Dr R Moreau, Laboratoire Ampère, INSA-Lyon, 25 Avenue Jean Capelle, 69621 Villeurbanne Cedex, France. Email email@example.com
This paper aims to highlight the benefits of simulator training in obstetric manipulations such as forceps blade placement. The BirthSIM simulator is used to mimic operative vaginal deliveries. To characterise forceps blade placement, we studied the curvature of forceps path. The orientation of the forceps blades are studied in the quaternion unit space to ensure time-independent analysis. The results showed progress for all novices in forceps blade placement. Simulator training helps them to develop their self-confidence and acquire experience before working in the delivery room.
One major difficulty in obstetric training is the transfer of knowledge and skills. In order to learn, junior obstetricians observe senior obstetricians while they perform obstetric manipulations and try, under supervision, to repeat what they have seen. It is, however, difficult to correctly assimilate such manipulations under the constraints of both emergency and stressful conditions with these manipulations not easily visible as they are performed inside the maternal body.
Simulators in obstetrics are available to give initial training before moving on to hands on training. Simulators can be divided into three groups.
Anatomical simulators, featuring anthropomorphic manikins, are often used in midwifery and medical schools. The first known simulator was constructed in 1759 by Madame Du Coudray.1 It consisted of a female pelvis and the body of a fetus. The dummy fetus was made of tissue papers and a real pelvis was used. Nowadays, plastic simulators are currently commercially available, including a pelvis with a full-term fetus (Simulaïds Inc., Saugerties, NY, USA) and a perineum with an anal sphincter (Limbs & Things Inc., Bristol, UK). These simulators make it possible to teach the anatomy of the pelvis and fetal head as well as provide an overall view of the delivery process.
These make it possible to observe the path of the fetus through the pelvis.2 On a screen, operators can watch both the progression of the fetus during its expulsion and the activity of pelvic muscles. Some of these simulators employ haptic feedback systems,3 so that the trainee can sense the forces involved. Most of the time, however, the interface is a joystick, which is not realistic and thus not useful for operators.
Instrumented anatomical simulators
This last type of simulator is much more attractive because it integrates the functional characteristics of both the above types. Such a simulator makes it possible to assess operator performance repeated over several exercises, and the simulation of different cases. They also provide users with greater interactivity. Several simulators have been developed and one particular simulator is used to provide training in the management of shoulder dystocia.4 With regard to training in instrumental deliveries, some instrumented simulators offer training methods for learning the correct instrument movements.5–7 None, however, offer the objective evaluation of the manipulations carried out. With these simulators, the evaluation is made by an experienced obstetrician who determines whether or not a junior obstetrician has sufficient experience to perform such manipulations by him(her)self.
The BirthSIM simulator was developed at the Hospices Civils de Lyon (HCL) and by Laboratoire Ampère and has been designed to fulfil this lack.8 In the long term, this tool will ensure risk-free training of junior obstetricians and midwives, allowing them to acquire initial skills before moving on to hands on delivery room training.
Materials and methods
The BirthSIM simulator integrates the following three distinct components (Figure 1).
The mechanical component of the BirthSIM simulator consists of anthropomorphic dummies with the main anatomical markers: ischial spines, coccyx, sacrum, and pubis for the pelvis and sutures, fontanels, and ears for the fetus. Only the head of the fetus is used as the model for the fetus. We assume the fetus is in a cephalic presentation and that, once the head has been extracted, the rest of the body is usually expelled without any complication. An electromagnetic position sensor capable of measuring the 6 degrees of freedom (dof) is mounted inside the head of the fetus. A similar sensor is also placed on each blade of a pair of obstetric forceps. These three sensors allow the position and orientation of the fetal head and forceps blades to be monitored in real time with respect to a fixed frame attached to the pelvis.9
The BirthSIM simulator integrates an electropneumatic system, which is responsible for simulating the dynamics of the childbirth process. This system consists of a servovalve and a pneumatic cylinder. Pressure sensors are mounted near each chamber of the pneumatic actuator. Different control laws of this system make it possible not only to position the head of the fetus but also to reproduce the different childbirth-related forces. A force sensor is mounted between the fetal head and the extremity of the pneumatic actuator rod to measure the force exerted on the head of the fetus by the operator. In this paper, we focus mainly on forceps blade placement. The electropneumatic component is only used to ensure the repeatability of the initial placement of the fetal head.
A display interface is available on the BirthSIM simulator. It offers real-time visualisation of the position of both the obstetric instruments and the fetal head with respect to the parturient’s pelvis from several viewpoints (Figure 2). The BirthSIM simulator offers operators the possibility of viewing inside the parturient’s pelvis, thereby making a manipulation inside the maternal body visible.
Only forceps blade placement is studied in this paper. Junior obstetricians must first correctly acquire the placement technique before proceeding to the extraction technique. Previous studies have focused on forceps blade placement but only blade positions were analysed,10 whereas their orientations are also crucial.
Three junior obstetricians were trained on the BirthSIM simulator in collaboration with Hospices Civils de Lyon. A junior obstetrician is defined as an obstetrician with less than 12 months of obstetric experience. BirthSIM simulator training was supervised by the first author, a senior obstetrician, acting as an instructor. In the literature, there is no agreed definition for ‘an experienced obstetrician’. For the purpose of this study, we have arbitrarily defined an experienced obstetrician as a senior obstetrician with more than 10 years of experience and who uses obstetric forceps in over 80% of his or her vaginal instrumental deliveries. The position and station of the fetal head was Occipital Anterior and spines +4 cm (OA+4), respectively, according to the American College of Obstetricians and Gynecologists classification.11
Training consists of 2 minutes of sessions with a 7-day gap. During each session, trainees performed 15 forceps blade placements during 1 hour. Initially, the junior obstetric trainees placed forceps as they normally would in the delivery room, without any guidance from the instructor. The instructor then explained to them progressively how to place the forceps correctly using the BirthSIM simulator. The recording protocol begins with the left blade (the blade held by the left hand) when the trainee is ready and ends when he estimates the left blade is correctly placed. Then, the trainee grasps the right blade (the blade held by the right hand) and the recording starts when he is ready to place it and it ends when both blades are locked. It is noteworthy that the recording of the left blade trajectories recording does not take into account the movement related to the locking procedure, whereas the second blade trajectories include it.
Manipulations need to be analysed on the basis of criteria established by the medical team from an obstetric standpoint. These criteria are given below.
Independent analysis based on manipulation time: given the dynamics of manipulation, the time required to place forceps blades is usually not critical in real practice. In an emergency situation, the fetus has to be extracted as quickly as possible. A recent study has shown that a forceps delivery is twice as quick as a caesarean section.12 Furthermore, the action of positioning the forceps takes several seconds, which makes it a relatively short manoeuvre compared with a surgical procedure such as a caesarean section. Manipulation time should therefore not be employed in the analysis of this study.
The manipulation should be fully analysed not just in relation to specific points. Forceps are effectively almost always in contact with both the pelvic muscles and the head of the fetus inside the parturient’s pelvis. There is therefore a permanent risk of injuring the parturient, the fetus, or both.
Analysis must take into account not only the positions but also the orientations of the forceps blades. For this reason, we used sensors that are capable of measuring 6 dof. While performing the analysis, both positions and orientations must be studied.
Three methods have been developed to analyse forceps blade placement. The first is used to study operator reliability by calculating the distance between specific points from different paths.13 This analysis is a time-independent approach, but it only takes into consideration a few specific points. The second method developed considers global manipulation by calculating the integral error compared with a reference manipulation defined by a senior obstetrician.14 This calculation requires data standardisation, which can result in data modification, especially when the time difference is large, as in the case of junior obstetricians. Both methods have a common inconvenient: they cannot be applied to study the orientations. In response to the requirements of obstetricians, a third method has been developed.15 It evaluates manipulations by comparing their curvature correlation. To ensure time independence, the data are first expressed as a function of cumulative chord length, before calculating the curvature. Once the curvature has been calculated for each trajectory, the results are compared by correlation using the Pearson coefficient.16 This coefficient makes it possible to calculate the similarity between the two manipulation curvatures. This method offers another advantage: it can be adapted to orientations. In this case, data are first expressed in quaternion unit space to ensure time independence. Quaternions make it possible to study any general three-dimensional rotation that can effectively be transformed into a unit norm quaternion. One major difficulty in studying angular trajectories is to visualise them in space. The quaternion unit space can be represented as a unit sphere on which the angular trajectory can be plotted and, thus, be visually compared. Calculating the curvature of the angular trajectory makes it possible to obtain a parameter that fulfils the desired criteria: time independence, complete trajectory and orientations are taken into account. Technical details are given in Moreau et al.17
Figure 3 shows the orientations represented on the quaternion unit space, which is an unit sphere. This figure shows that post-training paths are closer to the senior obstetrician’s paths than before training. We noticed that the novices’ paths were less hesitant, closer to the reference path and smoother at the end of the training. Only the paths of the left forceps blade are plotted, but similar results were obtained for the right forceps blade. The reference path represented is spread over one-eighth of the quaternion sphere. It means that the trajectory covers a rotation of about 90°: from the initial vertical position of the blade to the final horizontal position of the blade. Novices’ trajectories at the beginning of the training are not as spread as the expert one, which means that they did not respect the major rotation of the blade. Their initial blade placement is nearly the same than their final which is horizontal, the manipulations are thus not correct.
The curvature correlation study allows this improvement with forceps blade placement to be quantified. Table 1 shows the difference between the first training session and the second. These values are the correlation coefficients of path curvature compared with the reference manipulation curvature. Each correlation coefficient is the average of all the manipulations recorded during the training session. In Table 1, LFB means left forceps blade and RFB means right forceps blade. Both sessions are compared to observe the novices’ progression while being trained on the simulator. Their progressions between the two sessions are also calculated.
Table 1. Correlation coefficient variation during training
Correlation coefficient (%)
Progression between the two training sessions in %
Training Session 1
Training Session 2
All the orientation correlation coefficients were higher during the second training session than during the first training session. On some trajectories, novices manage to obtain a correlation coefficient around the expert performance, that is, around 60% for both blades. They, however, did not manage to obtain regularly such results; therefore, the average of their 15 manipulations is quite low. These junior obstetricians gained experience and are able to carry out a forceps placement as an expert does but they still required further training to enhance their performance. Unfortunately, we did not have time to perform longer experiments.
This paper provides an overview of the BirthSIM simulator and the results obtained while training junior obstetricians. Training was divided into two stages. The first stage was to allow junior obstetricians to correctly position the forceps. The experimental results clearly showed that the junior obstetricians were able to improve manipulation and gradually approach the results of a senior obstetrician. The second stage concerned the extraction manipulation, but in this paper only forceps blade placement was studied.
With regard to the results, the rise in the correlation coefficient was greater with the right forceps blade (+60.2 versus +71.4%), although the correlation coefficient during the first training session was higher for the left blade than for the right (17.5 versus 11.5%). During the second training session, the correlation coefficient was also higher for the left blade than the right (28 versus 19.8%). This was because of the fact that the recording protocol takes into account the locking procedure during the right blade placement, whereas for the left blade it is not taken into account. This locking procedure requires a certain experience to carry it out correctly. As our analysis is based on the curvature of the trajectories and as the locking procedure has an influence on the curvature, it explains the difference between both blades for novices that did not appear for experts. Moreover, the right blade is placed after the left blade so it is more difficult to perform a correct gesture because the left blade, which is already placed inside the body, may perturb the placement of the right blade.
The BirthSIM simulator is a tool for training junior obstetricians by providing them with the opportunity to overcome conventional training-related constraints. The proposed training allows junior obstetricians to acquire an initial experience by performing obstetric manipulations with no time or stress constraints. A training method based on the visualisation interface showed on Figure 2 is under development to provide a pedagogical method to correctly place forceps. The next step of the training consists in performing the delivery manipulation once the placement technique has been correctly acquired. With regard to the future, our results naturally need to be complemented by further training sessions with more junior obstetricians. These initial results are nevertheless promising and highlight the value of a simulator in obstetrics as a means of gaining initial experience.
Currently, our results show that novices manage to improve their gestures on the simulator and that they are more self-confident after simulator training. However, it is difficult to know the influence of this kind of training on their performance in the delivery room. To assess this issue will need prospective study using clinical end-point such as maternal or neonatal morbidity.
Disclosure of interests
There are no disclosure of interests concerning this paper because for the moment the simulator is not commercially available.
Contribution to authorship
Each co-author contributed to conception and analysis of data and revised this article critically for important intellectual contents.
Details of ethics approval
As this study was performed on inanimate simulator, no ethic committee approbation was requested.
The prototype of the BirthSIM simulator used for this study has been developed thanks to the Hospices Civils de Lyon and the CNRS (Centre National de la Recherche Scientifique – French national institute of research) funding.
The authors thank Professor Pierre Boulanger and Victor Ochoa from the Advanced Man Machine Interface (AMMI) Laboratory, Computer Science Department, University of Alberta, Canada for their useful help concerning the quaternion theory.
Reviewers’ Commentary on ‘Emerging technologies in obstetrics–simulation training’
Simulation has had great potential for transferring obstetric skills to junior doctors for at least 250 years. Madame Du Coudray, a French midwife, was commissioned in 1759 by Louis XV to travel throughout France teaching birthing skills: she wrote ‘… when difficulties arise they [birth attendants] are absolutely unskilled, and until long experience instructs them they are the witness or the cause of many misfortunes, … one learns on the machine in little time how to prevent such accidents.’
Potentially avoidable intrapartum morbidity and mortality is still being reported (Perinatal Mortality 2006, CEMACH). Evidence from simulation suggests that obstetricians and midwives may not possess advanced intrapartum management skills (Crofts, et al. Obstet Gynecol 2006;108:1477–85). In parallel, reductions in junior doctors’ hours, combined with an increased caesarean rate, have reduced the exposure of trainees to potentially difficult vaginal deliveries. Providing accoucheurs with the skills to manage these complex cases remains a challenge (Patel and Murphy. BMJ 2004;328:1302–5), and simulation is again recommended as a potential solution (The Kings Fund. Safe Births: Everybody’s Business 2008).
However, the exact role of simulation remains unclear. The articles by Dupuis et al. and Bahl et al. add to the increasing experience in the use of simulation-based technologies within obstetrics.
Bahl et al. have attempted to capture the subtleties of experience using a combination of semistructured interviews and videos of simulated ventouse deliveries. These two methods allowed them to define a comprehensive skills list, which could be used to both assess trainee’s performance and inform future training programmes.
Dupuis et al. identify experience in instrumental delivery as key to reducing perinatal morbidity and mortality. They have developed a Virtual Reality Birth Sim simulator to train junior obstetricians how to correctly place forceps blades.
The findings of both articles appear promising, although both studies recognise that training should be validated in a clinical setting.
Simulation training using mannequins has been demonstrated to improve the simulated management of instrumental delivery (Barnfield et al. Simul Healthcare 2007;2:209), eclampsia (Ellis et al. Obstet Gynecol 2008;111:723–31), postpartum haemorrhage, vaginal breech delivery (Deering et al. Obstet Gynecol 2004;103:1224–8), and shoulder dystocia (Crofts et al. Obstet Gynecol 2006;108:1477–85; Deering et al. Obstet Gynecol 2004;103:1224–8). Evidence suggests that shoulder dystocia training is more effective when a specifically designed, high-fidelity mannequin is used when compared to training on simple mannequins (Crofts et al. Obstet Gynecol 2006;108:1477–85). However, simulation using expensive mannequins may not be universally applicable; training using a patient-actor appears to improve communication more than training using a full-bodied mannequin (Crofts et al. Qual Saf Health Care 2008;17:20–4). Similarly, in one randomised study, training locally within obstetric departments with relatively low-technology and low-cost simulation was as effective as training in a simulation centre (Ellis et al. Obstet Gynecol 2008;111:723–31).
Obstetric simulation training has been associated with real improvements in perinatal outcome. Significant reductions in the rate of low Apgar scores, hypoxic ischaemic encephalopathy and obstetric brachial plexus injuries have been demonstrated (Draycott et al. BJOG 2006;113:177–82; Draycott et al. Obstet Gynecol 2008;112:14–20). Reductions in malpractice claims have also been noted after the introduction of training (Pratt and Sachs, 2006, http://www.webmm.ahrq.gov/perspective.aspx?perspectiveID=21). However, not all training appears to be equally effective; some studies have not found a significant clinical benefit to training (Whittle et al. BJOG 2008;115:125; MacKenzie et al. Obstet Gynecol 2007;110:1059–68).
As a profession, we need to increase research into new educational technologies to provide the best training for our trainees and avoid unnecessary morbidity for our patients. The challenge is to determine which educational technologies work best, where, and at what cost.
Disclosure of Interests
Mr Timothy J. Draycott is a consultant to Limbs and Things Ltd, manufacturers of the PROMPT Birthing Simulator. Dr Joanna F. Crofts has no conflict of interest to declare.
TJ Draycott,a JF Croftsb a Southmead Hospital, Bristol, UK b Musgrove Park Hospital, Taunton, UK