Dr JF Crofts, The SaFE Study, Department of Obstetrics and Gynaecology, Southmead Hospital, North Bristol NHS Trust, Bristol BS10 5NB, UK. Email email@example.com
Objective To investigate completeness and accuracy of record keeping by comparison of documentation and actual events, recorded on video and through a force-monitoring device, during simulated shoulder dystocia.
Design An observational study.
Setting Six maternity units in South West of England.
Population Seventy-one midwives and 39 doctors.
Methods Doctors and midwives documented their management of a shoulder dystocia simulation on paper used in their hospital (simple notepaper or preformatted form). Documentation was compared with video recording of each simulation and an electronic record of force applied during delivery.
Main outcome measures Documentation of head-to-body delivery time (and comparison with actual delivery time). Documentation of force (and comparison with actual applied force).
Results A total of 110 participants documented their actions, 70.9% used a preformatted sheet, 29.1% used hospital notepaper. Fifty-six percent documented head-to-body delivery interval (HBDI) with 56% overestimating the time by more than 1 minute. Force used during the simulation was documented by 70.9%, with no relationship between the subjective description of force applied and the maximum recorded force. The anterior shoulder was documented by 78.2% and correctly identified in 80%. Documentation of force was more likely if a preformatted sheet was used (88 versus 53%, P = 0.016). Documentation of the laterality of the anterior shoulder was tended to be more accurate with plain hospital notepaper (93 versus 78%, P = 0.3526).
Conclusions Manoeuvres performed were well documented. HBDI and force applied were not documented accurately in the majority of simulated deliveries. Use of a preformatted sheet appeared to improve completeness, but not accuracy, of documentation.
The medical record of obstetric emergencies is important for communication between healthcare professionals, audit of care, and medico-legal investigations.1 In 1991, Acker2 stated that careful documentation of a shoulder dystocia delivery should include the exact times of events, description of the manoeuvres used, and an estimation of the traction forces exerted. A 1998 UK review of 56 perinatal deaths that followed shoulder dystocia concluded that events during delivery were often inadequately recorded and emphasised the need for a clear complete contemporaneous record.3
For documentation to be useful, it is must be both comprehensive and accurate. The Royal College of Obstetricians and Gynaecologists (RCOG) Shoulder Dystocia Guideline included a preformatted sheet for the documentation of care given during shoulder dystocia.4 Although this might be expected to act as aide-memoire for completeness of documentation, its use might not necessarily improve accuracy.
The aim of this study was to investigate the completeness and accuracy of note keeping following simulated shoulder dystocia by comparison of a written record with a video tape of the simulation and an electronic record of applied force, and furthermore to compare the quality of record keeping with and without the use of a preformatted sheet.
This study forms part of a large trial (Simulation and Fire-drill Evaluation [SaFE]), commissioned by the Department of Health of England and Wales to investigate training for obstetric emergencies. As part of the SaFE study, each participant managed a standardised shoulder dystocia scenario before and after training. We have previously reported improvement (from 43 to 83%, P < 0.001) in management of the scenario after training.5
Participants (junior and senior midwives [less than and greater than 5 years midwifery, respectively] and junior and senior doctors [less than and greater than 3 years obstetrics, respectively]) were recruited from six hospitals in the South West of England as previously described.5 Training included a 40-minute practical workshop on the management and documentation of shoulder dystocia. For simulations, a prototype PROMPT Birthing Trainer (Limbs and Things Ltd, Bristol, UK) was used together with a patient-actor on a delivery bed. Participants were aware that they would be asked to manage an obstetric emergency, which would be videoed, but were not aware of its nature. Participants were taken into a delivery room individually, given a standardised description of the scenario, and asked to complete the delivery.5 If a participant called for help, an assistant would attend and perform suprapubic pressure if requested but would not take over or complete the delivery. The assistant did not document any timings during the scenario. The simulation was continued until delivery had been achieved through delivery of the posterior arm, the participant had chosen to stop, or 5 minutes had elapsed. Each simulated delivery was recorded on video using four overhead cameras. The fetal mannequin housed a strain gauge embedded in the upper torso of the fetal mannequin connected to the centre of the fetal head by a steel cable. Force data from each simulation were transferred to a bespoke computer program through the umbilical cord and then as a Comma-Separated-Variable file to an Excel® spreadsheet for later analysis. Force was measured in Newtons (N).6
At the end of the simulation, each participant was asked to document the delivery. If the base hospital in which the participant worked used a preformatted sheet, this was used for documentation, but if the hospital did not use a preformatted sheet, participants were asked to use standard hospital notepaper. Preformatted sheets used in the four different hospitals each differed slightly, but all were based on an example contained in the RCOG Shoulder Dystocia Guideline.4
This analysis was confined to those participants who achieved delivery after training. The accuracy and comprehensiveness of documentation were assessed using the following outcome criteria by comparing those documented with those derived from digital recordings of the simulations: (1) delivery manoeuvres, (2) head-to-body delivery interval (HBDI), (3) fetal head position during shoulder dystocia (to determine the anterior shoulder with reference to the maternal pelvis), and (4) force applied. Comparisons between documentation on preformatted sheets and plain hospital notepaper were also made.
Statistical analyses comparing the proportion of participants documenting actions on preformatted sheets and plain hospital notepaper were performed using a chi-square test. A 5% level of significance was used throughout. The statistical software used was Stata vs 8 (StataCorp, College Station, TX, USA, 2003).
Ethical approval was granted by the Regional Research Ethics Committee and five local Research Ethics Committees granted site-specific approval. Research and Development approval was granted by each Healthcare Trust.
The simulation was documented by all 110 participants (36 junior midwives, 35 senior midwives, 19 junior doctors, and 20 senior doctors) who had achieved delivery. All records were suitable for analysis. Four of the six participating hospitals used a preformatted sheet for the documentation of shoulder dystocia as recommended by the RCOG; 78 (70.9%) of 110 participants completed documentation on a preformatted sheet and 32 (29.1%) used plain hospital notepaper.
The term shoulder dystocia was used explicitly in 90/110 (81.8%) of all deliveries. In deliveries in which the term shoulder dystocia was not used, ‘difficulty with the shoulders’ was documented in 12 (10.9%), but there was no mention of delivery of the shoulders in 8 (7.3%). Shoulder dystocia was more likely to be used if a preformatted sheet was used 78/78 (100%) compared with 12/32 (38%) when the delivery was documented on plain hospital notepaper (P < 0.0001).
McRoberts’ position was adopted in 109/110 simulations and documented to have been performed in 107/110 simulations. For the delivery in which McRoberts’ position was not used, the delivery record correctly reflected this omission. For the two simulations in which it was performed but not recorded, both were written on plain hospital notepaper rather than on a preformatted sheet.
Suprapubic pressure was not performed in 6 of the 110 simulations. Three of these six deliveries (all on a preformatted sheet) documented that suprapubic pressure had been performed when it had not. Conversely, after one delivery, the participant, using hospital notepaper, failed to record that suprapubic pressure had been performed when it had been used. Fundal pressure was performed during two deliveries, but this was not recorded in either case, in both cases with a preformatted sheet.
Overall, the HBDI was documented after 64 (58.2%) deliveries. There was no significant difference in the proportion of deliveries in which the HBDI was documented using a preformatted sheet compared with plain hospital notepaper (44/78 [56%] versus 20/32 [63%]; P = 0.707). The documented HBDI was compared with the actual delivery time measured from the video recording. One HBDI, recorded as negative time (head delivered at 13:48 and body delivered at 13:44), was excluded from the analysis. Twenty-six of the 63 (41%) documented the HBDI to within 1 minute of the true value (range −46 to 58 seconds). Thirty-five (56%) of 63 overestimated the HBDI by more than 1 minute (mean overestimation 139 seconds, range 5–497 seconds). Two (3%) of 63 underestimated the HBDI by more than 1 minute (75 and 77 seconds) (Figure 1).
The force used during resolution of the shoulder dystocia was described by 78 (70.9%) of the 110 participants. Force applied was more likely to be mentioned if a preformatted sheet was used than if the written record was made on plain hospital notepaper (61/78 [88%] versus 17/32 [53%]; P = 0.016). Force was described as ‘minimal’ by 2 of 78 (3%), ‘gentle’ in 3 (4%), ‘normal’ in 3 (4%), ‘routine’ in 62 (79%), ‘moderate’ in 7 (9%), and simply as ‘traction applied’ in 1 (1%). There was no simple relationship between the description of the traction applied and the maximum force recorded during simulations (Table 1); 4/5 (80%) of those who stated gentle or minimal applied more than 100 N, a force that has been associated with development of brachial plexus injury in vivo.7 Twelve (10.9%) participants applied more than 200 N, twice the nominal threshold, but none described the applied force as being excessive.
Table 1. Mean maximum applied force according to description of traction
Description of force
Mean actual maximum applied force (range) (N)
n = 30 force not recorded in two simulated deliveries.
n = 61 force not recorded in two simulated deliveries.
Overall, the laterality of the anterior arm at the time of delivery was documented by 86 (78.2%) of the 110 participants. Use of a preformatted sheet was associated with a greater likelihood that the fetal position was recorded; 72/78 (92%) for a preformatted sheet compared with 14/32 (44%) for plain hospital notepaper (P < 0.0001). Of those who documented the anterior shoulder at the time of delivery, 69/86 (80%) identified the correct side. Documentation of the laterality of the anterior arm was more accurate if plain hospital notepaper was used, 13/14 (93%) compared with 56/72 (78%) using a preformatted sheet, but this was not statistically significant (P = 0.3526).
Shoulder dystocia is an unpredictable emergency with a high risk of serious maternal and neonatal complications. The medical record is used to inform decisions about the management of future deliveries, and it may be relied upon to resolve medico-legal actions. A key question in law is scriptum verum factum –does the document reflect the facts? Our data suggest that the immediate written recollection of events by the accoucheur is often both incomplete and inaccurate.
The data are likely to be representative of current UK practice with a large sample of midwives and doctors randomly allocated to the SaFE study from six hospitals across a large health region. Availability of video and force recording for each simulation enabled direct comparison with actual events. Even so, interpretation of the findings should be undertaken with some caution. This was a simulated shoulder dystocia and might not be representative of real life. Our impression is that the participants took the exercise seriously, and it was evident that they experienced many of the feelings raised by the real emergencies. It could also be argued that shoulder dystocia is usually managed by a team rather than by a single individual. Some might propose that the medical record is a collaborative effort but that would not necessarily overcome the problem of inaccurate recall.
A previous evaluation of delivery notes after simulated shoulder dystocia also found that notes often lacked critical elements,8 and the authors recommended training in documentation as part of learning packages for shoulder dystocia. One advantage of our study is that all the participants had been given guidance on documentation of shoulder dystocia during a 40-minute training session 1–3 weeks prior to this evaluation. The study was not designed to determine the effectiveness of our training package, but it is interesting to note that our participants achieved higher rates of compliance compared with previous published findings;8 78% documented laterality of the anterior shoulder, for example, compared with just 18% in the study by Deering et al.8 These data suggest that our participants were not unusually poor at documentation but were perhaps better than average.
Completeness of the description of manoeuvres was generally good, but there were significant deficiencies for other aspects of the delivery. Almost half failed to record the HBDI and one in five failed to record the laterality of the anterior shoulder. The use of a preformatted sheet was associated with a significantly better rate of compliance for the anterior shoulder, but it appeared to have no effect on documentation of the HBDI.
The accuracy of the documentation also gave cause for concern. One in five incorrectly identified the contralateral shoulder as being anterior. This has particular repercussions for medico-legal case analysis as brachial plexus injury of the posterior shoulder is considered unlikely to be the result of clinical negligence.9,10
The participants also had difficulty in estimating the duration of the shoulder dystocia with less than half of participants documenting time to within 1 minute of the actual time. Nearly, six in ten of participants who documented the HBDI overestimated the time by more than 1 minute, with the greatest overestimation of the HBDI being by more than 8 minutes. The tendency to overestimate the HBDI may be a result of stress. A relationship between anxiety and overestimation of time is well established,11 with more stressful situations tending to produce a greater overestimation than lesser ones.12 At the end of the simulation, many participants said ‘I looked at the clock but I didn’t register what it said’. Attentional narrowing during acute stress has also been described,13 and looking-without-seeing during management of emergencies may be the result of this phenomenon. Given the relatively poor estimation of duration, it should be questioned whether mental noting of the time is an additional stress burden without value. The documented HBDI does not necessarily inform future deliveries and importantly does not accurately reflect the severity of any neonatal acidosis with some babies dying within 5 minutes and others being born in good condition long after this time.3,14
Even more striking is the difficulty participants had in subjectively assessing the degree of force used. There was no relationship between the adjective used and the mean maximum electronically logged force. This may be partly due to an inability to approximate the applied force, especially during a stressful situation. Gurewitsch et al.15 have previously demonstrated that while clinicians are intuitive about their applied traction in laboratory conditions, during simulated shoulder dystocia, clinicians’ actual applied force was over twice their own estimate. Our data tend to support that view. Force data collected during the SaFE evaluation simulations demonstrated that larger forces tended to be applied as the delivery time increased,6 raising the possibility that increasing stress may distort perception of the applied force. It is striking that none of the written records in our study reflected that. It is possible that the preformatted documentation sheets used in the participating hospitals may have been restrictive and did not offer the accoucheur the option of describing their applied force in their own words. A free text box labelled ‘description of applied delivery force’ might solve this problem, although other alternatives might also have to be explored as documentation of force on plain notepaper was equally as inaccurate as the proformas.
We wished to compare the standard of note keeping with preformatted sheets and plain paper. As the participants already had experience of using one or the other, this was not suitable for a random allocation trial. We adopted a pragmatic design using the standard documentation used in the base hospital. Documentation on preformatted sheets tended to be more comprehensive, more likely to state that shoulder dystocia had occurred, the magnitude of the force applied, and the fetal position, but not the HBDI. The increase in completeness of the medical record with the aid of a preformatted sheet might be at the expense of accuracy. Three participants using a preformatted sheet documented use of suprapubic pressure when it had not been used. In addition, 93% of participants who documented the fetal position on hospital notepaper recorded the fetal position correctly compared with only 78% of participants who used preformatted sheets. Each system appears to have different merits: writing longhand may encourage more careful thought than simply ticking preformatted boxes. A preformatted sheet appears to act as an aide-memoire for more comprehensive documentation. However, it may encourage accoucheurs to document things of which they have no true recollection. Equally, issues not covered by a preformatted sheet may be overlooked such as the failure to record application of fundal pressure.
It is intuitive that documentation of shoulder dystocia management may be more accurate if it is completed contemporaneously by a scribe who is not directly involved in the emergency, although this hypothesis has not been tested. We believe shoulder dystocia management training should be multiprofessional and should highlight that if there are enough personnel in attendance, one member of the team should document what is happening as, and when, it is occurs.
Documentation is an important component of health care; inaccurate or incomplete record keeping could negatively influence future care and the outcome of litigation. Our data suggest that recording emergencies is not straightforward and that the quantifiable elements such as delivery interval and force applied are particularly difficult. A preformatted sheet appears to bring the advantage of improved completeness but perhaps at the expense of reduced accuracy in some areas. The challenge is to produce a comprehensive sheet that acts as an aide-memoire but without prompting inaccurate documentation.
Conflict of interests
T.J.D. invented the mannequin used in the study and is a consultant to Limbs and Things Ltd, manufacturers of the PROMPT Birthing Simulator. None of the other authors own stock, or hold stock options, in any obstetric emergency training companies.
This study was funded as part of the SaFE study (Simulation and Fire-drill Evaluation) by the National Patient Safety Research Programme. The research team are independent of the National Patient Safety Research Programme.
Details of ethics approval
Ethical approval was granted from a Regional Research Ethics Committee (04/Q2103/68) and site-specific ethical approval was granted by the five appropriate local Research Ethics Committees. Research and Development approval was granted by all the Healthcare Trusts involved in the study.
Contribution to authorship
J.F.C., D.E., and C.B. conducted the study. J.F.C. analysed the data. T.J.D. designed the study. J.F.C., T.J.D., and R.F. wrote the paper.
Training Team: Fiona Donald (Consultant Obstetric Anaesthetist), Mark James (Consultant Obstetrician and Gynaecologist), Imogen Montague (Consultant Obstetrician and Gynaecologist), Cathy Winter (Midwife)
Local Hospital Support: Cheltenham General Hospital: Penny Watson (Midwife), Anne McCrum (Consultant Obstetrician and Gynaecologist); Gloucestershire Royal Hospital: Sarah Read (Midwife); Taunton and Somerset Hospital: Heather Smart (Midwife), Melanie Robson (Consultant Obstetrician and Gynaecologist); Royal Devon and Exeter Hospital: Katie Harrison (Midwife), Neil Liversedge (Consultant Obstetrician and Gynaecologist); Royal Cornwall Hospital: Joanne Crocker (Midwife), Simon Grant (Consultant Obstetrician and Gynaecologist)
The authors would like to thank all staff who participated in, or helped to conduct, the SaFE study.