A multidisciplinary 90-min round-table took place in person on 8 October 2025 during the 9th NASCE Congress at Orsi Academy. Participants represented a broad international spectrum of the simulation ecosystem, including senior surgeons with different backgrounds, surgical trainees, medical students, biomedical engineers, industry developers, simulation center directors, and leaders of European and national simulation societies. Participants were selected by invitation, with the aim of ensuring a representative sample of all key stakeholders involved in surgical training and simulation.

In preparation for the event, invited congress faculty were asked to share their perspectives on the question: “How do you envision the future of healthcare simulation by 2050?” Their reflections were consolidated into a PowerPoint presentation and showcased as discussion prompts during the session, setting the stage for a rich and engaging dialogue. Contributors were selected to ensure broad representation across the simulation community. Invitations were extended to the presidents of related professional societies as well as representatives from surgical trainee organizations and key industry partners advancing simulation technologies and practices.

The round-table was audio-recorded and transcribed verbatim using Microsoft Teams (Microsoft Corporation, Redmond, WA, USA). Rapid qualitative analysis was performed using inductive coding to identify recurring ideas and tensions. Two reviewers independently coded the material and integrated it with the pre-meeting statements. Discrepancies between reviewers were resolved through discussion and the final coding was reviewed and validated by all authors, all of whom were present at the meeting. The synthesis highlights the most representative and divergent opinions, maintaining the spirit of an open, forward-looking dialogue rather than enforcing consensus.

The action items and recommendations presented were initially outlined during the closing phase of the round-table session, where key emerging themes were collectively summarized. Following the meeting, a structured draft of the recommendations was developed by the two lead authors and subsequently reviewed and approved by all co-authors, all of whom had participated in the round-table discussion. It should be noted that the recommendations were not formally validated by all round-table participants, and should therefore be interpreted as author-synthesized reflections of the discussion rather than formally endorsed conclusions.

This scholarly activity constitutes expert consultation and did not involve patient data. Participants provided verbal consent to the use of anonymized excerpts and that, on a voluntary basis, some shared their contact details to review and approve the summarized document after drafting.

Many participants expressed the view that simulation should precede clinical exposure, emphasizing structured proficiency-based progression (PBP) with performance validated against expert benchmarks. This approach minimizes patient risk and standardizes surgical education by fostering skill acquisition through deliberate practice in a controlled environment, rather than relying on variable clinical exposure^[5,6]. Repetitive achievement and benchmarking, rather than fixed score thresholds, were viewed as essential indicators of readiness.

The dialogue also acknowledged that current evidence in simulation research emphasizes efficiency, procedure time, and technical performance, while relatively few studies directly link simulation to patient outcomes^[7]. Participants underscored that the ultimate goal of simulation is not training per se, but to increase patient safety and improve healthcare system efficiency. In essence, simulation-based education is justified only when improvements in trainee performance translate into safer, more efficient, and higher-quality patient care.

Access to high-quality simulation remains uneven across institutions and countries^[8]. Funding responsibilities - whether borne by governments, hospitals, or industry - were actively debated. While several participants argued that making simulation a mandatory credentialing requirement could naturally drive the creation of financial and regulatory frameworks, others felt that this step remains premature, as decision-makers have not yet endorsed it. Overall, many participants converged on the direction but acknowledged some uncertainty about the timeline and feasibility of implementation.

Participants also criticized the “drop-and-forget” approach to simulator acquisition, in which technology is introduced without systematic curricular integration, defined learning objectives, or mentorship support. They emphasized that such isolated implementation fails to deliver meaningful educational or clinical impact.

AI-based coaching, digital learning records, and cognitive-load-based personalization were recognized as transformative opportunities. Participants envisioned “virtual coaches” and adaptive learning systems within five years. Simulation environments represent the ideal setting for deploying AI-driven coaching and adaptive learning, as they allow safe, repeatable data collection, real-time feedback, and standardized skill progression before such tools are translated into the operating room. However, caution was expressed about the lack of validation of several emerging tools. Participants emphasized the need for evidence linking simulation outcomes to clinical performance and patient safety. Haptic feedback was viewed as optional for experts in robotics and minimally invasive surgery but valuable for novices and selected specialties. The conceptual framework itself is not new: Kirkpatrick’s model of training evaluation was first published in 1959 and has remained the standard reference for over sixty years. The gap has therefore never been theoretical; it has been practical. What differentiates the current opportunity is the technical feasibility of achieving Level 4 evidence at scale, enabled by the convergence of robotic platforms, advanced recording hardware, and AI-driven analytics, which now allow systematic, granular capture of intraoperative metrics, including instrument motion economy, tremor, coagulation time, and phase-specific timing, that were previously difficult or impossible to obtain in open or conventional laparoscopic settings.

Equity was identified as both a moral and practical imperative. Resource-constrained and conflict settings face severe infrastructure limitations, demanding low-cost, low-power, or offline alternatives. Participants emphasized that simulation technology should not deepen global inequities. Partnerships, open-source curricula, and regional hubs were proposed to ensure inclusivity. Several advocated combining high-tech with low-fidelity and cadaveric models to preserve tactile learning in diverse contexts.

While many participants expressed the view that standardized curricula are needed to ensure equity and comparability across institutions, training cannot be one-size-fits-all. Needs may vary by country, hospital, and individual profiles. Some trainees require refinement of psychomotor (“hard”) skills, others require improvement in decision-making or communication (“soft”) skills. AI can support adaptive scoring, feedback, and longitudinal tracking to standardize benchmarks, but training pathways should be integrated with human mentorship to interpret outcomes and tailor learning plans.

While AI can streamline quantitative assessment, human mentorship remains indispensable for qualitative feedback, debriefing, and emotional intelligence. Participants discussed the risk of “training empathy out of healthcare” through over-digitalization. Simulation, they argued, should also cultivate empathy, teamwork, and communication. The narrative should shift from “training surgeons” to “enhancing patient safety and professional humanity”. This reflects a division of roles rather than a contradiction: AI is best suited to repeatable, quantitative tasks such as error detection, data tracking, and standardized scoring, while human mentors remain uniquely positioned to address the emotional and motivational dimensions of learning, including frustration and setbacks that arise during skill acquisition.

Under the pre-training stage, the need for structured curricular integration and avoidance of “drop-and-forget” implementation is essential. Simulation may become a requirement for credentialing and standardizing training, using structured and reproducible approaches such as PBP.

Unequal access to simulation and uncertainty in funding and regulatory frameworks remain significant challenges. International collaboration and regulatory support are needed to bridge this gap, while emerging technologies such as AI-driven solutions and low-cost or hybrid models may help improve accessibility in the future.

In the peri-training stage, evaluation should be an integral part of surgical training. It is essential to have scalable, low-cost solutions that combine human mentorship with AI to monitor training progress during peri-training phases. Digital learning records and longitudinal performance tracking will be key to supporting continuous learning throughout the operative process. This ensures that peri-training is monitored, structured, and continuously refined.

Under the post-training stage, the narrative around simulation must center on patient outcomes and healthcare efficiency, not just skill acquisition. Therefore, it is essential to track training impact on clinical outcomes, following principles like Kirkpatrick Level 4. In addition, post-training monitoring, feedback loops, and continuous learning systems ensure that training programs are refined based on real clinical performance, ultimately improving patient care.

Simulation is indispensable but must be validated, standardized, and sustainably funded. AI, including coaching, adaptive learning, and cognitive personalization, can support training, but human mentorship remains essential. Mentors are needed to interpret data, personalize learning, and foster empathy, communication, and other non-technical skills. Additionally, scalable solutions like low-cost or hybrid models could help achieve global equity, but these technologies must first be validated before widespread adoption. Digital learning records and performance tracking will also support continuous learning across stages.

The recommendations derived from this NASCE meeting are summarized in Table 2, providing clear, actionable guidance tailored to different stakeholders, with a focus on integrating simulation, ensuring equity, and fostering validated, human-centered approaches.

In the short term, there is an urgent need to recognize the importance of simulation-based training throughout the entire clinical career, from medical students to residents, surgeons, and subspecialists, with professional societies playing a central role in driving this recognition. In the mid-term, validation of these tools should be solidified, along with integration of AI, distinguishing what remains human-driven versus AI-supported. Personalized ecosystems will start to develop. In the long term, reducing costs and fostering global collaboration between institutions will be essential, with fully personalized ecosystems ultimately approved and adopted on an international scale. Figure 1 provides a schematic overview of this roadmap across the three time horizons.

The themes emerging from this discussion also prompted the authors to identify a number of related considerations that, while not derived from the roundtable itself, merit attention in future work. These should not be interpreted as recommendations of the roundtable, but rather as author reflections stimulated by the discussion. Future roundtables and consensus efforts should include patients or patient advocacy representatives, involved from the design phase of the discussion, to ensure that the patient perspective informs future research priorities and recommendations. Similarly, defining minimum evidence standards and validation trial requirements for AI-based simulation tools, including Kirkpatrick Level 4 outcomes, would benefit from a dedicated, formally structured consensus process such as a Delphi study. Future consensus efforts should also define the operational aspects of digital learning records, including data content, provenance, storage, and governance. The credentialing and legal implications of integrating AI-driven simulation into surgical training likewise warrant dedicated discussion in future consensus initiatives. Finally, the development of illustrative training scenarios and standardized debriefing frameworks, combining quantitative AI metrics with qualitative emotional and behavioral feedback, remains an important direction for future dedicated work.

This study presents several strengths. The multidisciplinary composition of the round-table, encompassing senior surgeons, surgical trainees, simulation center directors, industry representatives, and leaders of European and national simulation societies, ensured a broad range of perspectives across different professional backgrounds and levels of experience. The methodological approach, combining pre-meeting written inputs with a live audio-recorded and verbatim-transcribed discussion analyzed by two independent reviewers, adds rigor to the qualitative synthesis. Furthermore, while several themes addressed in this paper have been previously discussed in the international literature, this work offers a distinctive contribution by revisiting them through a multidisciplinary lens and within a specifically European context, while also introducing emerging perspectives not yet widely explored.

Nonetheless, several limitations should be acknowledged. The heterogeneous composition of the group, while a source of breadth, also introduces the potential for divergent or competing interests: industry representatives, for instance, may have commercial incentives that could influence their perspectives on technology adoption or certification requirements. The total number of participants remains limited (n = 24), and despite the international representation, the group was predominantly European, which may limit the transferability of the findings to other regional and cultural contexts. Future work should aim to overcome these regional barriers, for example through cross-continental working groups and larger participant groups, and extend the discussion to include limited-resource settings. Additionally, the rapid qualitative thematic analysis employed, while appropriate for this type of scholarly activity, is inherently interpretive and does not constitute formal consensus; the recommendations should therefore be understood as emerging expert opinions rather than validated guidelines. Finally, the absence of patient or patient advocate representation is a notable gap, particularly given the paper’s emphasis on patient safety and patient-centered care.

Taken together, this white paper should be interpreted as a synthesis of multidisciplinary perspectives emerging from a structured roundtable discussion, and does not represent a formal guideline or consensus statement. The recommendations presented herein reflect emerging expert opinions intended to inform and stimulate further debate rather than to prescribe standardized practice.