Impact of simulation training on decision to delivery interval in cord prolapse

Abstract

Background

Umbilical cord prolapse is a rare obstetric emergency requiring rapid coordination of a multidisciplinary team to effect urgent delivery. The decision to delivery interval (DDI) is a marker of quality of teamwork. Multidisciplinary team simulation-based training can be used to improve clinical and teamwork performance.

Aim

To assess the DDI for cord prolapse before and after the introduction of simulation-based training at a quaternary maternity unit in Australia.

Method

A retrospective, observational cohort study comparing the DDI before and after the introduction of simulation-based training activities. The general linear model was used to estimate the association between DDI and simulation training while adjusting for potential confounders including model of care (public or private) and time of birth (regular or after hours).

Results

After the introduction of simulation training, mean DDI decreased by 4.1 min (difference −4.1, 95% CI −6.2 to −1.9), after adjustment for confounding factors. Despite this, there was no difference in selected neonatal outcomes including Apgar score at 5 min and arterial cord pH.

Conclusions

The introduction of simulation-based training was associated with a decrease in the DDI in the setting of cord prolapse.

What is already known on this subject.

• Previous studies into the impact of simulation training on decision to delivery interval (DDI) for management of cord prolapse have yielded conflicting results. Interprofessional simulation-based training has been demonstrated to improve DDI for cord prolapse in a large maternity unit in England; however, these findings were not reproduced in a more recent Australian study.

What this study adds.

• This study demonstrated that the introduction of simulation-based training was associated with a decrease in the DDI in the setting of cord prolapse. This decrease suggests improvements in multidisciplinary teamwork in the setting of a rare obstetric emergency.

Introduction

Umbilical cord prolapse is a rare obstetric emergency occurring in <0.5% of deliveries and is associated with increased perinatal morbidity and mortality.¹ Urgent management is required to relieve pressure on the cord and deliver the baby to ensure optimal outcome for the baby.² Unless cord prolapse has occurred at full dilatation and vaginal delivery is imminent, this necessitates rapid coordination of a multidisciplinary team required for an emergency caesarean section. The time frame between decision for caesarean section and delivery of the neonate, or decision to delivery interval (DDI), is considered a marker of quality of teamwork.³ Short DDIs are reflective of rapid and effective clinician communication and coordination across the birth suite and the operating theatres and therefore are a reasonable marker of teamwork more generally.

Simulation training is increasingly used to improve teamwork in healthcare and has been demonstrated to be associated with improvements in patient management,^4–6 patient outcomes^7–9 and patient perceptions of care.^{10 11} More recently, simulation-based training has been demonstrated to reduce maternal mortality in a country with a high maternal mortality ratio.¹² Interprofessional simulation-based training has been demonstrated to improve DDI for cord prolapse in a large maternity unit in England, but this has not been established in an Australian setting.^{3 4} A recent Australian-based retrospective cohort study assessing DDI preintroduction and postintroduction of multidisciplinary simulation-based training demonstrated improvements in documentation but no difference in DDI.⁴ More broadly, improvements in patient outcomes relating to multiprofessional simulation training have been demonstrated in a UK setting following introduction of the PRactical Obstetric Multi-Professional Training (PROMPT) programme^{8 13}; however, the effect on clinical outcomes following PROMPT introduction in an Australian setting was less.¹⁴

Despite these mixed results, professional bodies overseeing obstetric training specifically recommend multidisciplinary ‘team rehearsals’ for optimisation of management of cord prolapse.²

This study aims to provide further data to assess the potential impact of simulation training on teamwork and patient care in an Australian setting by examining DDI for cord prolapse before and after the introduction of an interprofessional simulation-based training programme.

Materials and methods

A retrospective, observational cohort study was completed comparing DDI prior and subsequent to the introduction of simulation-based training at a quaternary obstetric unit.

Setting

This Australian hospital is a mixed public/private unit delivering approximately 10 000 babies annually. Three types of simulation-based training activities were commenced concurrently in 2013: the Maternity Emergency Management (MEM) course (an 8-hour course for midwives, obstetric doctors and anaesthetic doctors held monthly at the colocated simulation centre), monthly ‘mini-MEMs’ (or ‘pop up simulations’, lasting 15 min) held in the clinical areas and advanced MEMs held quarterly in the simulation centre, which focused on advanced teamwork and clinical skills.

Population

The study population includes all women who underwent caesarean delivery for cord prolapse from 24+0 weeks’ gestation between 2008 and 2019. Cases were excluded if the pregnancy was affected by a known lethal abnormality or if fetal demise occurred prior to or at the time of diagnosis of cord prolapse. Data were extracted from an obstetric database and checked for accuracy against the health record. Twenty per cent of charts were reviewed independently by a second researcher to verify the recorded DDI.

The primary study outcome was DDI in minutes. Additional information collected included demographic data and potential confounders of DDI including whether the birth occurred during or after hours, model of care (public/private), body mass index (BMI), previous caesarean section (yes/no) and mode of anaesthesia (regional/general). A priori it was acknowledged that a change in after-hours obstetric cover for public patients that occurred in 2013 (a move to public consultant on-site 24/7) may have affected results, and subgroup analysis by model of care and time of birth was planned. Additional outcome measures collected were neonatal Apgar score at 5 min, arterial cord pH at birth and admission to the neonatal nursery (if weight ≥2500 g).

Statistical methods

Data are presented as number (%) for categorical variables and mean (SD) for continuous variables. Differences in clinical characteristics and neonatal outcomes other than DDI presimulation or postsimulation training were tested using Pearson’s χ² test. The general linear model was used to estimate the unadjusted and adjusted association between DDI and simulation training. Log transformation of DDI was considered but judged not to be necessary because the distribution of DDI was reasonably close to normal, the sample size was adequate and linear models are robust to minor deviations in normality. It was also considered important to estimate the effects in terms of minutes rather than proportions, which could not be done if data were log transformed.

Variables in the adjusted model were model of care (public/private), time of birth (during or after hours) and any additional variables that were either associated with DDI at p<0.05 or confounded the association between DDI and simulation training. A three-way interaction term between model of care, time of birth and simulation training was tested to determine whether the association between DDI and simulation training was different for public and private patients during and after hours. Statistical analysis was performed using Stata V.15.1 (Stata Corp), and a significance level of 0.05 was used throughout inferential analysis.

Results

There were no substantial differences in model of care (public or private), time of birth (regular or after hours), BMI or history of a previous caesarean section between births that occurred before and after the implementation of simulation training. There was a higher proportion of nulliparous women, induced labour, preterm babies and low birth weight babies in the postsimulation training cohort (table 1). The type of anaesthetic used (general or regional) was not different between cohorts. However, it was noted that regional anaesthesia was more commonly used after hours (18 of 59 after hours births, 31%) than it was during regular hours (4 of 43 births during regular hours, 9%, p=0.01). This proportion was similar before and after simulation training.

Table 1.

Characteristics of 102 births in women who underwent caesarean section for cord prolapse before (2008–2012) and after (2013–2019) the implementation of simulation training

	Total	Pre-SIM	Post-SIM	P value*
	N=102	N=41	N=61
Public or private, n (%)				0.39
Public	72 (71)	27 (66)	45 (74)
Private	30 (29)	14 (34)	16 (26)
After hours birth, n (%)				0.13
Regular hours	43 (42)	21 (51)	22 (36)
After hours	59 (58)	20 (49)	39 (64)
Anaesthetic type, n (%)				0.94
General anaesthesia	80 (78)	32 (78)	48 (79)
Regional anaesthesia	22 (22)	9 (22)	13 (21)
Maternal BMI, n (%)†				0.58
18–<25 kg/m2	49 (51)	16 (44)	33 (55)
25–<30 kg/m2	27 (28)	12 (33)	15 (25)
≥30 kg/m2	20 (21)	8 (22)	12 (20)
Previous CS, n (%)				0.64
No previous CS	89 (87)	35 (85)	54 (89)
Previous CS	13 (13)	6 (15)	7 (11)
Parity, n (%)				0.011
0	49 (48)	13 (32)	36 (59)
1–2	38 (37)	18 (44)	20 (33)
3–7	15 (15)	10 (24)	5 (8)
Gestational age, n (%)				0.041
Preterm (<28 weeks)	33 (32)	18 (44)	15 (25)
Birth weight, n (%)				0.029
Low birth weight (<2500 g)	30 (29)	17 (41)	13 (21)
Induction of labour, n (%)				0.022
Spontaneous	38 (37)	18 (44)	20 (33)
Induced	41 (40)	10 (24)	31 (51)
No labour – CS	23 (23)	13 (32)	10 (16)

*Pearson’s χ² test.

†Missing for n=6.

BMI, body mass index; CS, caesarean section;; SIM, simulation training.

Decision to delivery interval

Unadjusted results for DDI presimulation and postsimulation training are presented in table 2 and graphically represented in figure 1. Variables in the final adjusted model included public or private, during or after hours and general or regional anaesthesia. In the adjusted model, DDI was on average 4.1 min shorter after the implementation of simulation training (95% CI −6.2 to −1.9, p=0.002). DDI for private patients was on average 2.5 min shorter than for public patients (95% CI −4.8 to −0.2 p=0.03), and DDI for regional anaesthesia was 3.9 min longer than for general anaesthesia (95% CI 1.3 to 6.5, p=0.004). There were minimal differences in DDI between women who presented during regular versus after hours (1.5 min longer, 95% CI −0.7 to 3.7, p=0.17).

Table 2.

Unadjusted decision to delivery interval (DDI) for 102 births in women who underwent caesarean section for cord prolapse before (2008–2012) and after (2013–2019) the implementation of simulation training

	DDI (min)		P value
	Mean (SD)	Difference in means (95% CI)
Simulation training
Pre	17.8 (7.2)	Reference
Post	14.1 (4.4)	−3.7 (−5.9 to −1.4)	0.002
After hours
No	14.4 (6.0)
Yes	16.5 (5.7)	2.1 (−0.2 to 4.4)	0.08
Patient type
Public	16.3 (6.4)	Reference
Private	14.0 (4.2)	−2.3 (−4.8 to 0.2)	0.07
Anaesthesia
General	14.7 (5.0)	Reference
Regional	19.0 (7.6)	4.4 (1.6 to 7.1)	0.002
Previous CS
No	15.4 (5.9)	Reference
Yes	16.8 (6.0)	1.4 (−2.1 to 4.8)	0.44
BMI
18–<25 kg/m2	15.6 (6.8)	Reference
25–<30 kg/m2	14.7 (4.5)	−0.9 (−3.8 to 2.0)
≥30 kg/m2	16.4 (5.5)	0.8 (−2.4 to 4.0)	0.63

General linear model with regression of DDI on each of the independent variables separately.

BMI, body mass index;;CS, caesarean section.;

Figure 1.

Histograms of decision to delivery interval (DDI, min) for 102 births in women who underwent caesarean section for cord prolapse before (2008–2012) and after (2013–2019) the implementation of simulation training (SIM).

Change in on-call arrangements

A change in after-hours obstetric cover for public patients (a move to public consultant on-site 24/7) occurred at a similar time to the simulation training in 2013 so it was important to confirm that the decrease in DDI was not solely in public patients with after-hours births. An interaction term between simulation training, model of care and time of birth was not statistically significant (p=0.25) indicating that any differences between these groups could be due to chance. However, because a priori there was interest in the differences between these groups, we examined the mean DDI before and after simulation training separately by model of care and time of birth.

On average, DDI decreased in all groups, and DDI after simulation training was similar in all groups (table 3). However, private patients with births during regular hours had shorter DDIs than the other groups before the introduction of simulation training and DDI did not decrease as much for this group after simulation training as it did for private patients with after-hours births or for public patients.

Table 3.

Decision to delivery interval for 102 births in women who underwent caesarean section for cord prolapse before (2008–2012) and after (2013–2019) the implementation of simulation training, by model of care and time of birth

	Decision to delivery interval (min, mean (SD))		Difference (min)
	Pre	Post
Public
Regular hours	17.2 (7.7)	13.1 (6.0)	4.1
After hours	20.7 (7.7)	15.3 (4.1)	5.4
Private
Regular hours	13.9 (4.6)	12.5 (2.7)	1.4
After hours	17.5 (6.0)	12.9 (2.4)	4.6

Additional outcomes

There were no differences in neonatal Apgar score at 5 min (0–6 or 7–9), arterial pH (<7.1 or ≥7.1) or admission to the nursery (for neonates with birth weight ≥2500 g) between neonates that were born before and after the implementation of simulation training (table 4).

Table 4.

Neonatal outcomes of 102 births in women who underwent caesarean section for cord prolapse before (2008–2012) and after (2013–2019) the implementation of simulation training

	Total	Pre-SIM	Post-SIM	P value*
	N=102	N=41	N=61
	N (%)	N (%)	N (%)
Apgar score at 5 min				0.87
7–9	83 (82)	34 (83)	49 (82)
0–6	18 (18)	7 (17)	11 (18)
Arterial pH<7.1				0.74
No	59 (70)	19 (68)	40 (71)
Yes	25 (30)	9 (32)	16 (29)
Admitted to nursery (for neonates ≥2500 g only)				0.60
No	48 (67)	17 (71)	31 (65)
Yes	24 (33)	7 (29)	17 (35)

*Pearson’s χ² test. Missing for n=1 (Apgar score), 18 (arterial pH) and 30 (admitted to nursery: appropriately missing for 30 neonates <2500 g).

SIM, simulation training.

Discussion

Previous studies into the impact of simulation training on DDI have yielded conflicting results.^{3 4} The findings of this study support the findings of a similar study in England, which also demonstrated improvements in DDI for caesarean delivery for cord prolapse following the introduction of simulation training.³ As shown in figure 1, not only did the DDI shift overall to the left postsimulation training, there was an elimination of DDI outliers that might be considered more clinically significant than a modest overall reduction in DDI.

We can be confident that changes in DDI were not just due to changes in on-call arrangements for public patients since changes were also seen in public and private patients with births during regular hours. Private patients during regular hours did not have as large a decrease in DDI as other groups, which could be because they had shorter DDIs to start with and so further improvement would be more difficult to achieve.

Despite improvement in DDI, there was no change in measured neonatal outcomes that again was consistent with both the English study findings³ and those of a previous Australian-based study.⁴ Previous studies have also observed a poor correlation between DDI and neonatal cord pH.¹⁵ Although the goal of simulation training is to improve patient safety, the focus of this study was DDI, and therefore, further research into the impact on neonatal outcome measures following simulation training is warranted.

Despite the lack of improvement seen in neonatal outcomes, DDI has historically been regarded as a marker of efficient teamwork¹⁵ and quality maternity care.¹⁶ A recent systematic review assessed the impact of simulation training in obstetrics using the Kirkpatrick model of evaluation.¹⁷ This model can be applied to classify the effectiveness of simulation training across increasing translational levels: level 1: reactions, level 2: learning, level 3: behaviour and level 4: outcomes. The improvement in DDI could be considered a level 4 impact using the Kirkpatrick model based on the improved outcome as a reflection of quality team work and in keeping with classification of similar studies assessed by the systematic review.¹⁷

Assessment of behavioural changes remain important as UK National Enquiries and previous studies have identified behavioural deficiencies including poor communication and role confusion to be associated with poor outcomes,^{4 18} with improvements in teamwork behaviours associated with improved outcomes.³ Obstetric simulation training has been demonstrated in other studies to improve team awareness, knowledge and markers of clinical care.^{5 13} While this study did not specifically examine teamwork behaviours, an unpublished survey in 2018 of midwives and doctors working in maternity care within this unit reported 94% of staff agreed or strongly agreed that patients were safer as a result of simulation activities (data available on request from corresponding author). While these results are encouraging, objective evidence of neonatal care improvements at Kirkpatrick level 4 is still required to demonstrate return on investment for simulation.

An unexpected finding in this study was the greater number of patients receiving regional anaesthesia after hours compared with those who birthed during regular hours. The reason for this difference is unclear from this study. Overall, the number of general anaesthetics performed for caesarean section is low and declining, which may result in reduced exposure of anaesthetic trainees to this technique.^{19 20} It may be that junior anaesthetic staff who may be working without direct consultant supervision may be more comfortable with a regional anaesthetic as their first-line approach. Contradictory to our findings, a recent study demonstrated an increase in general anaesthetics for births after hours.¹⁹ This raises an essential area of potential future research into the impacts of afterhours anaesthetic decision making.

Strengths of this study include the comparable number of cases preintervention and postintervention, relative to other studies examining for impact of simulation training on DDI and neonatal outcomes. The robust nature of the data collection strategy was another strength of this study. An upgrade occurred within the perinatal data collection system prior to the commencement of the time period being assessed (2008), which allowed for reliable identification of relevant cases. Regular auditing of the perinatal data system occurs within the unit to ensure precision of data. Individual case data were cross-checked for accuracy against the health record, and an additional 20% of charts were independently reviewed by a second researcher to limit bias in data collection. DDI for cord prolapse was chosen as the outcome in this study as it is felt that it is the least susceptible (although not immune) to the influence of other factors. Cord prolapse management has not changed over the last several decades, and diagnosis and management are unambiguous, making heightened awareness of the diagnosis or management unlikely to have changed over this time period.

Despite these strengths, studies performed over long time periods may have limited ability to attribute outcomes to a single intervention. As this was a retrospective observational study evaluating DDI before and after the implementation of simulation training, we cannot be sure that simulation training caused the reduction in DDI. It is possible that other changes not accounted for in this study could have confounded the results. While birth numbers have increased over time, resulting in an increased workload, staff-to-patient ratios have not changed in the birth suite environment where women are provided with 1:1 care. In the setting of competing workload, cord prolapse is such an acute emergency; it would almost universally take priority over other cases being concurrently managed. The change to a public on-site consultant afterhours was examined separately, and while this may have contributed to a reduction in DDI in public women birthing after hours, similar reductions in DDI were seen in births, which were unaffected by this change.

Conclusion

DDI for women who underwent caesarean section for cord prolapse was shorter after the implementation of simulation training suggesting improved teamwork. There were no differences in selected neonatal outcomes. Ongoing research into the impacts of simulation training on patient management and patient outcomes is required.

Data availability statement

Data are available on reasonable request. Deidentified individual participant data that underlie the results reported in this article will be available for up to 3 years after publication. Data will be shared with investigators who propose use of the data under a methodologically sound proposal. Proposals should be sent to gillian.gallagher@health.qld.gov.au. To gain access, data requestors will be required to sign a data access agreement.

Ethics statements

Patient consent for publication

Not required.

Ethics approval

Ethics approval for the study was obtained (HREC/MML/55947).