British Journal of Neurosurgery, December 2014; 28(6): 707–712 © 2014 The Neurosurgical Foundation ISSN: 0268-8697 print / ISSN 1360-046X online DOI: 10.3109/02688697.2014.913775
ORIGINAL ARTICLE
Development of a modelled anatomical replica for training young neurosurgeons Claudia Craven1,2, David Baxter2,3, Martyn Cooke2, Lydia Carline2,4, Samuel J. M. M. Alberti2, Jonathan Beard2 & Mary Murphy2 1Department of Neurosurgery, Addenbrooke’s Hospital, Cambridge, UK, 2Royal College of Surgeons England, London, UK, 3Department of Medicine, Imperial College London, London, UK, and 4School of Fine Art, Royal College of Art, London, UK
Abstract Introduction. The Modelled Anatomical Replica for Training Young Neurosurgeons (MARTYN) is a novel simulation model developed by the Royal College of Surgeons England (RCSEng). This study describes the development of the model and aims to determine its feasibility as a potential future training tool. Methods and materials. Traditional model-making methods were used to develop a prototype. Initial procedural trials tested the feasibility of the model. Eighteen participants, grouped by experience (nine novices, four intermediates and five experienced), completed two tasks: a craniotomy and a burr hole followed by insertion of an external ventricular drain (EVD). Subjective data on confidence, usefulness, realism and preference to other training modalities were collected via a standardised questionnaire and a 5-point Likert scale. Results. Preliminary trials of the model prototype demonstrated feasibility. The novice group had the greatest self-reported benefit from MARTYN training, with significant increases in self-rated confidence in both the craniotomy (p ⬍ 0.01) and EVD insertion (p ⬍ 0.05) procedures. MARTYN was reported to having good visual and tactile realism overall with the bone component being considered highly realistic. The model was reported to be a useful training tool. When asked to rank preferred training modalities, operative experience was chosen first with cadaveric training and MARTYN consistently scoring a second choice. Conclusions. MARTYN was developed with the intention to fill the current niche for an inexpensive synthetic model head. This study shows that the use of MARTYN for training is both feasible and realistic. We demonstrate a preliminary face and construct validity of the model in this pilot study. With the reduction in working hours, we believe this model will be a suitable supplement to the current ST 1–3 level cadaveric training and will have a positive impact on patient safety.
Introduction Current surgical trainees are receiving less exposure than in the past throughout all stages of their careers as a result of contemporary approaches to hospital service delivery, consultant-driven care and rigid implementation of the European Working Time Directive.1 In order to help mitigate problems associated with lack of operative experience, it is essential that trainees should have access to a variety of different training tools outside of the operating theatre to allow them to develop the skills necessary for future patient care. Virtual reality and haptic simulators hold great promise, with rapid development with respect to both visual and tactile realism.2 Nonetheless, their use is associated with considerable expense and it remains difficult to emulate the tactile experience using haptics alone. Although cadavers have traditionally been utilized to fill this void, they are disadvantaged due to the difficulty in replicating a bleed. Aboud et al.3 devised an ingenious solution by vascularising fresh cadaveric heads appropriate for practising a variety of neurosurgical procedures such as aneurysm repair, craniotomies, skull base approaches, tumour resection and ventricular endoscopy; this approach has since been applied to fixed cadaveric specimens.4 Nonetheless, cadaveric use in this regard has been low in the UK compared to other countries due to possible lack of resources and a requirement for pre-mortem consent.5 Simulated surgery using models provides a valid alternative. In 1998, Abe et al. developed skulls, produced from bone-like polyamid-nylon and glass beads, containing realistic brain-like tissue, dura and cavernous sinuses.6 They used patient computed tomography scans and stereolithography for model construction, thus opening up the possibility of using these models for surgical planning in addition to a training function. Ono and Co.7 developed the Kezlex model; arguably the most advanced model of its kind. The model costs upwards of £1,0007 and whilst such realistic models are
Keywords: model; neurosurgery; simulation; training
Correspondence: Dr Claudia Craven, Foundation Year 2 Doctor, Department of Neurosurgery, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 0QQ, UK. Tel: ⫹ 0044745436688. E-mail: [email protected] Received for publication 7 October 2013; accepted 6 April 2014
707
708
C. Craven et al.
ideal for the experienced surgeon, there is currently not an equivalent model for the needs of the junior trainee. We have developed a Modelled Anatomical Replica for Training Young Neurosurgeons (MARTYN) with the intention to fill the current niche for a distributable and inexpensive synthetic model head. MARTYN was developed at the Royal College of Surgeons of England (RCSEng) for junior neurosurgical trainees to practice surgical skills outside of working hours, with the aim of having a positive impact on patient safety. This paper describes the development of this model, the procedures that can be practised on the current prototypes and its preliminary subjective validation as a training tool.
Methods and materials Our aim was to construct a realistic but simple simulation model containing salient structures such as temporalis muscle, bone, dura mater, cerebrospinal fluid, brain and ventricles.
Materials Traditional moulding methods of model development were used to develop models. The skull and brain model components were cast in silicone from treated cadaveric specimens obtained from the RCSEng Wellcome Museum collection. The brain mould included removable clay insets to allow space for the lateral ventricles; the brain was cast using a gelatin composite base. Paraffin cerebro-spinal fluid (to prevent gelatin dissolution) and was injected in the lateral ventricles. Skulls were cast in thickened polyurethane resin in halves and later fixed around the brain using resin. The latex dura mater was painted to the inside of the cast
skulls. The temporalis was handcrafted in clay and then cast. Originally produced from gelatine, the second-generation temporalis was re-designed in silicone.
Procedural trials At least 40 operations were performed using the MARTYN model for training purposes; with 30 of them as part of the ‘Head Injury and Integrated Approach’ course conducted at RCSEng. Procedures performed on this model include frontal and temporal craniotomies and insertion of external ventricular drains (EVD) via burr holes.
Preliminary validation We ran a preliminary study of MARTYN models with an aim to determine the feasibility of this model as a potential training tool and establish face and construct validity. Eighteen participants; nine novices (ST 1 neurosurgical trainees and non-neurosurgeons, having done less than 10 burr holes no operative experience), four intermediates (ST 2–3 neurosurgical trainees or those who have done more than 10 burr holes and craniotomies) and five experienced trainees (those who had completed more than 30 craniotomies, which included senior registrars and consultants) completed two tasks: a craniotomy and a burr hole followed by the insertion of an EVD. Subjective data were collected via a standardised questionnaire and a 5-point Likert Scale. The questionnaire evaluated subjective confidence rating before and after MARTYN training and realism (both tactile and visual). The results were analysed using a Wilcoxon matched-pairs signed rank test and presented as the median. A p-value of less than 0.05 was considered as significant. Subjective opinions of the model overall (including its usefulness) were collected and responses were coded.
Fig. 1. Two training procedures performed on MARTYN models (A) Temporal craniotomy with revealing of the latex dura mater. (B) Reflection of dura and cortex exposure. (C) External ventricular drain insertion followed by (D) Aspiration of cerebo-spinal fluid from the left lateral ventricle (with CSF dyed red for demonstration purposes). Images courtesy of John Carr, RCSEng.
Developing MARTYN Finally, we asked all individuals to rank how the MARTYN model compares in preference to four other major training modalities; direct operative experience, cadaveric-based training, haptic trainers and other.
Results Procedural trials The model was successfully used for a variety of skills training including patient positioning and approach, skin flaps and suturing, craniotomy using both high-speed drills, perforator drill and craniotomes, dural opening, bone flap replacement, EVD and CSF aspiration (Fig. 1). The second-generation models had a temporalis made from silicone rather than a gelatine composite. We found this improved realism and allowed the muscle to be retracted without tearing. The thickened polyurethaneresin used for the skull accurately mimicked real bone as did the latex dura mater, which replicated the egg-shell peeling sensation encountered when separating real dura from bone. Similarly to actual surgery, plentiful irrigation and aspiration were beneficial during drilling as it prevented the collection of plastic bone dust and enhanced the realism of the model.
Novice Craniotomy (n=9)
Preliminary validation Confidence was assessed for two procedures, an emergency temporal craniotomy and the insertion of an EVD. The trainees were asked to rate their confidence (1 ⫽ Not confident and 5 ⫽ Confident) in both procedures, prior to training on MARTYN and after. We observed changes in selfassessed confidence ratings in the three groups: ‘novices’, ‘intermediates’ and ‘experienced’ both before and after the use of the MARTYN model. All individuals in the novice group experienced a significant increase in confidence to perform both procedures after training on the MARTYN model (Fig. 2A and B). Those in the intermediate group reported a similar increase in confidence when using MARTYN for craniotomy training (Fig. 3A); however this improvement was not observed post training for EVD insertion (Fig. 3B). Predictably, those in the ‘experienced’ group (senior registrars and above) reported no significant increase in their selfassessed confidence despite MARTYN training in craniotomy and EVD insertion (Fig. 4A and B, respectively). MARTYN provided most benefit to those in the novice group, and provided some additional self-reported benefit to some more experienced surgeons. Overall this data suggests that the MARTYN model is better suited for ST 1 level trainees.
(A) Confidence Rating (1–5)
Confidence Rating (1–5)
(A) 5 4
3
2
5
Intermediate Craniotomy (n=4)
4
3
2
1
1 Before
Before
After
(B) 5
After MARTYN Training
MARTYN Training
(B)
Novice EVD (n=9)
Confidence Rating (1–5)
Confidence Rating (1–5)
709
4
3
2
5
Intermediate EVD (n=4)
4
3
2
1
1 Before
After MARTYN Training
Fig. 2. Subjective Confidence Rating Scores before and after MARTYN training in nine novice neurosurgical trainees. (A) Confidence in performing an emergency temporal craniotomy. (B) Confidence in placing an External Ventricular Drain (EVD). All data are presented as the median, *p ⬍ 0.05 and **p ⬍ 0.01.
Before
After MARTYN Training
Fig. 3. Subjective Confidence Rating Scores before and after MARTYN training in four intermediate neurosurgical trainees. (A) Confidence in performing an emergency temporal craniotomy. (B) Confidence in placing an External Ventricular Drain (EVD). All data are presented as the median.
C. Craven et al.
All trainees (n ⫽ 18) were asked to rate the visual and tactile realism of MARTYN in the following question using a 5-point Likert scale (1 ⫽ Unrealistic and 5 ⫽ Highly realistic). MARTYN was rated as having a high visual and tactile realism (visual median ⫽ 4, tactile median ⫽ 4) for individual components and overall. The results from one trainee were excluded due to inadequate completion of the questionnaire (Fig. 5). The bone component of MARTYN was considered most realistic (visual median ⫽ 4; tactile median ⫽ 4). The CSF and ventricles were considered less realistic (perhaps predictably, as the CSF was dyed red to make the successful drain insertions more apparent), along with the temporalis and dura mater components. No component was considered to be unrealistic. These results suggest that MARTYN is most realistic in terms of appearance and feel of the skull and is likely to be sufficiently realistic for training novices in basic procedures. In an attempt to investigate the face validity of MARTYN, we asked individuals whether they would recommend use of the MARTYN model to your colleagues and whether it was useful. All the questionnaire takers perceived MARTYN as being at least “useful”, with 11 responders describing the model as “extremely useful” (or an equivalent term) for neurosurgical training. The experienced and intermediate
(A)
Experienced Craniotomy (n=5)
4
Discussion There is currently a wide availability of simulator types for training in Neurosurgery. We have developed a model with
(A) 5
Visual Realism (n=17)
4
3
2
3 ll ra O
ve
C & s le
ur
Ve
nt
ric
D
Te
SF
n ai Br
at M a
po m
2
er
lis ra
ne
1 Bo
Confidence Rating (1–5)
5
groups were the most generous, reporting it to be very or extremely useful. No one described MARTYN as having no benefit, and furthermore, no one believed that training on MARTYN would be counterproductive. The trainees were also asked to rank their preference of training modalities (1 ⫽ First Choice and 5 ⫽ Last Choice) comparing operative experience with cadaveric training, MARTYN training, haptic models and other modalities (left blank by all responders expect one who indicted animal models). Operative experience was chosen as first choice, with cadaveric training and MARTYN consistently scoring in second and third place. Training on MARTYN for craniotomy and EVD insertion was favoured by novices, intermediates and experienced trainees over other simulation modalities such as haptic and other (including animal models).
Realism Score (1–5)
710
1 Before
(B)
Experienced EVD (n=5)
(B) 5 Realism Score (1–5)
5
4
3
Tactile Realism (n=17)
4
3
2
2
Fig. 4. Subjective Confidence Rating Scores before and after MARTYN training in five experienced neurosurgical trainees. (A) Confidence in performing an emergency temporal craniotomy. (B) Confidence in placing an External Ventricular Drain (EVD). All data are presented as the median.
ve
le s
O
&
ra ll
SF C
ai n Br
at e ur a
M
ra l
Ve
nt
ric
D
After MARTYN Training
m po
Before
Te
Bo
1
r
is
1 ne
Confidence Rating (1–5)
MARTYN Component
After MARTYN Training
MARTYN Component Fig. 5. Face Validity assessment of MARTYN components using Likerttype realism scores from 17 neurosurgeons. (A) Visual realism scores. (B) Tactile realism scores. All data are presented as the median.
Developing MARTYN the intention to fill an available niche for an accessible and inexpensive synthetic model head, which focuses on the training needs of neurosurgeons in early stages of the neurosurgical curricula.
Development Our initial development and trials show that models can be used successfully for basic procedures including emergency craniotomy and insertion of an EVD. Traditional model-making methods have enabled relatively low material costs for MARTYN compared to other simulation models and haptic devices. The benefit of bespoke models also allows for pathology by request and specific adaptations as required by the trainer (Fig. 6). The new generation models now include gelatine-based extradural haematomas that are amenable to evacuation (Fig. 6). We are currently developing subdural haematomas, meningiomas and other intraparenchymal pathologies. MARTYN models have also been successfully used for training in the removal of foreign objects such as nails and screwdrivers. A low-cost version was simultaneously produced involving a board-mounted half-skull, thus avoiding the costly joining process. The simplicity of MARTYN allows it to remain affordable and appropriate for junior trainees, whilst maintaining adequate realism making it specifically useful for training individuals on drills and insertion of extra ventricular drains. A completed MARTYN head costs 250 pounds, on which multiple operations can be performed. The models are portable, and unlike cadavers, can be used in any room (or outside), and do not come under any tissue act restrictions,
711
thus removing the need for a wet lab or specific simulation environment. A disadvantage of using such moulding methods is time. It can take up to 8 hours for one skilled personnel to produce one complete MARTYN head (including CSF-filled ventricles). However, the aforementioned half-skull version can be produced faster and with fewer resource implications. These versions of the model are more portable and affordable (Fig. 6). Another disadvantage is the inevitable degree of minor variability, limiting the standardisation of models. This may also be viewed as an advantage however, reflecting normal anatomical variance.
Feasibility and validation We are able to demonstrate face validity and a preliminary construct validity of the MARTYN model in this pilot study. The effect on confidence in novices was significant in basic procedures and as predicted, of less benefit to those with more experience. Thus, there is evidence that the model subjectively differentiates between novices and experienced trainees and is likely to hold such differences in objective studies. However effect may also be due to an increased familiarity in the procedure. The bone component of MARTYN was rated most realistic (having both visual and tactile realism) which is appropriate for a training model designed to familiarise novices with craniotomes and perforator drills. All the responders to the questionnaire felt that MARTYN was useful for training and the individual feedback supported the benefit of MARTYN, particularly for learning the basics of drilling techniques. Operative experience was ranked as the most favourable form of training. The experience of real operative training is
Fig. 6. Variations of MARTYN. (A) Isolated MARTYN brain (ventricles included) for demonstration. (B) First Generation MARTYN, with half skull removed. For demonstration, the CSF is red, revealing location of the lateral ventricles. (C) Half version MARTYN: Half skull mounted on a board including brain, skull and temporalis. (D) Extradural haematoma (following evacuation of the clot).
712 C. Craven et al. incomparable in terms of anatomical realism and pressure. Of the simulation methods that aim to emulate operative experience, MARTYN was favoured above haptic trainers and other methods (including animal models). Whilst more vigorous validation is required before any concrete conclusions can be drawn, these initial results are promising. The preliminary validation shows that the use of the MARTYN model for training and rehearsal is feasible, realistic and has potential to be used in a more formal training setting. Furthermore, the overall impression from our feedback was that this model would be of most benefit to ST 1 level trainees. In future studies we would aim to compare MARTYN training methods to traditional methods such as observation, and measure predictive validity of the model.
We believe this model will be a suitable supplement to the current ST 1–3 level cadaveric training. As students’ progress in their training, the details required in the training models will naturally need to increase. No doubt to be most effective, MARTYN will be used in conjunction with a variety of other simulators14 and training methods during ST 2–5.
Model-based simulation and the future of neurosurgical training
Declaration of interest: C Craven, M Cooke and L Carline manufactured the models. At the time of development the authors were employees or volunteers at the RCSEng. Material costs for model development were funded by the RCSEng. The authors report no personal or financial interest in any of the materials or devices described in this article.
Although simulation research has a large emphasis on realism, rehearsal is equally, if not more, important for trainees.8 Regular rehearsal with realistic tactile feedback can be provided by easily accessible and affordable simulators. Simulator training has been correlated with improved surgical performance in ENT,9 laparoscopic10 and gynaecological surgery,11 demonstrating reduced error, reduced operative times, greater accuracy and efficiently and improved confidence. Thus one might also expect to see similar outcomes following neurosurgical simulation. One of the major benefits of model-based simulation is the portability and potential to be integrated into a simulation scenario, allowing the testing of both technical and nontechnical skills. In addition to this benefit, a recent public demonstration of MARTYN has elucidated an additional potential role in family education and public engagement in neurosurgical procedures. A recent survey demonstrated that 85% of UK neurosurgical trainees had used simulation-training tools.12 However, this use of simulation tools mostly occurred in courses; we feel that MARTYN is particularly suitable for use in the hospital environment outside working hours. The authors also reported that 65% of neurosurgical trainees felt simulation should be an integral part of the current curriculum. Current Research in the United States appears to be drawing similar conclusions regarding the need for simulation in neurosurgery. A survey of 99 US neurosurgery program directors concluded that simulation should be integrated in the Neurosurgical training curricula, with a preference for models over cadavers.13
Conclusion We have described the development of the MARTYN model and the procedures that can be taught and practised on it. In addition we have collected subjective response to the current prototypes. Whilst still improving and in development, MARTYN can be already used for a variety of basic neurosurgical training, allowing juniors to become familiar with technical skills and try various operative approaches whilst not jeopardising patient safety.
Acknowledgements The authors would like to give special thanks to Mr Mark Wilson who provided expert feedback on the Model development for the Head Injury Course and subsequent public engagement events.
References 1. Bates T, Slade D. The impact of the European Working Time Directive on operative exposure. Ann R Coll Surg Engl 2007;89:452. 2. Alaraj A , Charbel FT, Birk D, et al. Role of cranial and spinal virtual and augmented reality simulation using immersive touch modules in neurosurgical training. Neurosurgery 2013; 72:115–23. 3. Aboud E, Al-Mefty O, Yasargil MG. New laboratory model for neurosurgical training that simulates live surgery. J Neurosurg 2002;97:1367–72. 4. Guvencer M, Sayhan S, Ay Dereli N, et al. Simulation of cerebrovascular circulation in the human cadaver for surgical neuroanatomy training. Turk Neurosurg 2007;17:243–6. 5. Heasley R, Farrow L, Ahmad F, Lovell M. Effect of the Human Tissue Act on UK Orthopaedic Cadaveric Research. Bull R Coll Surg Engl 2011;93:32–3. 6. Abe M, Tabuchi K, Goto M, Uchino A . Model-based surgical planning and simulation of cranial base surgery. Neuro Med Chir 1998;38:746–50; discussion 750–1. 7. Ishikawa T, Yasui N, Ono H. Novel brain model for training of deep microvascular anastomosis. Neuro Med Chir 2010;50:627–9. 8. Marcus H, Vakharia V, Kirkman MA , Murphy M, Nandi D. Practice makes perfect? The role of simulation-based deliberate practice and script-based mental rehearsal in the acquisition and maintenance of operative neurosurgical skills. Neurosurgery 2013;72:124–30. 9. Fried MP, Sadoughi B, Gibber MJ, et al. From virtual reality to the operating room: the endoscopic sinus surgery simulator experiment. Otolaryngol Head Neck Surg 2010;142:202–7. 10. Matsuda T, McDougall EM, Ono Y, et al. Positive correlation between motion analysis data on the LapMentor virtual reality laparoscopic surgical simulator and the results from videotape assessment of real laparoscopic surgeries. J Endourol 2012;26: 1506–11. 11. Culligan P, Gurshumov E, Lewis C, et al. Predictive validity of a training protocol using a robotic surgery simulator. Female Pelvic Med Reconstr Surg 2014;20:48–51. 12. Coulter IC, Brennan PM. Simulation in neurosurgery: A survey of experiences and perceptions in the UK. Bull R Coll Surg Engl 2013;95:304–7. 13. Ganju A , Aoun SG, Daou MR, et al. The role of simulation in neurosurgical education: a survey of 99 United States neurosurgery program directors. World Neurosurg 2012;80:e1–8. 14. Delorme S, Laroche D, DiRaddo R, Del Maestro RF. NeuroTouch: a physics-based virtual simulator for cranial microneurosurgery training. Neurosurgery 2012;71:32–42.
Copyright of British Journal of Neurosurgery is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
ORIGINAL ARTICLE
Development of a modelled anatomical replica for training young neurosurgeons Claudia Craven1,2, David Baxter2,3, Martyn Cooke2, Lydia Carline2,4, Samuel J. M. M. Alberti2, Jonathan Beard2 & Mary Murphy2 1Department of Neurosurgery, Addenbrooke’s Hospital, Cambridge, UK, 2Royal College of Surgeons England, London, UK, 3Department of Medicine, Imperial College London, London, UK, and 4School of Fine Art, Royal College of Art, London, UK
Abstract Introduction. The Modelled Anatomical Replica for Training Young Neurosurgeons (MARTYN) is a novel simulation model developed by the Royal College of Surgeons England (RCSEng). This study describes the development of the model and aims to determine its feasibility as a potential future training tool. Methods and materials. Traditional model-making methods were used to develop a prototype. Initial procedural trials tested the feasibility of the model. Eighteen participants, grouped by experience (nine novices, four intermediates and five experienced), completed two tasks: a craniotomy and a burr hole followed by insertion of an external ventricular drain (EVD). Subjective data on confidence, usefulness, realism and preference to other training modalities were collected via a standardised questionnaire and a 5-point Likert scale. Results. Preliminary trials of the model prototype demonstrated feasibility. The novice group had the greatest self-reported benefit from MARTYN training, with significant increases in self-rated confidence in both the craniotomy (p ⬍ 0.01) and EVD insertion (p ⬍ 0.05) procedures. MARTYN was reported to having good visual and tactile realism overall with the bone component being considered highly realistic. The model was reported to be a useful training tool. When asked to rank preferred training modalities, operative experience was chosen first with cadaveric training and MARTYN consistently scoring a second choice. Conclusions. MARTYN was developed with the intention to fill the current niche for an inexpensive synthetic model head. This study shows that the use of MARTYN for training is both feasible and realistic. We demonstrate a preliminary face and construct validity of the model in this pilot study. With the reduction in working hours, we believe this model will be a suitable supplement to the current ST 1–3 level cadaveric training and will have a positive impact on patient safety.
Introduction Current surgical trainees are receiving less exposure than in the past throughout all stages of their careers as a result of contemporary approaches to hospital service delivery, consultant-driven care and rigid implementation of the European Working Time Directive.1 In order to help mitigate problems associated with lack of operative experience, it is essential that trainees should have access to a variety of different training tools outside of the operating theatre to allow them to develop the skills necessary for future patient care. Virtual reality and haptic simulators hold great promise, with rapid development with respect to both visual and tactile realism.2 Nonetheless, their use is associated with considerable expense and it remains difficult to emulate the tactile experience using haptics alone. Although cadavers have traditionally been utilized to fill this void, they are disadvantaged due to the difficulty in replicating a bleed. Aboud et al.3 devised an ingenious solution by vascularising fresh cadaveric heads appropriate for practising a variety of neurosurgical procedures such as aneurysm repair, craniotomies, skull base approaches, tumour resection and ventricular endoscopy; this approach has since been applied to fixed cadaveric specimens.4 Nonetheless, cadaveric use in this regard has been low in the UK compared to other countries due to possible lack of resources and a requirement for pre-mortem consent.5 Simulated surgery using models provides a valid alternative. In 1998, Abe et al. developed skulls, produced from bone-like polyamid-nylon and glass beads, containing realistic brain-like tissue, dura and cavernous sinuses.6 They used patient computed tomography scans and stereolithography for model construction, thus opening up the possibility of using these models for surgical planning in addition to a training function. Ono and Co.7 developed the Kezlex model; arguably the most advanced model of its kind. The model costs upwards of £1,0007 and whilst such realistic models are
Keywords: model; neurosurgery; simulation; training
Correspondence: Dr Claudia Craven, Foundation Year 2 Doctor, Department of Neurosurgery, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 0QQ, UK. Tel: ⫹ 0044745436688. E-mail: [email protected] Received for publication 7 October 2013; accepted 6 April 2014
707
708
C. Craven et al.
ideal for the experienced surgeon, there is currently not an equivalent model for the needs of the junior trainee. We have developed a Modelled Anatomical Replica for Training Young Neurosurgeons (MARTYN) with the intention to fill the current niche for a distributable and inexpensive synthetic model head. MARTYN was developed at the Royal College of Surgeons of England (RCSEng) for junior neurosurgical trainees to practice surgical skills outside of working hours, with the aim of having a positive impact on patient safety. This paper describes the development of this model, the procedures that can be practised on the current prototypes and its preliminary subjective validation as a training tool.
Methods and materials Our aim was to construct a realistic but simple simulation model containing salient structures such as temporalis muscle, bone, dura mater, cerebrospinal fluid, brain and ventricles.
Materials Traditional moulding methods of model development were used to develop models. The skull and brain model components were cast in silicone from treated cadaveric specimens obtained from the RCSEng Wellcome Museum collection. The brain mould included removable clay insets to allow space for the lateral ventricles; the brain was cast using a gelatin composite base. Paraffin cerebro-spinal fluid (to prevent gelatin dissolution) and was injected in the lateral ventricles. Skulls were cast in thickened polyurethane resin in halves and later fixed around the brain using resin. The latex dura mater was painted to the inside of the cast
skulls. The temporalis was handcrafted in clay and then cast. Originally produced from gelatine, the second-generation temporalis was re-designed in silicone.
Procedural trials At least 40 operations were performed using the MARTYN model for training purposes; with 30 of them as part of the ‘Head Injury and Integrated Approach’ course conducted at RCSEng. Procedures performed on this model include frontal and temporal craniotomies and insertion of external ventricular drains (EVD) via burr holes.
Preliminary validation We ran a preliminary study of MARTYN models with an aim to determine the feasibility of this model as a potential training tool and establish face and construct validity. Eighteen participants; nine novices (ST 1 neurosurgical trainees and non-neurosurgeons, having done less than 10 burr holes no operative experience), four intermediates (ST 2–3 neurosurgical trainees or those who have done more than 10 burr holes and craniotomies) and five experienced trainees (those who had completed more than 30 craniotomies, which included senior registrars and consultants) completed two tasks: a craniotomy and a burr hole followed by the insertion of an EVD. Subjective data were collected via a standardised questionnaire and a 5-point Likert Scale. The questionnaire evaluated subjective confidence rating before and after MARTYN training and realism (both tactile and visual). The results were analysed using a Wilcoxon matched-pairs signed rank test and presented as the median. A p-value of less than 0.05 was considered as significant. Subjective opinions of the model overall (including its usefulness) were collected and responses were coded.
Fig. 1. Two training procedures performed on MARTYN models (A) Temporal craniotomy with revealing of the latex dura mater. (B) Reflection of dura and cortex exposure. (C) External ventricular drain insertion followed by (D) Aspiration of cerebo-spinal fluid from the left lateral ventricle (with CSF dyed red for demonstration purposes). Images courtesy of John Carr, RCSEng.
Developing MARTYN Finally, we asked all individuals to rank how the MARTYN model compares in preference to four other major training modalities; direct operative experience, cadaveric-based training, haptic trainers and other.
Results Procedural trials The model was successfully used for a variety of skills training including patient positioning and approach, skin flaps and suturing, craniotomy using both high-speed drills, perforator drill and craniotomes, dural opening, bone flap replacement, EVD and CSF aspiration (Fig. 1). The second-generation models had a temporalis made from silicone rather than a gelatine composite. We found this improved realism and allowed the muscle to be retracted without tearing. The thickened polyurethaneresin used for the skull accurately mimicked real bone as did the latex dura mater, which replicated the egg-shell peeling sensation encountered when separating real dura from bone. Similarly to actual surgery, plentiful irrigation and aspiration were beneficial during drilling as it prevented the collection of plastic bone dust and enhanced the realism of the model.
Novice Craniotomy (n=9)
Preliminary validation Confidence was assessed for two procedures, an emergency temporal craniotomy and the insertion of an EVD. The trainees were asked to rate their confidence (1 ⫽ Not confident and 5 ⫽ Confident) in both procedures, prior to training on MARTYN and after. We observed changes in selfassessed confidence ratings in the three groups: ‘novices’, ‘intermediates’ and ‘experienced’ both before and after the use of the MARTYN model. All individuals in the novice group experienced a significant increase in confidence to perform both procedures after training on the MARTYN model (Fig. 2A and B). Those in the intermediate group reported a similar increase in confidence when using MARTYN for craniotomy training (Fig. 3A); however this improvement was not observed post training for EVD insertion (Fig. 3B). Predictably, those in the ‘experienced’ group (senior registrars and above) reported no significant increase in their selfassessed confidence despite MARTYN training in craniotomy and EVD insertion (Fig. 4A and B, respectively). MARTYN provided most benefit to those in the novice group, and provided some additional self-reported benefit to some more experienced surgeons. Overall this data suggests that the MARTYN model is better suited for ST 1 level trainees.
(A) Confidence Rating (1–5)
Confidence Rating (1–5)
(A) 5 4
3
2
5
Intermediate Craniotomy (n=4)
4
3
2
1
1 Before
Before
After
(B) 5
After MARTYN Training
MARTYN Training
(B)
Novice EVD (n=9)
Confidence Rating (1–5)
Confidence Rating (1–5)
709
4
3
2
5
Intermediate EVD (n=4)
4
3
2
1
1 Before
After MARTYN Training
Fig. 2. Subjective Confidence Rating Scores before and after MARTYN training in nine novice neurosurgical trainees. (A) Confidence in performing an emergency temporal craniotomy. (B) Confidence in placing an External Ventricular Drain (EVD). All data are presented as the median, *p ⬍ 0.05 and **p ⬍ 0.01.
Before
After MARTYN Training
Fig. 3. Subjective Confidence Rating Scores before and after MARTYN training in four intermediate neurosurgical trainees. (A) Confidence in performing an emergency temporal craniotomy. (B) Confidence in placing an External Ventricular Drain (EVD). All data are presented as the median.
C. Craven et al.
All trainees (n ⫽ 18) were asked to rate the visual and tactile realism of MARTYN in the following question using a 5-point Likert scale (1 ⫽ Unrealistic and 5 ⫽ Highly realistic). MARTYN was rated as having a high visual and tactile realism (visual median ⫽ 4, tactile median ⫽ 4) for individual components and overall. The results from one trainee were excluded due to inadequate completion of the questionnaire (Fig. 5). The bone component of MARTYN was considered most realistic (visual median ⫽ 4; tactile median ⫽ 4). The CSF and ventricles were considered less realistic (perhaps predictably, as the CSF was dyed red to make the successful drain insertions more apparent), along with the temporalis and dura mater components. No component was considered to be unrealistic. These results suggest that MARTYN is most realistic in terms of appearance and feel of the skull and is likely to be sufficiently realistic for training novices in basic procedures. In an attempt to investigate the face validity of MARTYN, we asked individuals whether they would recommend use of the MARTYN model to your colleagues and whether it was useful. All the questionnaire takers perceived MARTYN as being at least “useful”, with 11 responders describing the model as “extremely useful” (or an equivalent term) for neurosurgical training. The experienced and intermediate
(A)
Experienced Craniotomy (n=5)
4
Discussion There is currently a wide availability of simulator types for training in Neurosurgery. We have developed a model with
(A) 5
Visual Realism (n=17)
4
3
2
3 ll ra O
ve
C & s le
ur
Ve
nt
ric
D
Te
SF
n ai Br
at M a
po m
2
er
lis ra
ne
1 Bo
Confidence Rating (1–5)
5
groups were the most generous, reporting it to be very or extremely useful. No one described MARTYN as having no benefit, and furthermore, no one believed that training on MARTYN would be counterproductive. The trainees were also asked to rank their preference of training modalities (1 ⫽ First Choice and 5 ⫽ Last Choice) comparing operative experience with cadaveric training, MARTYN training, haptic models and other modalities (left blank by all responders expect one who indicted animal models). Operative experience was chosen as first choice, with cadaveric training and MARTYN consistently scoring in second and third place. Training on MARTYN for craniotomy and EVD insertion was favoured by novices, intermediates and experienced trainees over other simulation modalities such as haptic and other (including animal models).
Realism Score (1–5)
710
1 Before
(B)
Experienced EVD (n=5)
(B) 5 Realism Score (1–5)
5
4
3
Tactile Realism (n=17)
4
3
2
2
Fig. 4. Subjective Confidence Rating Scores before and after MARTYN training in five experienced neurosurgical trainees. (A) Confidence in performing an emergency temporal craniotomy. (B) Confidence in placing an External Ventricular Drain (EVD). All data are presented as the median.
ve
le s
O
&
ra ll
SF C
ai n Br
at e ur a
M
ra l
Ve
nt
ric
D
After MARTYN Training
m po
Before
Te
Bo
1
r
is
1 ne
Confidence Rating (1–5)
MARTYN Component
After MARTYN Training
MARTYN Component Fig. 5. Face Validity assessment of MARTYN components using Likerttype realism scores from 17 neurosurgeons. (A) Visual realism scores. (B) Tactile realism scores. All data are presented as the median.
Developing MARTYN the intention to fill an available niche for an accessible and inexpensive synthetic model head, which focuses on the training needs of neurosurgeons in early stages of the neurosurgical curricula.
Development Our initial development and trials show that models can be used successfully for basic procedures including emergency craniotomy and insertion of an EVD. Traditional model-making methods have enabled relatively low material costs for MARTYN compared to other simulation models and haptic devices. The benefit of bespoke models also allows for pathology by request and specific adaptations as required by the trainer (Fig. 6). The new generation models now include gelatine-based extradural haematomas that are amenable to evacuation (Fig. 6). We are currently developing subdural haematomas, meningiomas and other intraparenchymal pathologies. MARTYN models have also been successfully used for training in the removal of foreign objects such as nails and screwdrivers. A low-cost version was simultaneously produced involving a board-mounted half-skull, thus avoiding the costly joining process. The simplicity of MARTYN allows it to remain affordable and appropriate for junior trainees, whilst maintaining adequate realism making it specifically useful for training individuals on drills and insertion of extra ventricular drains. A completed MARTYN head costs 250 pounds, on which multiple operations can be performed. The models are portable, and unlike cadavers, can be used in any room (or outside), and do not come under any tissue act restrictions,
711
thus removing the need for a wet lab or specific simulation environment. A disadvantage of using such moulding methods is time. It can take up to 8 hours for one skilled personnel to produce one complete MARTYN head (including CSF-filled ventricles). However, the aforementioned half-skull version can be produced faster and with fewer resource implications. These versions of the model are more portable and affordable (Fig. 6). Another disadvantage is the inevitable degree of minor variability, limiting the standardisation of models. This may also be viewed as an advantage however, reflecting normal anatomical variance.
Feasibility and validation We are able to demonstrate face validity and a preliminary construct validity of the MARTYN model in this pilot study. The effect on confidence in novices was significant in basic procedures and as predicted, of less benefit to those with more experience. Thus, there is evidence that the model subjectively differentiates between novices and experienced trainees and is likely to hold such differences in objective studies. However effect may also be due to an increased familiarity in the procedure. The bone component of MARTYN was rated most realistic (having both visual and tactile realism) which is appropriate for a training model designed to familiarise novices with craniotomes and perforator drills. All the responders to the questionnaire felt that MARTYN was useful for training and the individual feedback supported the benefit of MARTYN, particularly for learning the basics of drilling techniques. Operative experience was ranked as the most favourable form of training. The experience of real operative training is
Fig. 6. Variations of MARTYN. (A) Isolated MARTYN brain (ventricles included) for demonstration. (B) First Generation MARTYN, with half skull removed. For demonstration, the CSF is red, revealing location of the lateral ventricles. (C) Half version MARTYN: Half skull mounted on a board including brain, skull and temporalis. (D) Extradural haematoma (following evacuation of the clot).
712 C. Craven et al. incomparable in terms of anatomical realism and pressure. Of the simulation methods that aim to emulate operative experience, MARTYN was favoured above haptic trainers and other methods (including animal models). Whilst more vigorous validation is required before any concrete conclusions can be drawn, these initial results are promising. The preliminary validation shows that the use of the MARTYN model for training and rehearsal is feasible, realistic and has potential to be used in a more formal training setting. Furthermore, the overall impression from our feedback was that this model would be of most benefit to ST 1 level trainees. In future studies we would aim to compare MARTYN training methods to traditional methods such as observation, and measure predictive validity of the model.
We believe this model will be a suitable supplement to the current ST 1–3 level cadaveric training. As students’ progress in their training, the details required in the training models will naturally need to increase. No doubt to be most effective, MARTYN will be used in conjunction with a variety of other simulators14 and training methods during ST 2–5.
Model-based simulation and the future of neurosurgical training
Declaration of interest: C Craven, M Cooke and L Carline manufactured the models. At the time of development the authors were employees or volunteers at the RCSEng. Material costs for model development were funded by the RCSEng. The authors report no personal or financial interest in any of the materials or devices described in this article.
Although simulation research has a large emphasis on realism, rehearsal is equally, if not more, important for trainees.8 Regular rehearsal with realistic tactile feedback can be provided by easily accessible and affordable simulators. Simulator training has been correlated with improved surgical performance in ENT,9 laparoscopic10 and gynaecological surgery,11 demonstrating reduced error, reduced operative times, greater accuracy and efficiently and improved confidence. Thus one might also expect to see similar outcomes following neurosurgical simulation. One of the major benefits of model-based simulation is the portability and potential to be integrated into a simulation scenario, allowing the testing of both technical and nontechnical skills. In addition to this benefit, a recent public demonstration of MARTYN has elucidated an additional potential role in family education and public engagement in neurosurgical procedures. A recent survey demonstrated that 85% of UK neurosurgical trainees had used simulation-training tools.12 However, this use of simulation tools mostly occurred in courses; we feel that MARTYN is particularly suitable for use in the hospital environment outside working hours. The authors also reported that 65% of neurosurgical trainees felt simulation should be an integral part of the current curriculum. Current Research in the United States appears to be drawing similar conclusions regarding the need for simulation in neurosurgery. A survey of 99 US neurosurgery program directors concluded that simulation should be integrated in the Neurosurgical training curricula, with a preference for models over cadavers.13
Conclusion We have described the development of the MARTYN model and the procedures that can be taught and practised on it. In addition we have collected subjective response to the current prototypes. Whilst still improving and in development, MARTYN can be already used for a variety of basic neurosurgical training, allowing juniors to become familiar with technical skills and try various operative approaches whilst not jeopardising patient safety.
Acknowledgements The authors would like to give special thanks to Mr Mark Wilson who provided expert feedback on the Model development for the Head Injury Course and subsequent public engagement events.
References 1. Bates T, Slade D. The impact of the European Working Time Directive on operative exposure. Ann R Coll Surg Engl 2007;89:452. 2. Alaraj A , Charbel FT, Birk D, et al. Role of cranial and spinal virtual and augmented reality simulation using immersive touch modules in neurosurgical training. Neurosurgery 2013; 72:115–23. 3. Aboud E, Al-Mefty O, Yasargil MG. New laboratory model for neurosurgical training that simulates live surgery. J Neurosurg 2002;97:1367–72. 4. Guvencer M, Sayhan S, Ay Dereli N, et al. Simulation of cerebrovascular circulation in the human cadaver for surgical neuroanatomy training. Turk Neurosurg 2007;17:243–6. 5. Heasley R, Farrow L, Ahmad F, Lovell M. Effect of the Human Tissue Act on UK Orthopaedic Cadaveric Research. Bull R Coll Surg Engl 2011;93:32–3. 6. Abe M, Tabuchi K, Goto M, Uchino A . Model-based surgical planning and simulation of cranial base surgery. Neuro Med Chir 1998;38:746–50; discussion 750–1. 7. Ishikawa T, Yasui N, Ono H. Novel brain model for training of deep microvascular anastomosis. Neuro Med Chir 2010;50:627–9. 8. Marcus H, Vakharia V, Kirkman MA , Murphy M, Nandi D. Practice makes perfect? The role of simulation-based deliberate practice and script-based mental rehearsal in the acquisition and maintenance of operative neurosurgical skills. Neurosurgery 2013;72:124–30. 9. Fried MP, Sadoughi B, Gibber MJ, et al. From virtual reality to the operating room: the endoscopic sinus surgery simulator experiment. Otolaryngol Head Neck Surg 2010;142:202–7. 10. Matsuda T, McDougall EM, Ono Y, et al. Positive correlation between motion analysis data on the LapMentor virtual reality laparoscopic surgical simulator and the results from videotape assessment of real laparoscopic surgeries. J Endourol 2012;26: 1506–11. 11. Culligan P, Gurshumov E, Lewis C, et al. Predictive validity of a training protocol using a robotic surgery simulator. Female Pelvic Med Reconstr Surg 2014;20:48–51. 12. Coulter IC, Brennan PM. Simulation in neurosurgery: A survey of experiences and perceptions in the UK. Bull R Coll Surg Engl 2013;95:304–7. 13. Ganju A , Aoun SG, Daou MR, et al. The role of simulation in neurosurgical education: a survey of 99 United States neurosurgery program directors. World Neurosurg 2012;80:e1–8. 14. Delorme S, Laroche D, DiRaddo R, Del Maestro RF. NeuroTouch: a physics-based virtual simulator for cranial microneurosurgery training. Neurosurgery 2012;71:32–42.
Copyright of British Journal of Neurosurgery is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
