Mr James Ansell, BSc, MBBCh, MRCS, Royal College of Surgeons Clinical Research Fellow in Simulation, Welsh Institute for Minimal Access Therapy (WIMAT), Cardiff Medicentre, Heath Park, Cardiff CF14 4UJ, UK. E-mail:email@example.com
Aim Simulation allows the acquisition of complex skills within a safe environment. Endoscopic polypectomy has a long learning curve. Our novel polypectomy simulator may be a useful adjunct for training. The aim of this study was to assess its content validity.
Method The Welsh Institute for Minimal Access Therapy (WIMAT) endoscopy suitcase was designed to simulate colonic polypectomy. Participants from regional and national courses were recruited into the study. Each undertook a standardized simulated polypectomy and completed a seven-point Likert scale questionnaire examining its realism.
Results In all, 17 participants completed the questionnaire: 15 (88.2%) gastroenterologists, one (5.9%) colorectal surgeon and one (5.9%) experienced endoscopic nurse specialist. Of the gastroenterologists, seven (46.7%) were consultants and eight (53.3%) were senior trainees or Post CCT (Certificate of Completion of Training) fellows. The mean number of real-life polypectomies performed by the cohort was 156 (95% CI 35–355). The highest scores were for ‘mucosal realism’ (median score 6.0, P =0.001), ‘endoscopic snare control’ (median score 6.0, P =0.001), ‘handling the polyp’ (median score 6.0, P =0.001) and ‘raising mucosa’ (median score 6.0, P <0.001). Of the 15 parameters examined only three were not statistically significant in favour of the simulator. These were ‘anatomical realism of sessile polyps’, ‘resistance of scope movement’ and ‘paradoxical motion’. The overall score for the simulation was 6.0 (P <0.001). There was no significant difference between the level of difficulty of the simulator compared with real life (median score 4.0, P =0.559).
Conclusion The WIMAT colonoscopy suitcase model has excellent content validity for several parameters. This may have potential applications in medical training and assessment.
This paper details a new way of simulating endoscopic polypectomy for use in medical training. This is a novel ex vivo animal model with the potential to teach a complex procedure. The paper highlights its content validity.
Simulation is widely used in medical training and assessment. The advantage of simulation is that it enables the development of practical skills in a controlled environment. This has the potential to improve patient care via a reduction in procedural complication rates.
The first endoscopic simulators were described between 1969 and 1970 [1,2]. Since then, there have been significant developments in the field. Simulators now exist with the capacity to teach several endoscopic procedures , including upper and lower gastrointestinal endoscopy, endoscopic retrograde cholangiopancreatography and endoscopic ultrasound. There are also different types of simulators available, ranging from bench models and ex vivo animal platforms to virtual reality (VR) trainers [4–7]. For colonoscopy in particular there is a large weight of evidence which focuses on VR simulation [8–16]. Several VR validation studies have been conducted on commercially available products [8–16]. VR simulators [GI mentor II (Simbionix), Accutouch HT immersion, Olympus Endo Ts-1 2nd Generation] allow participants to practice multiple computer-based modules with varying degrees of difficulty. These simulators emphasize scope navigation and loop management with some limited capacity for therapeutic interventions. There is evidence to suggest that VR has the potential to oversimplify complex tasks [4,17–19]. This is illustrated by reports that these simulators are often unable to distinguish between participants with different levels of expertise [17–19]. Therefore, at present their most effective use could be for junior trainees with minimal experience. Others have questioned the usefulness of the multiple parameters that the VR simulators measure and their relevance to evaluating performance .
An alternative approach to VR is the use of ex vivo animal tissue simulation. Sedlack et al.  describe the validation of a novel bovine colonoscopy simulation for use in skills assessment and as an adjunct to senior training. The paper focuses on the skill of performing a colonoscopy but does not consider the simulation of therapeutic measures. There is a lack of current evidence for the validation of ex vivo animal models that focus on the area of therapeutic colonoscopic intervention.
Simulators should undergo a formal validation process before widespread adoption in training or assessment . This allows inferences to be made regarding effectiveness and in justifying its investment . The first stage of any validation process is to establish a construct . This defines exactly what needs to be examined by a new training tool . In colonoscopy, for example, the construct should clarify whether the simulator is being validated to assess or to teach its users . The next step is to establish face and content validity. This is usually assessed by surveying experts regarding a given simulator’s realism. Following this, the simulator should be assessed on its ability to distinguish between levels of expertise. This provides evidence for its construct validity. The final element to prove is criterion validity. This includes predictive validity (the ability of a tool to predict future performance) and concurrent validity (the correlation between the assessment tool and the ‘gold standard’) .
The aim of this study was develop an ex vivo animal model to simulate the procedure of colonic polypectomy. The Welsh Institute for Minimal Access Therapy (WIMAT) colonoscopy suitcase model is a novel porcine simulator which allows the participant to practice snare polypectomy of sessile and pedunculated (bleeding and non-bleeding) polyps. We report the evidence for its face and content validity.
The WIMAT colonoscopy suitcase was developed at the WIMAT centre. Frozen porcine colonic specimens were sourced from a regional company (Fresh Tissue Supplies Ltd, Heathfield, East Sussex, UK) . These originated from low risk category three animal by-products . All animal samples were handled and disposed of according to a strict internal protocol. The specimens were initially defrosted and everted (Fig. 1a) to expose the mucosa, and the internal aspect of the colon underwent a standardized cleaning process. Three types of polyps were then constructed: sessile, pedunculated non-bleeding and pedunculated bleeding polyps. Sessile polyps were created by injecting a standardized volume of a polyp mix into the bowel submucosal layer. The polyp mix is a solution that solidifies at room temperature and does not break down when refrozen. The mix was warmed using a T.ARE heating magnetic stirrer (VELP®) to a temperature of 90°C . Once liquefied, it was injected and cooled in situ to seal its position under the mucosa as a sessile polyp. The pedunculated non-bleeding polyps were constructed by using a thin layer of sausage skin which was filled with a standardized volume of liquid polyp mix (Fig. 1d) and allowed to solidify. To create the pedunculated bleeding polyps this process was repeated and, in addition, a standardized length of porcine ureter was attached inside each polyp and cannulated with a small plastic catheter (Fig. 1c and e). This catheter was subsequently attached to a 50 ml Luer-Lock syringe containing simulated blood. All pedunculated polyps were attached to the bowel mucosa in a standardized way producing the view illustrated by Fig. 1a. The whole specimen was then inverted so that the polyps were transferred to the internal aspect of the lumen of the bowel (Fig. 1b).
The simulator casing
The porcine bowel was housed in a portable polymer suitcase (Storm Case Im2600, Hardigg®) with a hole made in one end to simulate the anus (Fig. 2a) . This hole was cannulated with a 15 mm Ethicon XL port (Johnson and Johnson) and secured internally with a plastic ring clip. This allowed the passage of the colonoscope into the suitcase and provided an airtight seal to enable insufflation of the colon. Inside the suitcase we placed a removable metal mesh base. This accommodated a crocodile clip which was connected to a diathermy unit, for use during the simulated polypectomy. Foam segmentors were mounted onto metal rails and the whole device was secured with wing nuts onto the mesh base (Fig. 2). The porcine bowel was then passed through the inside of the foam segmentors and a curved piece of 55 mm standard exhaust piping and then through a second set of segmentors (Fig. 2b) (this represented the sigmoid bend). The anal end of the specimen was attached to the Ethicon XL port using cable ties and the oral end was secured airtight. When the bowel was inflated, the foam segmentors indented on the serosal aspect of the bowel (to represent haustral folds) to give it a realistic luminal appearance (Fig. 3).
A standard Olympus Cf-Q140L colonoscope, CLV-U40 light, CV-240 processor and OEV 203 monitor were used for each procedure (Fig. 2e). All endoscopy equipment was dedicated to animal endoscopy teaching alone. Standardized endoscopic 25 mm snares (AcuSnare®) were supplied by Cook® (Wilson Cook Medical GI Endoscopy) . The energy source was from Valley LabTM EZc and was placed on a cutting setting of 40 W . If a mucosal lifting agent was required we used a water based dyed substance which was developed at the training centre. This contrasted with the colour of the polyps.
Participants were recruited from regional and national endoscopic training courses where the simulator was being demonstrated. All participants were experienced in the skill of colonic polypectomy. Each completed a snare polypectomy on a simulated pedunculated (bleeding/non-bleeding) and sessile polyp. All simulators and endoscopic equipment were standardized throughout. Following the procedure, each participant was asked to complete a 15-question realism survey based on a seven-point Likert scale. The questionnaire was adapted from that of Sedlack et al.  and reconstructed according to expert opinion at our research centre. Questions 1–13 of the survey were divided into the following three areas: visual realism, anatomical realism, mechanical realism (1 = strongly disagree, 4 = neutral, 7 = strongly agree). Question 14 focused on the overall degree of similarity between the simulated polypectomy and ‘real-life’ polypectomy (1 = strongly disagree, 4 = neutral, 7 = strongly agree). Question 15 compared the technical difficulty of human polypectomy with the simulation (1 = much easier, 4 = same, 7 = much more difficult).
Data analysis and power calculations
Assuming that the seven-point scales had a standard deviation of 1.0, 17 participants would give > 90% power to detect a difference of 1 point or more on the survey scale against a hypothetical mean of 4. The realism surveys were analysed using the Wilcoxon signed-rank test on a PASW Statistics 18 (spss) for non-parametric data. Median values from the seven-point scale were compared with a hypothetical mean of 4 to determine statistical significance.
A total of 17 participants (male: female ratio 14:3) completed the questionnaire: 15 (88.2%) gastroenterologists, 1 (5.9%) colorectal surgeon and one (5.9%) experienced endoscopic nurse specialist. Of the gastroenterologists, 7 (46.7%) were consultants, 8 (53.3%) were ST6-7 level or Post CCT fellows (Table 1). All participants were experienced in performing colonoscopy, polypectomy and polyp biopsies. The mean numbers of previous procedures performed by the cohort were 371 (95% CI 179–689) colonoscopies, 156 (95% CI 35–355) polypectomies and 165 (95% CI 42–360) biopsies (Table 1). The majority of the cohort had previous experience of using several different polypectomy simulators (Table 1).
Table 1. Experience levels of participants.
Level of participant
Number of participants (%)
Type of procedure
Number of real life procedures performed (mean with 95% CI)
Type of simulation
Average previous number of times simulators used by delegates (mean with 95% CI)
The highest scores were for ‘mucosal realism’ (median score 6.0, P =0.001), ‘endoscopic snare control’ (median score 6.0, P =0.001), ‘handling the polyp’ (median score 6, P =0.001) and ‘raising mucosa’ (median score 6.0, P <0.001) (Table 2). Six parameters scored a median score of 5 with statistically significant results (Table 2). These were ‘endoscopic view’ (P =0.001), ‘polyp realism’ (P <0.001), ‘bleeding realism’ (P =0.013), ‘haustral folds’ (P =0.029), ‘anatomical realism of pedunculated polyps’ (P =0.01) and ‘diathermy of the polyp’ (P =0.026) (Table 2). Of the 15 parameters examined only three were not statistically significant in favour of the simulator. These were ‘anatomical realism of sessile polyps’ (P =0.08), ‘resistance of scope movement’ (P =0.406) and ‘paradoxical motion’ (P =0.055).
Table 2. Results of realism survey [median scores, 25%–75% interquartile range (IQR) for realism parameters using a seven-point Likert scale; Wilcoxon sign-rank testing to compare actual median with a hypothetical median neutral score of 4, P <0.05].
Median score (IQR) (n =17)
Resistance to scope
Handling the polyp
Diathermy of polyp
Difficulty compared with reality
The overall score for the simulation was statistically significant compared with a neutral score (median score 6.0, P <0.001). When participants were asked to compare the level of difficulty of the simulator compared with real life the result was not significantly different (median score 4, P =0.559).
The simulation of practical procedures may be a valuable adjunct for medical training, particularly in the light of reduced working hours . In colonoscopy, simulation is a growing field . In recent years more research to support the use of ex vivo animal models has been published [4,29].
The results of this study have demonstrated that the WIMAT colonoscopy suitcase has excellent face and content validity across a range of parameters. A cohort of participants, experienced in the skill of colonoscopy and polypectomy, awarded the model favourable scores for visual, anatomical and mechanical realism. All of the measured parameters for visual realism scored well enough to produce a statistically significant result. Most encouragingly was the statistically significant score for the overall realism of the model and the non-statistically significant score comparing the difficulty of the simulation to actual reality. For anatomical realism, it was interesting to note that the scoring for simulated pedunculated polyps was more favourable than that for the simulated sessile polyps. However, when participants were asked to comment on the realism of performing a mucosal lift on a sessile polyp, a statistically significant favourable result was achieved. This would imply that the reduced level of anatomical realism of the sessile polyp did not significantly impact on the process of performing the polypectomy. The other non-statistically significant parameters were for ‘resistance to scope movement’ and ‘paradoxical motion’. We would agree that this may in fact be a limitation of the current model. However, this should not significantly affect the use of the simulator which is designed to focus on polypectomy training as opposed to navigation and endoscopic steering.
There are several proposed benefits of using our novel ex vivo animal tissue polypectomy trainer. First, the cost of the simulator is considerably less than VR, live animal and cadaveric models. This makes it a financially viable option for training large numbers of participants with varying levels of experience. Also, the model is portable, has a simple set-up process which requires little technical expertise and can be tailored according to the level of experience of the user. This means that it can be utilized at any training centre with minimal inconvenience to faculty and course administrators. Limitations of this model are that it is single use and requires a time consuming process of polyp insertion. Furthermore, porcine bowel can also be relatively thin, risking perforation and desufflation during the simulation. We have overcome this by using rectal tissue which is much thicker than other parts of the porcine large intestine. The model can be quickly and easily patched should a perforation occur.
In conclusion, this paper highlights the benefits of ex vivo animal simulation and introduces our novel porcine simulator. We have confirmed the face and content validity of this model. Future work will focus on demonstrating its construct and concurrent validity and on testing the capacity of the model to allow transfer of skills into reality.
Neil Warren, Roger Leicester, Sunil Dolwani, Neil Hawkes, Stuart Goddard and Konstantinos Arnaoutakis designed and constructed the WIMAT endoscopy suitcase. James Ansell and Jared Torkington have no conflicts of interest or financial ties to disclose.