We included studies on simulation-based interventions in medical education and training in Pakistan, with no language restrictions. Eligible studies: (1) reported implemented SBE interventions in the medical field with any level of learners (undergraduate, postgraduate or continuing professional development) and (2) were conducted within the Pakistani healthcare or educational system, with at least one Pakistan-based author. We excluded studies unrelated to simulation, not focused on medical training, conducted outside Pakistan or involving mixed-country participants without disaggregated Pakistani data. Studies without Pakistani-based authors, even if published locally, were excluded. Reviews, editorials, letters, short communications, opinion pieces, conference abstracts and studies that only proposed or advocated simulation without implementation were also excluded.

We conducted a series of electronic searches, guided by a medical librarian from a leading healthcare institute, to identify eligible studies. Searches were performed on 25 September 2023, 29 February 2024, 3 July 2025 and 22 July 2025 across PubMed, CINAHL, PsycINFO and PakMediNet [14]. PakMediNet, a Pakistani database, was searched to broaden coverage because many Pakistani authors publish in regional journals that are not indexed in major international databases.

The tailored search strategy for each database comprised three components: (1) simulation terms, (2) education/training terms and (3) Pakistan terms, combined using Boolean operators. Controlled vocabulary (e.g. MeSH and CINAHL Headings) and free-text terms were used where applicable. In PakMediNet, which does not support complex Boolean queries, we manually searched combinations of simulation and education terms derived from the main strategies. Full search strategies for all databases are provided in Supplementary Material Appendix A. In addition, we manually searched the reference lists of all included studies to identify further eligible articles.

Citations from literature searches and reference list checks were imported into Covidence, where duplicates were removed [15]. To establish inter-rater reliability, three researchers (FN, NF, MB) independently screened the titles and abstracts of the first five studies and resolved inconsistencies through discussion. Subsequently, two reviewers independently screened all remaining titles and abstracts, with MB resolving disagreements through discussion and consensus.

For full-text screening, inter‑rater reliability was again established on five studies, followed by independent full‑text screening by two reviewers, with MB mediating any disagreements to reach consensus on inclusion or exclusion. Reasons for exclusion were documented during full-text screening. The overall study selection process is summarised in the PRISMA flow diagram (Figure 1).

For data collection, a data extraction sheet was developed in Covidence, pilot-tested on five randomly selected included articles, and then refined. For establishing the inter-rater reliability, all researchers independently extracted data for five studies, with conflicts resolved through discussion. Remaining data extraction was done independently by two researchers, with MB resolving conflicts. Cohen’s kappa (κ = 0.741) for inter-rater agreement was determined by Covidence [15], demonstrating strong agreement on the data points gathered by the researcher team. While this value is generally considered substantial agreement, it may suggest non-trivial disagreements regarding study design and outcomes. However, Covidence registers variations in data formatting (e.g. capitalisation differences, comma placement etc.) as discrepancies, which have marginally influenced the kappa value.

Extracted variables included: publication year, author details, study aim, study design, profession and learner level, simulation modality and description of the simulation activity, evaluation or assessment tools, tools’ psychometric properties (where reported) and simulation-relevant barriers and facilitating factors. Outcomes of each study were categorised using Kirkpatrick’s model as level 1 (reaction/satisfaction), level 2 (learning/knowledge/skills), level 3 (behaviour/change in practice) or level 4 (results/system impact) [11].

Because of high variability in study designs, participants and outcomes, conducting a meta-analysis was not possible. Given that this review aimed to understand the nature and scope of simulation-based interventions in Pakistan rather than calculate pooled effects, we conducted descriptive statistics (frequencies and proportions) to summarise key features of the studies. We synthesised the knowledge about the study foci through thematic analysis, and each study was also subsequently examined using the DOI framework to map the current landscape of simulation and identify strategies to improve adoption and scalability.

The thematic analysis of the primary foci was conducted by one author (FN) using Braun and Clarke’s framework [16], and was verified by another author (MB), with discrepancies resolved through discussion and review of the original study. Outcomes were mapped to Kirkpatrick levels (L1–L4) during data extraction. Using the combination of study foci and Kirkpatrick outcome levels, two authors (MB, FN) then independently assigned each study to one of Rogers’ five DOI stages using the following criteria: 1.

Knowledge (Stage 1): initial introduction or report of simulation with no documented adoption evidence.

Disagreements were resolved through discussion and re-review of original studies, with MB mediating final decisions. When available, we used follow-up data or longitudinal documentation of sustained use (e.g. multi-year ongoing delivery, faculty adoption post-training) to inform DOI stage assignment. Therefore, DOI stage assignment prioritised evidence of adoption and sustainability. Studies reporting lower Kirkpatrick levels (L1–L2) but with documented sustained use or follow-up showing adoption were classified at higher DOI stages (Implementation/Confirmation). See Figure 2.

We assessed risk of bias using the Joanna Briggs Institute (JBI) critical appraisal tools [17] for cross-sectional studies, case reports, quasi-experimental studies and RCTs and the National Heart, Lung and Blood Institute (NHLBI) tools [18] for pre-post studies without control groups. Two reviewers (MB, FN) independently applied the tools, recording supporting information and justifications for each domain. Discrepancies were resolved through discussion, and full item‑level ratings and thresholds are shown in Supplementary Material Appendix B.

Each study was coded according to its central purpose, whether it introduced simulation tools, evaluated instructional effectiveness, supported faculty development, delivered teaching or explored simulation conceptually. Five themes were identified: 1.

Simulation as a commodity: the first theme, simulation as a commodity, focused on reporting the effectiveness of simulation as a commodity, highlighting the development, validation or use of simulation tools, simulation techniques, simulators and simulation centres, including wet lab setups, virtual platforms and task trainers (n = 23, 31%) [19,26,39,40,42,43,46,49,51,53,56,58,66,68,70,71,73,74,77,78,80,85,91]. The individual codes included in-house task trainers [19,40,56,58,71,85], commercial simulators [26,43,68], animal cadavers [42,46,53], three-dimensional printed models [51,91], simulation space including SCIL labs [66,80], virtual simulators [70,74,77,78], simulation technique [73] and adherence to treatment guidelines in specific clinical scenarios [39].

To address the third research question, we examined the reported barriers and facilitators influencing SBE in Pakistan, focusing on how they shaped progression through the five stages of adoption: Knowledge, Persuasion, Decision, Implementation and Confirmation.

Barriers were reported in 64% of the included studies (n = 47) [19,20,22–27,30–32,35–38,40,41,44,46,47–51,55,57,60–64,66–70,72,73,75,79,81–83,85,87,88,90] and affected all DOI stages. The most cited challenges were structural: lack of funding (n = 12, 16%) [26,27,41,47,48,55,61,64,66,81,82,83] and inadequate infrastructure (n = 12, 16%) [22,24,26,40,41,46,47,55,62,79,81,83], both of which typically limited institutions to the Knowledge or Decision stages, with few advancing towards sustained implementation. In many reports, funding concerns were framed around the perceived need for high‑cost facilities and equipment, even though effective simulation can also be delivered using lower‑cost, pedagogy‑driven approaches. Human resource constraints were also noted, including limited availability of trained faculty (n = 2, 3%) [51,69] and insufficient time for simulation development (n = 4, 5%) [21,40,46,56]. These issues often disrupted progression from Persuasion to Implementation.

Technical limitations related to fidelity were another recurring theme. Anatomical inaccuracies (n = 7, 10%) [30,40,48,70,81,82,88], absence of force feedback (n = 1, 1%) [30] and inconsistent simulated patient (SP) performance (n = 2, 3%) [44,87] negatively affected learner engagement and reduced confidence in simulation as an effective educational tool. These issues commonly stalled advancement at the Implementation stage.

Barriers to Confirmation – the stage representing routine, sustained use – were less frequently reported but significant. Chief among these were limitations in outcome evaluation, including absence of long-term follow-up (n = 7, 10%) [23,25,49,60,61,72,88] and rater inconsistency (n = 7, 10%) [27,32,57,60,63,73,87], suggesting a lack of robust systems for monitoring long-term effectiveness. Notably, 34% of studies (n = 25) [21,28,29,33,34,39,42,43,45,52,53,54,56,58,59,65,71,74,76,78,80,84,86,89,91], most of which were descriptive or in early adoption stages, did not report any barriers.

Facilitators, while less frequently mentioned, were most often reported in studies situated in the Persuasion, Implementation or Confirmation stages. The most frequently reported enabler was the use of cost-effective simulation tools and techniques (n = 18, 25%) [22,26,27,29,32,40,41,42,46,48,54,56,70,72,73,77,85,88], which supported wider adoption and institutional feasibility. Experiential learning models (n = 15, 21%) [23,29,35,50,53,54,63,72,73,75,81,83,86,88,91] and access to trained facilitators (n = 6, 8%) [28,31,41,43,74,75] were also key factors, enhancing learner engagement and instructional quality – both critical for successful implementation.

Other supports included validated assessment instruments (n = 5, 7%) [20,22,40,47,54] and structured debriefing practices (n = 2, 3%) [57,75], which contributed to iterative refinement and outcome measurement, aiding progress towards Confirmation. Additional facilitators included high-fidelity simulation platforms (n = 4, 5%) [30,34,53,88], user-friendly interfaces (n = 1, 1%) [70] and portable or updatable simulators (n = 3, 4%) [48,61,88], which improved usability and scalability.

A comparative analysis revealed that studies reaching later DOI stages were more likely to report facilitators and systemic support. In contrast, those in earlier stages more often emphasised barriers or omitted contextual factors altogether. This trend suggests that the presence of enabling structures – both technical and institutional – is closely associated with successful and sustained adoption of SBE in the Pakistani healthcare context.

Simulation adoption in Pakistan remains fragmented, with over 80% of interventions situated in the Knowledge or Persuasion stages of the DOI framework. Most are limited to workshops, pilot sessions or partial curricular modules and report lower levels of Kirkpatrick outcomes focused on satisfaction or knowledge gains. Fewer studies in the Decision, Implementation and Confirmation stages of DOI highlight a persistent gap between innovation and institutional integration, implying that despite growing interest, simulation remains largely exploratory rather than embedded in medical education. Without strategic investment, leadership and policy support, simulation risks remaining a peripheral innovation rather than a routine educational approach.

Tools such as low-cost task trainers and in-house models were significantly more common than high-fidelity or digital modalities, which, when present, were rarely integrated into assessment frameworks or policy. Ali et al. ‘SonoGames’ [20] exemplifies purposeful adoption, showing how gamified, team-based simulation, pre-/post- evaluation and multi-institutional collaboration can enable transition into the DOI Decision and Implementation stages. These adoption bottlenecks parallel the challenges observed in other LMICs in the region, where policy ambiguity, underfunded infrastructure and fragmented leadership restrict sustainable diffusion [92]. Nagarajappa et al. report persistent infrastructure and faculty constraints in India despite national skills lab mandates [93]. Del Castillo Miranda et al. describe early-stage, low-fidelity simulation dominance in sub-Saharan Africa (Ghana, Uganda, Kenya) mirroring Pakistan’s 84% Knowledge/Persuasion distribution [94]. Southeast Asian contexts face similar policy-implementation gaps. Robinson et al.’s synthesis of 97 studies across 50 LMICs confirms Pakistan occupies a representative mid-LMIC position – excelling in low-cost innovation yet trailing high-income countries’ systemic integration [95].

Structural and human resource barriers continue to impede simulation adoption. Infrastructure and funding deficits were cited in 65% of studies; inadequate facilities, insufficient equipment and suboptimal maintenance were commonly reported. Faculty shortages and the absence of accredited, locally relevant training programmes further hinder scale-up. These systemic constraints also deepen regional disparities in access to SBE. Without targeted investment in infrastructure and faculty development, especially in underserved areas, scaling simulation nationally will remain out of reach.

Conversely, interventions supported by engaged leadership, collaborative programming and structured mentorship, such as local ‘train-the-trainer’ initiatives, were more likely to progress beyond pilot phases. Facilitators, including cost-effective simulation modalities, tailored assessment tools and local adaptation, may have contributed to wider adoption in 18 studies. The role of faculty development emerged as a specific enabler in 27% of cases. To better understand the factors influencing simulation adoption, we mapped the thematic findings onto Rogers’ five DOI attributes (Table 2). This analysis highlights that while trialability and relative advantage can support early-stage adoption, low compatibility with existing pedagogies, high perceived complexity and limited observability remain critical barriers.

In this review, the distribution of studies across DOI stages mostly aligned with Kirkpatrick outcome levels, however, a few exceptions remained. The alignment between Kirkpatrick levels and DOI stage classification reflects the internal consistency of the mapping framework, as Kirkpatrick levels informed DOI stage assignment. However, these exceptions reveal a gap that extends beyond a measurement methodology, demonstrating that Kirkpatrick outcome levels alone are insufficient for gauging the uptake of simulation. This finding highlights SBE as a complex, real-world implementation problem requiring an implementation science framework to account for multiple contextual factors. The 84% concentration of Pakistani studies in Knowledge–Persuasion phases versus 16% in Decision–Implementation–Confirmation stages suggests structural barriers – funding, infrastructure, leadership – rather than outcome maturity alone, a phenomenon also supported by determinant theories in implementation science research on health innovation adoption in LMICs, including education contexts [96].

Late-stage progression requires organisational infrastructure and leadership alignment. Pakistan’s barriers mirror these documented systemic factors. Addressing them requires integrated implementation: simulation champions with protected time, institutional funding models and Ministry of Health guidelines embedding simulation into organisational routines. Building organisational infrastructure – the prerequisite for sustained adoption documented in implementation science – requires policy leverage. Pakistan’s regulatory structure through PMDC/HEC provides exactly this mechanism by providing the PMDC/HEC oversight on the accreditation and incentivising through regulatory requirements what otherwise relies solely on institutional initiative [5,6]. Linking SBE competencies to medical education performance expectations or accreditation criteria would distribute both mandate and resources across the medical education system – converting the competitive institutional landscape from a fragmentation risk into a coordination mechanism.

To advance SBE from early DOI stages (Knowledge/Persuasion, 84% of studies) towards sustained integration (Decision/Implementation/Confirmation, 16%), we propose consolidated, evidence-based recommendations distinguishing near-term strategies from systemic reforms. These target Pakistan’s barriers (e.g. funding and infrastructure in 65% of studies) and enablers (e.g. low-cost tools in 25%), align with DOI attributes (Table 2) and mirror challenges in other LMICs, e.g. India, Indonesia and sub-Saharan countries. National accrediting/educational bodies (e.g. PMDC/HEC in Pakistan) should lead, ensuring equitable trainee access and ultimate patient safety via Level 3+ transfer.