In-situ cardiac arrest simulations in a tertiary-care hospital in Pakistan: a feasibility study exploring challenges and future directions

Abstract

Background

In-hospital cardiac arrest (IHCA) survival remains poor in low-resource settings, partly due to skill decay, delayed responses, and inconsistent adherence to resuscitation guidelines. In-situ simulation has been proposed as a strategy to improve resuscitation performance and identify system gaps, but evidence from low- and middle-income countries is limited.

Objective

To assess the feasibility and implementation outcomes of an unannounced in-situ cardiac arrest simulation program in a tertiary care hospital in Pakistan, and to determine whether key IHCA processes and performance metrics can be reliably measured using a structured documentation tool.

Methods

We conducted a prospective, non-randomized feasibility study at a 710-bed academic hospital from December 2023 to March 2025. Unannounced in-situ cardiac arrest simulations were conducted 1–2 times per month across multiple hospital units. Participants included resident physicians, nurses, and rapid response team members with current American Heart Association (AHA) BLS/ACLS certification. Simulations used real clinical equipment and a high-fidelity manikin, followed by structured debriefing when feasible. Outcomes were evaluated using Proctor et al.’s implementation framework, focusing on feasibility, acceptability, penetration, fidelity, and sustainability. Clinical performance metrics were collected as secondary process measures.

Results

Fifty-one simulations were conducted; data from 44 were analyzed. Feasibility and penetration were high, with simulations successfully integrated across diverse clinical areas. Acceptability was strong, with participants rating simulations as realistic and educationally valuable (mean scores 4.2–4.6/5). Fidelity was variable, particularly for debriefing, which was fully completed in 50% of applicable simulations. Sustainability challenges included competing clinical demands and lack of protected time.

Conclusions

In-situ cardiac arrest simulation is feasible and acceptable in a low-resource hospital setting and enables systematic assessment of resuscitation processes. Sustained impact will require institutional support, protected time for debriefing, and integration into ongoing quality improvement efforts.

Introduction

Background

Despite advances in resuscitation methods in the last decade, in hospital cardiac arrest (IHCA) is associated with poor survival rates, especially in low resource settings. One study from our center, a tertiary care hospital in Pakistan revealed a survival rate of 28.3% for adults and 17.9% for pediatric patients.¹ The 'chain of survival' is a foundational concept for organizing systems of care, emphasizing the timely execution of key steps: recognition and activation, early cardiopulmonary resuscitation (CPR), rapid defibrillation, advanced resuscitation, and integrated post-cardiac arrest care aimed at recovery and quality of life.2, 3 Implementation of this chain requires multiple healthcare providers working in coordination.⁴ Hence, the training of these providers in resuscitation is essential to improving outcomes, as emphasized in the 2025 international resuscitation education guidelines.5, 6, 7 However, not only has decay in skill been reported to occur faster than decay in knowledge over time,⁸ but also advanced life support (ALS) skills have been shown to decrease faster than basic life support (BLS) skills, with 14% vs 58% retention after 12 months of training.⁹ As a result, existing literature advocates for designing resuscitation programs that are interprofessional, evidence-based (aligned with contemporary international guidelines from bodies such as International Liaison Committee on Resuscitation, ILCOR, American Heart Association, AHA, and European Resuscitation Council, ERC), utilize simulation and debriefings, and are sustainable.¹⁰ A recent international consensus outlined ten critical steps for improving IHCA quality of care and outcomes, highlighting the importance of measuring performance, providing structured debriefing, and fostering a culture of continuous learning.¹¹ In-situ simulation can be a powerful tool in all these areas.

A recent systematic review suggests that in situ simulation should be considered for resuscitation training.¹² In situ simulation is defined by the society for simulation in healthcare as the training that takes place in an actual patient care setting in an effort to achieve a high level of fidelity and realism.¹³ In-situ simulation does not only refer to a clinical environment, but should also be integrated with the staff who works there with accessible information and technology.¹⁴ In situ simulation offers advantages over traditional center-based simulation training, by promoting experiential learning closely aligned with healthcare providers’ actual work and improving training efficiency for institutions and participants.¹⁵ In situ simulation can also be used to analyze and improve the quality of patient care, to identify latent safety threats in actual healthcare environments, and to test new facilities.16, 17 However, further research is warranted to assess the feasibility and impact of such programs in low resource setting.¹² Earlier investigations assessing the feasibility of in-situ simulation in low resource setting have emphasized the need for ongoing or repeated simulation training, rather than a one-time session.¹⁸

Intervention and intended improvement

We implemented an in-situ cardiac arrest simulation program at the Aga Khan University Hospital in Karachi, Pakistan. By utilizing these in-situ cardiac arrest simulations to address IHCAs, we created a model to assess resuscitation performance and discuss measures for improvement. The objectives of our study were to (a) assess the feasibility of introducing in-situ simulation for cardiac arrest practices before planning a larger study and (b) assess if current IHCA practices (such as response times and staff competency) are measurable on the basis of a purpose build in-situ cardiac arrest simulation documentation form. Clinical performance metrics (e.g., CPR quality, guideline adherence) were collected as secondary process measures to inform debriefing and program refinement.

Methods

Study design, setting and participants

This feasibility study had a non-randomized design. It was conducted at the Aga Khan University Hospital (AKUH) in Karachi, Pakistan. AKUH has a total capacity of 710 beds including 137 beds for adult specialized (monitored) care and 28 beds for adult intensive care units. An organizational culture of patient safety is overseen by the Center for Patient Safety and the Patient Quality and Safety Department. AKUH runs a dedicated Rapid Response Team (RRT) that addresses the medical needs of all patients throughout the hospital, excluding those in the Intensive Care Units (ICUs). The primary goal of the RRT is to respond within a 5-min timeframe, assess patients using modified early warning signs and actively engage in resuscitation efforts. Additionally, a dedicated 24/7 Cardiac arrest Team, comprising Anesthesia, Cardiology, and Internal Medicine teams, ensures continuous cardiac arrest coverage. The Cardiac arrest Team remains involved until the Return of Spontaneous Circulation (ROSC) or the declaration of death. Detailed documentation of each Cardiac arrest event occurs through a specialized CPR, capturing patient demographics, existing medical conditions, recent RRT utilization, and the timeline for Cardiac arrest Team arrival within a 5-min.

The baseline population consisted of resident physicians (working in departments of medicine, cardiology, and anesthesiology) and rapid response team nurses and staff assigned on duty in the intervention area. All participants were required to have current American Heart Association (AHA) BLS certification, while some team members, such as designated team leaders, held AHA Advanced Cardiac Life Support (ACLS) certification. The intervention was implemented from December 2023 to March 2025. This study was designed and reported in accordance with the SQUIRE-SIM guidelines, with the understanding that its primary focus was on evaluating implementation rather than testing iterative system changes.¹⁹

Implementation

We implemented an in situ program consisting of cardiac arrests simulations. Each in-situ cardiac arrest simulation was facilitated and observed by an AHA instructor on a simulated case scenario in a clinical area of the hospital. Simulations were conducted 1–2 times per month, at various locations, during different times of the day and on different days of the week and weekends. The participants involved were not informed about the time of the simulated in-situ cardiac arrest simulation in order to maintain the element of surprise. The event was not videotaped. A total of 16 realistic scenarios were used over the course of the program; these were developed by the senior author (anesthesiologist) with expertise in resuscitation and simulation, and were reviewed for clinical accuracy and relevance by a multidisciplinary panel including internal medicine specialists. We designed our simulation cases to reflect variations in pathophysiology that come into play when managing cardiac arrest and to highlight scenarios that are of clinical relevance in the hospital setting.

A vacant patient bed in the selected area was chosen, and a CodeBlue3 manikin²⁰ was placed on the bed. The hospital’s rush call paging system was used to generate the cardiac arrest. Upon the arrival of the first participant, a scenario was narrated to them, for example “Mr. X, is a 73-year-old male patient, and has no pulse or respiration hence a cardiac arrest was generated”. The arriving team was instructed to self-organize and identify a team leader from among themselves, as would occur in a real event. Team roles were then assigned by this leader amongst the participants. Following cardiac rhythm interpretation, CPR and medical treatment were initiated according to the scenario and in compliance with the AHA guidelines. All medical equipment used during the scenario, such as intravenous medications (e.g. epinephrine), and defibrillators were the unit’s actual clinical equipment.

Each training session comprised 30 min to 1 h, with 20–30 min of interprofessional simulation followed by 15–30 min of structured debriefing. Debriefing was conducted in an adjacent ward discussion room or at the site of the simulation. It was facilitated by nurses who were experienced clinical simulation educators as well as AHA certified ACLS instructors routinely involved in assessment and feedback. The facilitators utilized the AHA promoted GAS model for debriefing.²¹ Evidence-based management of the case, compliance with the latest AHA guidelines and teamwork were emphasized.² Given the variety of experience among all the staff members, every in-situ cardiac arrest simulation stimulated new questions on different topics. The facilitator reviewed the sequence of events, highlighted the positive actions taken and informed the teams of potential areas of improvement. The debriefing concluded with a facilitator-led synthesis of key discussion points, allowing participants to reflect on their learning in relation to the intended objectives.

Variables and data measurement

Clinical performance metrics and team interaction during resuscitation were assessed using a purpose built in-situ cardiac arrest simulation documentation form based on the AHA guidelines.²⁰ This form was reviewed for content validity by the resuscitation committee at our institution. One research team member was designated as the scribe and could move around the room as needed. The scribe documented as much information as possible during the simulation to avoid relying on recall. The institutional CPR form was used to record data regarding the techniques. Variables recorded included patient status (including vital signs, rhythm, and responsiveness), time elapsed until start of CPR, CPR quality (including interruptions and rate), time to defibrillation for pulseless VT or VF, time to epinephrine for PEA or asystole and details about medications used (including dose and patient response). Sensors in the CodeBlue 3 manikin allowed us to record variables pertaining to chest compression and ventilation. During the debriefing, the AHA GAS model²¹ and AHA mega code checklist²² were utilized. A post-simulation feedback survey was administered to participants following the debriefing to assess acceptability and perceived learning. The survey used a 5-point Likert scale (1 = Strongly Disagree to 5 = Strongly Agree) and included items on realism, educational value, and psychological safety. To manage inherent variability across simulations, core scenario elements (e.g., initial rhythm, required interventions) were standardized within each scenario type, and all simulations were observed and documented using the same structured tool by a trained member of the research team.

Evaluation framework

We adopted Proctor et al.’s taxonomy of implementation outcomes to structure our evaluation, moving beyond a descriptive research narrative.²³ We focused on feasibility (the extent to which the in-situ simulation program could be successfully carried out within our hospital), acceptability (the perception among stakeholders that the in-situ simulation program was satisfactory), sustainability (the extent to which the program was maintained within the hospital’s ongoing operations), fidelity (the degree to which simulations were implemented as intended) and penetration (the integration of the program across various hospital units). Additional clinical performance metrics (e.g., CPR quality, response times) served as process measures to inform debriefing and program refinement, not as primary research outcomes.

Ethical consideration

This study was conducted in compliance with the principles of the Declaration of Helsinki, the principles of Good Clinical Practice and was granted exemption by the Aga Khan University Ethical Review Committee, IRB number 2024-10632-32694. Consent was obtained from each participant. The benefits to the participants included exposure to a new learning environment and being made to reflect on their performance, leading to them gaining new knowledge on CPR. Possible risks included the participants encountering a situation that was emotionally or mentally overwhelming for them. To overcome this challenge and to ensure psychological safety, we included simulation experts in our team. The study protocol was designed to minimize disruption to clinical care. Simulations were scheduled in consultation with a unit nurse to avoid periods of known high acuity. A fundamental rule was that any participant could immediately leave the simulation to attend to a real clinical emergency, at which point the simulation would be terminated. This contingency was factored into our feasibility assessment, and such interruptions were recorded as part of the logistical challenges of implementing in-situ simulation in an active clinical environment.

Results

The results of our study reveal critical insights into both the strengths and limitations of implementing cardiac arrest training in a low resource setting. As a feasibility study, we successfully demonstrated the logistical capacity to conduct 51 unannounced simulations across diverse clinical and non-clinical units between December 2023 and March 2025. Other studies have reported more ambitious training frequencies, such as over twice per week.²⁴ After excluding simulations with incomplete or missing data, the data of 44 simulations were included in our results. These in-situ cardiac arrest simulations were performed in a variety of settings inside the hospital, as shown in Fig. 1. Each simulation had up to six participants. Participants belonged to one of the following specialties: internal medicine, cardiology, anesthesiology, primary nursing, primary physician or rapid response team. 63 participants completed the post simulation feedback survey, with the results summarized in Table 1. Clinical performance metrics varied, as documented in Table 2 and potential explanations for observed performance gaps are addressed in the discussion.

Fig. 1.

Distribution of in-situ simulations across clinical environments.

Table 1.

Responses of 63 participants to post in-situ cardiac arrest simulation feedback survey from November 2023 to April 2024.

Question	Mean response according to the Likert scale*	Standard deviation
The simulation sessions were a realistic portrayal of clinical events	4.2	0.9
I was pressured to make decisions without enough information	3.3	1.2
The sessions were too simplistic	3.1	1.1
I did not have adequate information to make decisions	2.9	1.1
The mannequin’s responses to events were realistic	3.7	1.1
Simulation-based scenarios are a good tool for learning clinical scenarios	3.7	0.5
Simulation-based scenarios reinforce clinical concepts I have learned in ALS Training.	4.6	0.9
Simulation should be a mandatory part of continuous training.	4.6	0.5
Simulation-based scenarios were a generally useful experience	4.6	0.5
Simulation-based scenarios were a generally good experience	4.5	0.6
The sessions gave me confidence to take care of actual patients	4.5	0.6
The sessions were not very realistic	2.8	1.1
There should be more exposure to simulator during briefing	4.1	0.7
There was too much exposure to simulator during briefing	3.5	1.0
The simulation sessions were too short	2.9	1.1
The simulation sessions were too long in duration	3.0	1.0
The simulation sessions were too difficult	2.9	1.1
The debriefing sessions were generally useful	4.4	0.6
The debriefing sessions focused on systems issues and were non-judgmental	3.9	0.9
I felt embarrassed during the simulation sessions	2.7	1.1
I was self-conscious of my performance during the simulation sessions/debriefing	3.6	1.0
I found the surveys/questionnaires to be burdensome	3.2	1.1
The surveys were a helpful self-reflection tool	4.2	0.8
The surveys detracted from the educational value of the program	3.3	1.1

For clarity in future iterations, we recommend that all Likert-scale questions be phrased positively to avoid confusion in interpretation. In this table, a low score (e.g., 2.8) on a negatively phrased statement like “The sessions were not very realistic” indicates disagreement with the negative, i.e., participants found the sessions realistic.

^*Likert Scale: 5 = Strongly Agree, 4 = Agree, 3 = Neutral, 2 = Disagree, 1 = Strongly Disagree.

Table 2.

Clinical performance metrics.

Clinical performance metric	Applicable n (%)	Criteria met n (%)
Cardiology team arrival with 3–5 min of the rush call generated	37/44 (84.1%)	15/37 (40.5%)
Anesthesia team arrival with 3–5 min of the rush call generated	39/44 (88.6%)	26/39 (66.7%)
Medicine team arrival with 3–5 min of the rush call generated	37/44 (84.1%)	27/37 (73.0%)
Primary team arrival	20/44 (45.5%)	13/20 (65.0%)
The algorithm as per the AHA guidelines was followed	42/44 (95.5%)	6/42 (14.3%)
Chest compression fraction (>81%)	41/44 (93.2%)	17/41 (41.5%)
Early defibrillation where required	41/44 (93.2%)	9/41 (22.0%)
When ROSC is obtained the recommended immediate care	44/44 (100.0%)	35/44 (79.5%)
Timely documentation initiated on CPR record form	44/44 (100.0%)	23/44 (52.3%)
Infection control practices were followed during the code (where applicable)	44/44 (100.0%)	16/44 (36.4%)

In this study, in-situ simulation functioned primarily as an assessment tool within an implementation framework. While debriefings were structured using the GAS model and included discussion of systems issues (Table 3), the primary mechanism for capturing system-level data was the structured observer documentation form. Insights into equipment failures, role confusion, and workflow barriers were systematically recorded. These identified latent safety threats and system gaps represent a critical output of the feasibility phase, generating a prioritized list of targets for subsequent, dedicated QI cycles.

Table 3.

Strengths and learning needs identified during debriefing.

Strengths identified	Learning needs identified
Clear role identification and good team co-ordination	Infection control practices should be emphasized (e.g. gloves were not worn)
Communication being performed in a close loop manner	Rhythm should be identified correctly (e.g. SVT was identified as Sinus tachycardia)
Team members discussing and intervening for underlying causes of cardiac arrest	Team should know the difference of anesthesia rush call, general rush call, and cardiology rush call
Appropriate staffing with staff following team leader's prompts	Chest rise and fall should be assessed with each ventilation
Crash cart maintenance with all required equipment available	CPR form was not available
Adequate compression rate	Rebreathing bag was nonfunctioning, no alternative used
ROSC management taken by whole team; post ROSC care verbalized	Defibrillator did not work during half of the effort
Infection control policies were followed	Recalling AHA guidelines
	Nurse empowerment to lead and initiate interventions
	Coordination needs improvement in drug administration (Read back and Repeat Back)
	Monitor attachment was delayed
	Initial shock was not given
	Transcutaneous pacing leads were not present in the crash cart
	Ventilation rate was inadequate (e.g. 77%)
	Combo pads were attached on defibrillator leads

Implementation outcomes

Feasibility and penetration

Logistically, conducting 51 in-situ simulations over 15 months across nearly all hospital wards (Medicine, Surgery, Cardiology, Radiology, Emergency Department) proved achievable. Penetration was broad, demonstrating that the model could be integrated into diverse clinical environments. Key feasibility challenges included securing a portable high-fidelity manikin, coordinating schedules to avoid clinical care disruptions, and managing the “element of surprise” without causing undue stress. The need for a dedicated simulation operations specialist became apparent to manage equipment and logistics.

Acceptability

Participant feedback was collected via post-simulation surveys (n = 63). Responses indicated high acceptability. Participants strongly agreed that simulations were a realistic (mean 4.2 ± 0.9/5) and useful (mean 4.6 ± 0.5/5) learning experience that reinforced clinical concepts and increased confidence in managing real patients (mean 4.5 ± 0.6/5). Notably, most participants believed simulation should be a mandatory part of continuous training (mean 4.6 ± 0.5/5). However, some feedback highlighted areas for improvement, such as the desire for more pre-briefing on simulator functionality (mean 4.1 ± 0.7/5).

Fidelity of implementation

A critical finding was the variable fidelity of the debriefing component. While debriefing was planned for all simulations, it was fully executed in only 50.0% of cases (21/42 applicable simulations). Barriers to debriefing fidelity included immediate clinical demands, staff being called away, and time constraints. This inconsistency represents a significant compromise in the educational cycle and underscores the challenge of protecting time for reflective learning in a busy clinical environment.

Sustainability

Program sustainability emerged as a central challenge. While the initiative was initially supported by project enthusiasm and leadership backing, long-term sustainability requires dedicated institutional resources. This includes funding for equipment maintenance and replacement, protected time for simulation educators and participants, and integration of in-situ simulation responsibilities into job descriptions. The variable debriefing rate is a key indicator of sustainability risk, highlighting the tension between operational clinical duties and training commitments.

Discussion

As a feasibility study, the primary objective was not to determine whether clinical performance met external benchmarks, but rather to evaluate whether key resuscitation process measures could be reliably captured during unannounced in-situ simulations and subsequently used to inform learning and system improvement. The successful collection of CPR quality metrics, teamwork observations, and participant feedback across diverse clinical units demonstrates that structured measurement is achievable without major operational disruption. Variability in performance, therefore, should be interpreted as evidence of the measurement system’s sensitivity rather than as definitive assessment of clinical quality.

Interpretation of observed performance variability

While variability in algorithm adherence, CPR quality metrics, and time-sensitive actions was observed, these findings should be interpreted within the contextual realities of in-situ simulation in a tertiary care setting. Several plausible contributors may explain these results. First, variability in guideline adherence may reflect differences in team familiarity with ACLS updates, particularly among residents and nursing staff with heterogeneous prior training exposure. Gaps in formal recertification cycles and limited ongoing skills reinforcement can result in skill decay. Although a majority of participants had ALS training, our results suggest that certification alone does not ensure practical proficiency during high-stakes resuscitation. This reinforces previous evidence that healthcare worker’s knowledge about CPR is low in general²⁵ and both cognitive and psychomotor skills pertaining to cardiac arrest resuscitation decay significantly within 6–12 months of training.²⁶ Our analysis of CPR quality metrics such as CCF, early defibrillation, and ventilation rates further illustrates this gap between knowledge and practice. 41.5% of simulations achieved the recommended CCF >81%, a finding observed in other studies as well.²⁷ Only 22.0% of applicable simulations accomplished timely defibrillation. These findings align with the literature indicating that delays in key interventions, such as early defibrillation and medication administration, are often due to both individual performance and systemic workflow inefficiencies.²⁸

Second, incomplete algorithm adherence by team members may also reflect cognitive load and leadership variability during high-acuity events. Observed communication breakdowns and task duplication suggest that structured team role assignment and closed-loop communication require reinforcement. These findings highlight modifiable system-level factors rather than individual knowledge deficits alone. Although a team leader was identified in 98% of simulations, role confusion and lack of clear communication were reported in over 40% of scenarios. Closed-loop communication, a hallmark of effective crisis resource management, was fully implemented in fewer than half of the sessions. In hierarchical healthcare cultures such as ours, common in South Asian contexts, junior team members often hesitate to speak up or assume leadership roles, especially during high-pressure situations. This cultural aspect must be considered when designing simulation programs, with deliberate efforts to promote psychological safety, flatten hierarchy, and encourage communication across all roles. Leadership, teamwork, and effective communication are key components of team performance. In situ cardiac arrest simulations may help staff be better prepared and resuscitation areas fully equipped with improved communication.

Third, delayed defibrillation and suboptimal chest compression fraction (CCF) likely reflects systems-level rather than individual deficits. Equipment accessibility (e.g., defibrillator availability, crash cart standardization), role ambiguity during early arrest phases, and environmental constraints (such as crowded wards and competing clinical demands) are well-documented contributors to resuscitation delays. The in-situ format intentionally preserved these environmental variables, thereby exposing latent safety threats that may not be apparent in simulation center based training. Low knowledge about the use of technical equipment such as defibrillators in CPR simulations has been previously reported, which may lead to defibrillators not being used appropriately.²⁹ Resource-related constraints, including malfunctioning defibrillators were rarely notable in our setting and likely contributed to these delays. Such in-situ simulations underscore the importance of routine system audits and maintenance protocols, which are often inconsistently applied in low-resource hospitals due to funding limitations and competing priorities.³⁰

Lastly, the operational challenge of completing structured debriefings in only approximately half of simulations underscores the tension between clinical workload and reflective learning time. This represents a missed opportunity for reflective learning, which is one of the most powerful tools in simulation-based education. In real-world hospital environments, protected time for post-event analysis is often limited, directly affecting educational fidelity. Barriers to debriefing in our setting included time constraints, and an underdeveloped culture of non-punitive feedback. Embedding structured debriefing as a mandatory component of in situ simulation, alongside facilitator training, could significantly strengthen the educational yield of each in-situ cardiac arrest simulation.

Strengths

There are several strengths to our study. This is one of the first prospective studies to assess the feasibility of in-situ simulation across a long period of time in low resource setting. Conducting over 50 in-situ cardiac arrest simulations across diverse hospital settings, ranging from inpatient wards to radiology and the emergency department, demonstrates that such initiatives are scalable and logistically achievable, even in resource-limited contexts. Nevertheless, sustainability remains a concern. For in situ simulation to be embedded into the institutional culture, dedicated resources—including personnel, simulation equipment, and protected time—must be allocated consistently, which can be challenging in financially constrained healthcare systems. All of our participants were actively involved in the in-situ cardiac arrest simulations, unlike some studies in which the majority of the participants were passive observers.³¹ Additionally, our participants treated the in-situ cardiac arrest simulation as a real situation, whereas participants in similar studies would verbalize actions as is often done during BLS/ALS recertification.³² The number of cardiac arrest scenarios simulated longitudinally was 16, whereas other studies of a similar scale in developed countries simulated less than 10 scenarios.24, 27 We used a manikin of the highest fidelity throughout the study, however other studies have reported using low fidelity manikins for part of their study period.³¹ Unlike other studies, which used self-reporting as a tool to measure the outcome of the in-situ simulations, we used a validated in-situ cardiac arrest simulation documentation form based on the AHA guidelines.³³ According to a review of in situ simulations by Martin et al., the limitation in the evaluation methods used in most of the studies was the subjectivity of self-reported surveys.³⁴

Limitations

Our study also has several limitations. This was a single center study with an uncontrolled study design, therefore we are unable to rule out selection bias. Robust data collection over a long period of time proved to be challenging in this study and there were more instances of missing or incomplete data than we would have liked. Debriefing was not performed after every simulation, resulting in a missed opportunity for self-reflection and closing knowledge gaps. Occasionally, the participants were able to anticipate an in-situ cardiac arrest simulation was about to be simulated in their ward which would have compromised on the element of surprise. We did not conduct in-situ cardiac arrest simulations during the evening and night, which may have provided the opportunity to even more staff members to participate. Prioritizing sustainable simulation was an oversight on our part as clinical equipment utilized during the simulation was not reused for education (after sterilization), whereas such actions have been reported in the existing literature.³¹ We did not take into account clinical workload of the participants when planning the in-situ cardiac arrest simulations; had we done so, it may have ensured dedicated and uninterrupted participation and may have avoided the instances when debriefing was not conducted.²⁴

Mitigation strategies and future directions

The primary value of identifying performance gaps lies less in benchmarking performance, and more in informing targeted system improvements. Based on findings from this feasibility phase, several mitigation strategies are planned. Collaboration with hospital leadership to ensure equipment standardization can improve crash cart organization and ensure functioning defibrillator accessibility. Engagement with active duty hospital staff to formalize brief (15–20 min) structured post-simulation debrief windows can ensure simulation participants can receive effective feedback. Continued capture of process measures to assess trends over time will be essential to assess trajectory and assess if improvement is occurring. Our experience supports the growing evidence for a 'high-frequency, low-dose' simulation training model. Conducting shorter, more frequent in-situ simulations (1–2 per month) embedded in clinical workflow may be more sustainable and effective for skill retention than intensive, infrequent courses. Future iterations of our program should aim to increase this frequency to combat the rapid decay of resuscitation skills, particularly in resource-limited settings where access to traditional courses is constrained. Utilization of rapid-cycle deliberate practice based in-situ simulation can lead to progressive acquisition of targeted resuscitation skills such as optimization of chest compression fraction and time to defibrillation.³⁵

Future directions should include prospective studies linking in-situ cardiac arrest simulation participation with real-world patient outcomes, such as survival rates from in-hospital cardiac arrest or quality of post-arrest care. Additionally, cost-effectiveness analyses are urgently needed to justify long-term investment in such efforts, especially when compared to conventional didactic or off-site simulation programs. Finally, interdisciplinary expansion to include community-based health centers could help build a more robust, system-wide emergency response culture.

Conclusion

This feasibility study demonstrates that unannounced in-situ cardiac arrest simulations can generate actionable, high-fidelity process data in a resource-limited tertiary care setting. The value of the program lies less in the absolute performance values observed and more in the establishment of a sustainable mechanism for continuous measurement, reflection, and system-level learning. However, for its full potential to be realized, such programs must be supported by systemic interventions including equipment maintenance, structured debriefing, ongoing skill reinforcement, and a culture that values continuous learning and psychological safety. Tailored approaches that address the specific needs and constraints of resource-limited settings will be essential to optimizing both the educational and clinical impact of in-situ simulation.

CRediT authorship contribution statement

Nadeem Ahmed Siddiqui: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Abdullah Saeed Khan: Writing – original draft, Supervision, Methodology, Formal analysis, Data curation. Nazish Khowaja: Project administration, Methodology, Formal analysis, Data curation. Selina Hasan: Resources, Project administration, Methodology, Investigation, Data curation. Amber Sabeen: Writing – original draft, Supervision, Data curation, Conceptualization. Rozina Roshan: Supervision, Project administration, Methodology, Investigation, Data curation. Faisal Waseem Ismail: Resources, Project administration, Methodology, Investigation, Data curation. Muhammad Faisal Khan: Writing – original draft, Supervision, Resources, Project administration, Methodology, Investigation, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary material to this article: