Approximately 25% of cesarean deliveries for failure to progress occur at full dilatation.1 Variability in case mix, individual practice, and senior support make it difficult to estimate the effect internationally, but this may represent approximately 65,000 deliveries in the United States each year.2 The incidence of full dilatation cesarean delivery is reported to be rising disproportionately compared with overall cesarean delivery rates.3 The reasons for this are likely to be multifactorial; reduced working hours for junior doctors resulting in less confidence in operative vaginal deliveries, increasing epidural use, and fear of litigation have all been cited.4
Full dilatation cesarean deliveries can be technically challenging when the fetal head is deeply impacted into the maternal pelvis. Severe impaction is uncommon, but frequently this occurs without experienced assistance and the need for delivery is urgent. These cases can be associated with increased maternal morbidity including excessive manipulation of the stretched lower segment resulting in angle extensions and obstetric hemorrhage.4 Delay in delivering the impacted fetal head can contribute to increased requirement for immediate neonatal resuscitation and increased rates of neonatal admission to specialist care.4 Techniques for safe delivery vary internationally and there are no national guidelines or formal training programs despite research demonstrating that obstetric trainees (in the United Kingdom) feel it is required.5
Simulation in obstetrics and gynecology training is gaining popularity quickly,6 and increasingly evidence suggests it can improve management of obstetric emergencies.7 We have invented and developed “Desperate Debra,” a full dilatation cesarean delivery simulator, as shown in Figure 1 and in the Video available online at https://www.youtube.com/watch?v=spTVjcMH-N0.
The aims of this study were 1) to assess the face validity (the simulator represents what it is supposed to represent), content validity (the simulator teaches what it is supposed to teach), and construct validity of the simulator (the simulator can distinguish between experienced and inexperienced users); and 2) to establish expert consensus on the most important techniques for safe delivery when the fetal head is impacted.
MATERIALS AND METHODS
The simulator replicates delivery of a fetal head at full dilatation cesarean delivery through a transverse suprapubic skin incision (Fig. 1). The fetal head can be adjusted to simulate varying degrees of fetal head flexion, rotation, and impaction into the pelvis that can be assessed by vaginal examination.
The simulator has an analog scale of difficulty from minor impaction to very severe impaction. This is determined by the degree to which the fetal head is pushed into the maternal pelvis and thus the force required to elevate the head out of the pelvis. The adjustment is made using a variable spring mechanism that is quantified by length of an external screw as shown in Figure 2. Additional difficulty is created by adjusting the position of the head, which is indicated on an externally visible dial (Fig. 3). The degree of flexion of the fetal head can be manually adjusted and is shown on an external sliding scale (Fig. 3).
Three delivery scenarios with increasing degrees of difficulty were tested. In level 1 the fetal head was direct occiput posterior and lightly impacted in the pelvis. In level 2 it was left occiput transverse and moderately impacted in the pelvis. In level 3 it was right occiput transverse and severely impacted into the pelvis. For the purpose of this study, the flexion of the fetal head was in neutral in all three scenarios because this was not considered to be a critical variable to evaluate the difficulty of delivery and to limit the number variables being analyzed.
The simulator was taken to three National Health Service maternity departments (Guys and St Thomas' National Health Service Hospital Trust, West Middlesex University Hospital National Health Service Trust and Victoria Hospital, National Health Service Fife) between September and December 2012.
Participants attended with no prior notice of the simulation session. Any participant who had previously used the simulator was excluded. All participants received a brief introduction to the study before consenting to participate. No clinician declined to take part. They were instructed to deliver the fetal head like they would in real life. Participants were assigned a case number and data were collected on the number of years in the specialty. All levels of trainees and consultants present were invited to complete the three scenarios, which were filmed for later analysis. On completion, visual analog scale (VAS) scores were filled to rate the perceived difficulty of each scenario.
Data were collated on whether successful delivery was achieved and the time taken to do so. Results were analyzed according to predefined groups. The construct validity was assessed by comparing the number of successful deliveries and the time taken to deliver between those with less than 7 years of experience in the specialty (junior group-representative of doctors undergoing specialty training which is typically a 7-year program in the United Kingdom) to those with 7 or more years of experience (experienced group—representing those postcompletion of specialty training).
Face and content validity was assessed by completion of an anonymous questionnaire immediately after finishing the simulation exercise. This contained VAS on how realistic participants thought the simulator was and how useful they thought the simulator would be as a training device. All users were invited to provide additional general feedback on the simulator.
Expert opinion was sought through a questionnaire that was sent to obstetricians with more than 7 years of labor ward experience, who were active members of the British Maternal Fetal Medicine Society. They were asked to list the five most important techniques for safe delivery at full dilatation cesarean delivery. These were independently categorized into themes by two individuals and their frequency recorded. The results were then compared and discrepancies reviewed by a third party.
Statistical analysis was performed using Stata 11.2. Visual analog scores were summarized using median and quartiles. Where delivery of the fetal head was attempted and failed, the time taken to delivery was treated as censored at the time the attempt was stopped. All tasks were considered together using random effects interval-data regression modeling.8 Tests for main effects of task and level of experience (less than 7 or 7 or more years of obstetric experience) were carried out in addition to interactions. No formal power calculation was performed for this initial assessment, but a prospective decision was made to evaluate clinicians from three institutions. After discussion with our local research and development department, it was determined that, under section three of the Governance Arrangements for Research Ethics Committees changes to the remit of Research Ethics Committees (September 2011), there is no requirement for Research Ethics Committee review of this research.
From the three National Health Service trusts, 30 obstetricians participated in the simulation scenarios. No participants had prior experience with the simulator. Four participants (two from the junior group and two from the experienced group) did not attempt to deliver scenario 3 and therefore were excluded from the analysis of this scenario. In each case the reason cited was discomfort of the delivering hand or arm experienced at easier levels. The mean number of years practicing in this specialty was 7. Of these participants, 17 had less than 7 years (mean 3 years) and 13 had 7 or more years of experience (mean 12 years).
Overall 100% (n=30) of participants were able to deliver the fetal head in scenario 1, 90% (n=27) in scenario 2, and 62% (n=16) in scenario 3 (1–2 compared with 3 binomial regression, P<.05). Average time to delivery was 45 seconds for scenario 1, 43 seconds for scenario 2, and 109 seconds for scenario 3. Therefore, scenario 3 took 3.4 times longer to deliver than scenario 1 (confidence interval 2.3–5.1; P<.001). The degrees of impaction correlated with perceived difficulty of delivery (mean VAS 1, 29/100; VAS 2, 42/100; and VAS 3, 88/100, respectively; scenario 3 rated 42 points higher than scenarios 1 and 2 after allowing for success rates; standard error 5.3; P<.001). These values are shown in Table 1.
To assess the construct validity of the simulator, the results for experienced doctors were compared with junior doctors. It was shown that 100% of both groups could deliver scenario 1 with no significant difference in the time taken to delivery (scenario 1 mean time to delivery: experienced doctors 30 seconds compared with junior doctors 58 seconds; P=.08). There was also no significant difference in the rates of successful delivery between the experienced doctors and junior doctors during the more difficult scenarios (percentage achieving successful delivery in scenario 2: 85% compared with 100% P=.28, scenario 3: 55% compared with 67%, respectively; P=.34). With all scenarios combined, experienced doctors on average deliver 60% faster than junior doctors (standard error 0.18, P=.08). However, the only scenario where experienced doctors delivered significantly faster than junior doctors was scenario 3 (scenario 2 mean time to delivery: 42 seconds compared with 44 seconds, P=.53, scenario 3: 81 seconds compared with 131 seconds, respectively, P<.001).
When assessing face validity, 87% of participants found the simulation device to be quite realistic or very realistic (79% [n=10] of experienced doctors and 94% of juniors [n=16]). The overall realism of the simulator was rated as 58 of 100 (interquartile range 40–74). Those who successfully delivered all three scenarios thought it was more realistic than those who failed to deliver or did not attempt one or more scenarios (72/100 compared with 41/100, P<.001). The user feedback highlighted two main areas for potential improvement of the simulator. The materials used to create the artificial vagina were thought to be too firm making it difficult to push up from below. Users would like to see the addition of a mechanism that can tell you how much force you are applying to the fetal head and whether this may be associated with fetal injury.
When assessing content validity, 93% of participants thought it was a quite useful or very useful training device (93% [n=13] of experienced doctors and 94% [n=16] of juniors). Overall usefulness of the device for training was rated as 80 of 100 (interquartile range 55–97). Those who successfully delivered all three scenarios thought it was more useful as a training device than those who failed to deliver or did not attempt one or more scenarios (85/100 compared with 63/100, P=.02).
Responses to the questionnaire were received from 47 consultants (7 years or more of experience) with a mean 18 years of labor-ward experience. Responses that were stated by more than one consultant are shown in Table 2. Responses that were not specific techniques for safe delivery, for example ensuring senior support is available, were not reported.
The results of this initial evaluation have demonstrated the face and content validity of the simulator. We failed to show the construct validity of the device because there were no significant differences in the number of successful deliveries or the average time taken to deliver between experienced and junior doctors. It is possible that the lack of construct validity was because the study was underpowered to show a statistically significant difference or that it can be explained by the methodology that had to be used in view of the lack of clinically established end points to evaluate against.
In the group as a whole, there were fewer successful deliveries in scenario 3 than the other scenarios. This is most likely reflects the intended difficulty of the device setting. It was our intention for the hardest setting to be difficult and we would expect failures to mimic the full range of difficulty that can be experienced. This is required for the simulator to stimulate the development of techniques to increase successful delivery.
There was a tendency for junior doctors to have more successful deliveries in the difficult scenarios than experienced doctors. It is possible that this reflects current practice in the United Kingdom, where trainees are performing cesarean deliveries more frequently than their consultant colleagues. However, the results also demonstrate that junior doctors took significantly longer to deliver in scenario 3 than the experienced doctors. This could reflect that junior doctors persevered with delivery for longer periods of time, thus achieving higher rates of successful delivery or alternatively that experienced doctors either delivered successfully or gave up more quickly. It could be argued that this is because the more difficult setting was deemed unrealistic and therefore they stopped the scenario early or that they felt challenged by the difficult setting and therefore stopped trying.
Experienced doctors also considered the simulator to be less realistic than their junior colleagues. Although this could be because they felt that delivery was too difficult to be realistic, our results demonstrate that it could be because they had fewer successful deliveries or had struggled to deliver. Indeed, those who did not successfully deliver all three scenarios were significantly less likely to think that the simulator was realistic than those who delivered all three scenarios.
There are no national guidelines to inform clinicians on the best way to deliver in this scenario. Only one randomized controlled trial has compared the most commonly reported push and pull techniques.9 The questionnaire responses highlight the diversity of techniques used by experienced obstetricians to deliver at full dilatation cesarean delivery. Only two techniques, high uterine incision and assistance to push the fetal head up transvaginally, were reported by more than half of the experts. Our survey will be limited by the individuals sampled. However, current practice is based on individual advice when faced with the scenario in real life. Therefore, although our survey cannot provide training guidance, it does highlight the necessity for further research in this field, including a complete Delphi consensus. Future research needs to examine the techniques used in successful deliveries.
In conclusion, Desperate Debra, the full dilatation cesarean delivery simulator, is a valuable training device. Full dilatation cesarean deliveries are increasing in prevalence and associated with significant maternal and neonatal morbidities. There is no consensus on the best way to deliver. The nature of emergency deliveries makes it difficult to study different techniques, but their prevalence and importance justify investigation. Expert consensus could be used to create a structured training protocol similar to that used for other obstetric emergencies such as shoulder dystocia or breech presentation. Acute, serious conditions requiring skill and experience are ideal for a training simulator. Desperate Debra provides a tool to achieve this for impacted heads at full dilatation cesarean deliveries and should be used to develop consensus of accepted techniques that can be incorporated into training protocols.
1. Royal College of Obstetricians and Gynaecologists Clinical Effectiveness Support Unit. The national sentinel caesarean section audit report. London (UK): RCOG Press; 2001.
2. Martin JA, Hamilton BE, Osterman MJK, Curtin SC, Mathews TJ. Births: final data for 2012. Natl Vital Stat Rep 2013;62:9.
3. Unterscheider J, McMenamin M, Cullinane F. Rising rates of caesarean deliveries at full cervical dilatation: concerning trend. Eur J Obstet Gynaecol Reprod Biol 2011;157:141–4.
4. Murphy DJ, Liebling RE, Verity L, Swingler R, Patel R. Early maternal and neonatal morbidity associated with operative delivery in second stage of labour: a cohort study. Lancet 2001;358:1203–7.
5. Sethuram R, Jamjute P, Kevelighan E. Delivery of the deeply engaged head: a lacuna in training. J Obstet Gynaecol 2010;30:545–9.
6. Macedonia CR, Gherman RB, Satin AJ. Simulation laboratories for training in obstetrics and gynecology. Obstet Gynecol 2003;102:388–92.
7. Deering S, Poggi S, Macedonia C, Gherman R, Satin AJ. Improving resident competency in the management of shoulder dystocia with simulation training. Obstet Gynecol 2004;103:1224–8.
8. Skrondal A, Rabe-Hesketh S. Generalized latent variable modeling: multilevel, longitudinal, and structural equation models. Boca Raton (FL): Chapman & Hall/CRC; 2004.
9. Fasubaa OB, Ezechi OC, Orji EO, Ogunniyi SO, Akindele ST, Loto OM, et al.. Delivery of the impacted head of the fetus at caesarean section after prolonged obstructed labour: a randomised comparative study of two methods. Obstet Gynaecol 2002;22:375–8.