Over the past two decades, it has become increasingly recognized that the historically rooted "see one, do one, teach one" model is no longer sufficient for preparing modern surgeons(1)]. In the United Kingdom (UK), the General Surgery (GS) curriculum underwent significant reforms, including a more holistic approach and more flexibility within training pathways. Despite these efforts, challenges persist. These include limited low fidelity simulation practice, absence of research in surgical education and inconsistent guidance on training(2). The burden on trainees is intensified by financial pressures, staff shortages and lack of dedicated teaching times; all of which influence motivation and career decisions(3). These systemic issues require a change in the current curriculum to a more structured, uniform format with standardized and researched high-quality teaching methods across the UK. Without such modifications, not only does one risk trainee burnout and disparities in training but also nationwide inconsistencies in surgical competence and patient care.

Intercollegiate Surgical Curriculum Programme (ISCP)

The national UK Surgical training pathway begins with the two-year Foundation Programme (FY1–FY2) and is then followed by the Core Surgical Training (CST). The latter also spans two years and exposes trainees to broad surgical exposure to core technical and non-technical skills. The following entry into GS specialty training (ST3-8) is highly competitive and signals the movement towards independent clinical practice. The training path is structured by the ISCP and overseen by the Joint Committee on Surgical Training (JCST).

Competence progression must be demonstrated in key index procedures such as appendicectomies, emergency laparotomies and hernia repairs, through Work-based assessments (WBAs) and Procedure-Based Assessments (PBAs). Their purpose is to evaluate operative planning, performance and post-op reflections. To attain the ability for unsupervised practice, a trainee must demonstrate these skills to a level 4 PBA(4) in all index procedure and record operations in a log-book to lead to completion of training, also known as ‘Certificate of Completion of Training’ (CCT).

The current form of WBAs function as formative and summative are suggested to align with Miller’s Pyramid [Figure 1], a model to assess clinical competence that moves from the more theoretical based assessments to more real-life clinical performance(5, 6). However, research suggests that concerns remain over their effectiveness(7), prompting a need for further research.

UK GS training Issues

Further training issues include external pressure, especially the COVID-19 pandemic as it exacerbated pressure on an already strained system and therefore resulted in significant loss of operative exposure and diminished opportunities for training progression(9). Consequences of this led to disruptions in the ‘Annual Review of Competency Progression’ outcomes and thus extended training periods(10). A UK study found that 81% of surgical trainees reported a decline in mental well-being during the pandemic (11). This coincided with a broader dissatisfaction demonstrated in the 2023/24 JCST survey, demonstrating a disparity between GS and other specialities. Only 52%, a 5% further decline from the previous year, of GS trainees felt that theatre-based quality indicators were met; they scored the lowest among all surgical specialties(12). In addition, only 45% felt their clinical goals were achieved, with only vascular surgery scoring worse(12). Despite the formal requirement of two-hourly weekly protected teaching times, only 70% of trainees reported this as being met. Similarly, only 69% of GS trainees reported the opportunities for mandatory simulation training, in contrast to 92% of Ears, Nots and Throat and 86% of paediatric trainees.

The trainee dissatisfaction and inconsistencies across different specialities is demonstrating a variability in teaching quality across different specialities(12) and thus runs the risk of discrepancies in patient outcomes subjected to different specialities. These trends demonstrate an urgent demand to re-evaluate the GS educational programme.

Surgeon as Educators

Mastery of surgical skills does not develop from theory alone but through repetitive, hands-on practice(13, 14), hence the pivotal role of a senior in shaping the progression of a trainee through direct supervision. In spite of this expected role, surgeons are not required to obtain a qualification in pedagogy(15). NHS England’s 2024 review demonstrated that neither CST nor ST curriculum mandates such formal qualification(15, 16), this throws light on the systematic gap in the curriculum(17). With the aforementioned curriculum issues, this lack of standardized teaching runs the risk of undermining both patient care(18) and the sustainability of surgical training itself(19, 20).

Surgical Education Learning Theories

Understanding of the theoretical frameworks that underpin effective teaching is crucial in improving surgical educational outcomes. Kolb’s Experiential Learning Theory(21) is particularly suited to surgery, outlining a four-stage cycle: concrete experience, reflective observation, abstract conceptualisation, and active experimentation; demonstrated by a trainee performing a procedure, reflecting on it, identifying improvements, and applying these to future scenarios. Reflection is a central process widely promoted in the medical field and supports the transfer of cognitive to clinical knowledge and thus progression of learning(22). Dewe’s (1933) concept of reflective thinking depicts it as a response to cognitive doubt prompting problem-solving and has been a foundation in defining critical thinking(23). This suggests that critical reflection in surgical training could drive continual self-improvement and adaptation(24, 25). Reflection is seemingly incorporated into the progression of Miller’s Pyramid of competency[FIgure1] from cognitive knowledge to clinical performance(5). Further research into curriculum and teaching methods should be based on such foundational learning theories and pedagogy.

Gap in Literature

A review of current literature reveals a significant gap in GS curriculum and effective teaching methods. Whilst some research explores surgical assessment strategies(26), there is a lack of teaching methodologies focusing on GS trainees within the UK. This is significant given the notable discrepancies in reported satisfaction of GS trainees compared to other specialities(12). In addition, UK training carries its own hurdles that must be researched in isolation to other countries. To date, no scoping or systematic review has comprehensively examined teaching and assessment methods specific to UK GS to ascertain competence and whether the current WBA based strategies are most appropriate. This review seeks to address this disparity.

Aim, Research Question

A scoping review was performed due to limited volume of existing research, therefore making a full systematic review unsuitable(27). This approach allows the mapping and organization of a vast topic whilst also identifying current gaps within the current literature(28). The five-stage methodological framework outlined by Arksey and O'Malley(27) was followed, which includes (a)research question identification, (b) identification of relevant studies, (c) study selection, (d) data charting, and (e) collating and reporting the results. The review was conducted in accordance with PRISMA-ScR guidelines for scoping reviews(29).

Eligibility criteria

Studies required to meet the eligibility criteria to be included. Firstly, GS trainees in the UK were the target population as the research question is addressing this population specifically. International studies were excluded as they vary in curriculum, funding, training duration, available resources, training equipment and possibly teaching qualifications.

Medical students were also excluded due to discrepancies in their curriculum and learning objectives. The target trainees were defined as postgraduate doctors pursuing a career in GS, including those without an official training post (e.g., trust-grade doctors). Eligible teaching methods encompassed any form of simulation or formative assessment strategy that could lead to surgical skill acquisition.

A systematic search was performed on the 18th February 2025 of Pubmed/Medline, EMBASE, ERIC and Web of Science. The following Boolean operators were used to identify the papers via OVID: GB OR Great Britain OR UK OR United Kingdom) AND (Teaching OR Teach* OR training OR mentoring OR mentor*) AND(General surg* OR colorectal OR HPB OR hepato-pancreato-biliary OR UGI OR Upper gastrointestinal) AND (Simulation OR formative OR assessment) AND (Surg* skill* OR technical OR competence) AND (Surgical trainee* OR surgical resident* OR Student OR graduate). Web of Science was not accessible via OVID and hence the same search was completely using the same terms. The search was limited to the last ten years and to the English language. Additional relevant papers were identified through tracking references to enhance literature saturation. Grey literature (e.g. abstracts) were considered as these suggested innovative topics with robust outcomes.

Screening

The identified files were exported in RIS file and imported to Zotero reference manager(30). This facilitated the upload to Rayyan(31), a research collaboration platform that supports systematic reviewing of papers. The first author (AA) screened the eligibility of the retrieved papers based on title, abstract and then full paper; this was then followed by a second blinded reviewer (EH), using the same eligibility criteria, to increase reliability,

Charting

Data from the selected studies was tabulated in Microsoft Excel and categorized under major subheadings: paper ID, study design, aim and outcomes. A simple numerical analysis was conducted, followed by a qualitative narrative synthesis organizing the studies according to the four levels of training evaluation of Kirkpatrick’s framework. This widely used model assesses the effectiveness of educational programs(32) starting with the learners’ immediate reaction to learning (level 1); followed by the evaluation of learning by testing the gain of knowledge (level 2). In level 3, it evaluates application of a new skill, and finally, it measures the broader impact of learning e.g. patient outcomes (level 3)(32) (Fig. 2).

A list definition can be found in [Appendix 1].

The initial search generated 201 papers and an additional 3 were identified through alternative sources, after removing 48 duplicates in ovid and 5 in Rayyan, 151 papers remained. After screening the titles and abstracts against the eligibility criteria, 78 and 5 papers were excluded respectively to leave 67 records for full text analysis(Fig. 3). A further 36 records were removed during the full-text article review, due to incorrect geography (n = 29), incorrect population (n = 2) e.g. med students, and wrong intervention (n = 5) e.g. endoscopy. Despite endoscopy being performed by many GS surgeons, the training follows a different curriculum and was therefore excluded. Some studies often included other geography/population in their study, they were only included if the majority of studies were relevant. One urology paper was not excluded, as the procedure tested was equally performed by GS trainees. A vascular paper was not excluded as GS and vascular training paths only separated in 2022(64). One scoping review was removed, yet the decision was made to keep one systematic review as it matched this paper’s methodology perfectly, capturing all Kirpatrick’s levels. This left 31 final studies for the scoping review.

Study design and setting

The included studies fell into five broad methodological categories: self-report measures, objective or performance-based assessments, descriptive analyses, and multi-methods. Thirteen used self-reported data, three through interviews, seven via questionnaires, two with focus groups and one via polling. Seven studies employed objective assessments: five assessed technical performance, while others used tools such as S-QAT (a self-assessment tool appraising training quality), OSATS (evaluating surgical technical skills), STAI (measuring state and trait anxiety), debrief scores, and NOTSS (assessing non-technical surgical skills). Eight were descriptive in design, including six studies analysing ISCP and/or eLogbook data and two based on training reports. One paper was a systematic review.

Distribution by Kirkpatrick level

The papers were then categorised according to Kirkpatrick's Four Levels of Training Evaluation (Fig. 2); many papers assessment incorporated multiple levels. They were distributed most commonly in level 1 with eight papers, then eight in level 3 and seven were under a combination of levels 3 and 4. Three papers were categorized as level 2. Two fell into level2/3 and then one paper in each of the following categories: level1/2, level 1/2/3/4, level 4(Fig. 4).

Level 1[Table 1]

Eight studies [Figure 4] assessed trainee reactions to simulation or formative assessment interventions, focusing on perceptions, confidence, and engagement rather than measurable outcomes. Data were largely qualitative, drawn from interviews and questionnaires. Trainees often reported increased motivation and confidence, particularly following simulation-based teaching such as bootcamps and skills camps. There was strong support for structured, modular, and mental rehearsal-based approaches (40), though barriers to engagement included service demands, trainer resistance, and a tick-box WBA culture (e.g.33–35).

Level 2[Table 1]– Learning

Three studies [Figure 4] measured evaluation of learning through knowledge e.g., impact of 3D diagnostic skills(43) and visual aids(44). Outcomes were demonstrated with objective metrics like post-test scores, NOTTS, S-QAT and OSATS. Improvements were seen especially if combined with mental rehearsal (44) structured debrief (46). Multimedia, 3D visual aids, and mental imagery enhanced performance, confidence, and reduced errors(44).

Level 3[Table 1]– Behaviour

Eight studies assessed evaluation of skills [Figure 4]. Evidence was largely anecdotal or inferred, based on questionnaires or post-training interviews. A consensus was pro-simulation training. Reported changes included improved operative confidence and patient safety in simulations (45–46, 49–54). Improvements in independence were enhanced further by increased fidelity of model (50), dedicated training program.s (51) and deliberate practice (54). However, cost and limited trainer and training time were common barriers (47, 48, 53).

Mixed levels[Table 1]

Eleven papers used multi-methods to assess learning on multiple evaluation levels simultaneously allowing in-depth multimodal assessment of learning. Level3/4, being the most common combination, offered the opportunity to evaluate new skills in the clinical field, often via formative assessment PBAs. Competence progression across training years was shown with a gradual decrease in supervision, especially for less complex procedures (55–60). Learner’s increased level of autonomy could infer positive clinical outcome, but not guarantee it, hence the level3/4 grade as opposed to level4. Level2/3 tested NOTSS skills and tested these via objective post sim methods, coaching improved real-time decision-making and teamwork (45, 46). A training module was able to evaluate independent skill progression through questionnaires and test clinical outcome metrics objectively(61).

One cadaveric review also reviewed the Kirkpatrick levels and therefore (62) provided: strong trainee approval, knowledge gains, procedural improvement and modest real-world outcome effects, to address all four levels of the training experience(Fig. 2).

Level 4[Table 1] – Results

Only one study [Figure 4] evaluated solely the clinical impact (63), this study examined the impact of a national training programme and noted improved in-training competency and improved real-time patient outcomes. The remaining studies were tested in conjunction with level 4.

Strength of Evidence of each Kirkpatrick Level

In this scoping review, the strength of evidence for each Kirkpatrick Level was assessed using a combination of methodological accuracy, data type, and the depth of outcome measurement reported in included studies and which type of levels and teaching opportunities would be best to improve surgical skill acquisition[Table 2]. Eight papers were rated low and were in Level 1 as it relied on subjective trainee perceptions(33–40). Level 2 papers involvin were categorized as moderate, if studies included objective, quantifiable outcomes (41–46). Papers in level 3 with self-rated outcomes through questionnaires were graded moderate due to their clinical relevance (49–53). However, papers scored high when clinical outcomes were measured directly which only one paper qualified for (62) and was categorized as level 4.

Studies spanning multiple levels were assessed contextually, with the highest level of rigorously collected and validated outcome data informing their overall strength. Papers in level 3/4 were rated moderate-high and contained eight papers that relied on large-scale datasets such as national trainee progression records. PBAs ratings of independence could indirectly support good clinical outcome (55–60).

This evaluative strategy aligns with well-known best-practice frameworks for educational evidence synthesis and therefore could relate to effective skill acquisition and potential for further research.

This scoping review aimed to identify which teaching methods used by GS in simulation and formative assessment contribute most effectively to surgical skill acquisition in UK trainees and which gaps remain in the current limited research surrounding this topic. The Kirkpatrick model(32) and Miller’s Pyramid(7) provided a valuable conceptual scaffold to interpret the nature and depth of learning outcomes revealed in this scoping review, while the Kirkpatrick Model enabled categorisation of educational interventions based on their evaluative impact, Miller’s framework clarifies the type of competence being developed at each stage. By using Kirkpatrick’s model to interpret the outcomes, an uneven distribution of evidence strength and a dominance of lower-level evaluation was identified. By using Miller’s pyramid it was still possible to highlight the progression from "knows" and "knows how", through cognitive gains and improved procedural understanding and "shows how" via simulation or structured assessment(65–67).

Level 1(33–40) outcomes were the most frequently reported, capturing trainee perceptions, satisfaction, and confidence e.g., via Likert scales or interviews. Trainees often reported increased motivation and confidence, particularly following simulation-based teaching such as bootcamps and skills camps. While many trainees valued simulation such as teaching camps for building confidence, some questioned its realism and educational value(39). There was strong support for structured, modular, and MR approaches (40), though barriers to engagement included service demands, trainer resistance, and a tick-box WBA culture (35). MR was regarded as valuable across all levels(40).

Level 1 outcomes, although high in quantity, most were single-centre, underpowered and were reliant on subjective tools to determine outcome. As a result of the subjective nature of their outcome they were classified as a weaker form of evidence. Their generalisability was limited and they failed to assess actual knowledge attainment or long-term clinical relevance

The deliberate proactive engagement suggested in Kolb’s cycle of reflection leads to the transfer of cognitive knowledge to clinical knowledge(22), one could argue that this was executed through self-reporting outcomes. Insights into attitude could indicate confidence, motivation and readiness(Kirkpatrick level1) but offered little measurable indication of learning, hence unable to confirm knowledge acquisition was achieved (Miller’s base).

Level 2(41–46) studies demonstrated more objective improvements across the cognitive domain, particularly with moderate-fidelity simulation and blended training models. Objective learning outcomes were assessed via post-intervention knowledge tests, diagnostic skill assessments, NOTTS, S-QAT and OSATS. They offered a means to demonstrate statistically significant improvements in post-intervention test scores. This level of evaluation provided stronger evidence for the acquisition of knowledge. Significant gains were seen with support of coaching(45), 3D visual aids and mental imagery to enhance performance, confidence, and reduced errors(44).

The strength was rated moderate, when using objective tools, to assess technical and cognitive acquisition. Limitations of its effectiveness in teaching included lack of long-term follow-up and over-reliance on surrogate endpoints like confidence or task completion and lack of skills assessment in the clinical field limited mixed simulation fidelity constrained broader application.

Level 2 tools evaluated knowledge gains and skill acquisition and mapped with Miller’s "Knows” when testing knowledge(41–44) and the “Knows How" stage, when testing knowledge through clinical problem solving(45–46). Level 2 teaching methods can be considered more effective due to their reliance on objective testing to assess knowledge retention yet are still flawed due to their lack of direct skill assessment.

Level 3(47–54) appraised skill attainment and predominantly explored behaviour change in the workplace, often self-reported through post-course feedback engagement. These studies linked simulation training to improved workplace preparedness, particularly in procedural confidence and clinical decision-making(51). Objective assessment tools such as post-test scores, simulation metrics and timed laparoscopic tasks were occasionally employed to measure time, accuracy and error rates e.g., laparoscopic exercises(54). However, most relied on self-reported post-course questionnaires to demonstrate improved skill improvement, confidence(49, 51, 52, 53) and teamwork dynamics(46). Simulation courses led to increased confidence and perceived readiness for registrar-level tasks(49). Simulation was identified as a driver of workplace behavioural change when structured feedback and debrief were included(46).

The evidence was graded moderate as outcomes were mostly deducted from self-reported growth without observational objective verification. In addition, fidelity of a simulation could affect the transferability of skill. Furthermore, minimal long-term follow-ups undermined the reliability of these paper’s findings.

Interventions involving high-fidelity simulation, mental rehearsal, and modular training allowed trainees not only to understand a procedure but to demonstrate it in controlled environments. When learners demonstrate skills in a simulated or structured setting, they “show how” (Miller’s) behaviour could change (Kirkpatrick’s level3). One may interpret the potential for good learning progression to the intermediate levels by demonstrating some behaviour changes indicative of competence and skill development.

Level 4 was demonstrated, in isolation, only by one paper(63) and evaluated clinical impact. It highlighted the benefits of a dedicated training programme in laparoscopic colorectal cancer surgery through demonstrating both the clinical benefits through mortality and morbidity reduction; as well clinical performance through assessment tools(63). This extended on previous papers where trainees call for such dedicated programs and structures(36,46,48,51,52,61) and should be incorporated into further curriculum work. Trainees, almost unanimously, called for dedicated teaching time, trainers, and incorporation into the training programs(11, 36, 46, 48, 51, 52, 61).

Despite the low quantity of papers, the sole paper(63) was considered a higher grade in strength due to high fidelity and thus skill transferability. Measuring clinical patient outcomes is arguably the most reliable way of testing skill outcomes. Reliability was high due to the three years follow up to monitor trainee’s competency.

Level 4 of Kirpatrick model matched with the maximum competence (“Does”) stage of Miller pyramid by demonstrating patient outcomes. It could be assumed to be the most rigorous form of assessment of skill attainment. One can infer the rarity of such papers to be secondary to logistical, ethical and financial challenges in measuring patient outcomes. Overall, level 4 papers were underrepresented, reinforcing the need for more longitudinal, mixed-method studies; with the capability of delineating impact of simulation and formative assessments from training environments to real-world clinical practice.

Majority of papers used multiple methods to address multiple levels of learning categories, which allowed insights across multiple levels. For example, level2/3(45, 46) deducted objective skill acquisition assessment and measured the broader impact of training outcome. The most common form was level3/4, which used large-scale national datasets from ISCP and eLogbook platforms to demonstrate index operative experience and achievement of a third level 4 competence that leads to CCT(56–60). Being based on real patient operations one can assume, indirectly, that increased level of independent operating (behaviour) is related to better patient outcome. Their long-term follow-up was associated with increased reliability. Interestingly, many trainees failed to obtain WBAs indicative of independence in their index procedures, suggesting that indicative numbers alone could fail to confirm true competence thereby questioning the validity of CCT by using WBAs(57, 58, 59). Furthermore, WBAs were often pre-filled by the trainees and inconsistently validated by their trainers(38). A recurring concern was the perceived ineffectiveness of WBAs/PBAs, with trainees reporting these often being used as ‘tick-box’ assessments(35), countering their educational intent.

Mixed levels studies evidence strength depended on the type of outcomes measured. Level 4 outcomes, alone or with level 3, provided rigorous generalisable data and insight into long-term impact on learning due to their large data, duration and reliability. In contrast, these were often associated with subjective or inconsistent validation means thereby lowering their evidence strength. If behavioural change through clinical application in real-world settings was demonstrated, then the "Does" level of Miller’s Pyramid was reached. However, post-certification(CCT) performance was poorly tracked therefore, making it difficult to assess competency, at the ultimate phase of integration into practice. Arguably, mixing the forms of assessments may yield valuable data, which is less ethically challenging than pure level 4 studies. Interest persists in creating more higher level competence-based assignments and to revise current WBAs given the current collated negative feedback.

This review has several limitations, firstly, the inclusion of heterogeneous study designs, populations (e.g., CTs, STs, SAS), and simulation modalities complicated direct comparison. Secondly, many included studies were observational, underpowered, or single-institution, limiting the external validity of findings. Thirdly, the nature of a scoping review precludes formal quality appraisal or meta-analysis. Fourthly, data focused solely on UK-based data, which, while relevant to the research question, may omit valuable international insights. Finally, reporting bias may exist, as negative or inconclusive studies are less likely to be published.

By linking Kirkpatrick’s levels with Miller’s ascending hierarchy of competence, the review supports that true surgical skill acquisition is multidimensional, requiring not just improved learner satisfaction or knowledge gain, but demonstrable procedural capability and transfer into practice. Integrating both models in curriculum and teaching design enables a more comprehensive evaluation of medical education from trainee satisfaction to patient outcome. Pursuing research into level 3 and 4(/“Shows How” and “Does”) is likely to be the most effective means to identify solutions to improving surgical skill acquisition.

This scoping review maps the current landscape of simulation and formative assessment in UK general surgery training, using the Kirkpatrick model to evaluate effectiveness of educational interventions whilst also describing the progression of knowledge to performance with Miller’s Pyramid of competency. Linking Kirkpatrick’s model with Miller’s pyramid underscores the need for educational strategies that move beyond early-stage gains to demonstrate actual competence in clinical settings. While simulation and formative assessment improve trainee confidence and technical ability (Levels 1–2, “Knows”/“Knows How”), fewer studies assess behavioural change or real-world performance (Levels 3–4, “Shows How”/“Does”). Future research should prioritise prospective, longitudinal designs capable of tracking performance into independent practice and thereby focusing on Levels 3–4, “Shows How”/“Does”, criteria as it demonstrates more ingrained learning for surgical skill acquisition. This form is already part of curriculum in form of WBAs, yet they need further research to improve given plentiful complaints raised. Embedding effective, evidence-based strategies into cohesive, competency-based teaching methods, may ultimately support a more accountable framework for developing safe and more consistent GS training across the UK.