporting evidence.

The research team, and the stakeholder and PPI groups will all be involved in the process of data synthesis, development of CMOCs, and causal pathways, bringing their experience to the decision-making process and resolving discrepancies through discussion. Comparisons will be made across the pool of literature to understand how and why observed outcomes occur. For example, we will compare documents that describe different delivery methods of teaching in general practice or types of general practice placement, to understand how contexts affect GP educator role modelling and subsequently influence on undergraduate career choice. Findings will be assimilated as a refined programme theory, which will then be tested with a subsequent realist evaluation.