Interprofessional obstetric simulation training improves postpartum haemorrhage management and decreases maternal morbidity: a before‐and‐after study

To determine the impact of the Obstetric Simulation Training and Teamwork (OB‐STaT) curriculum on postpartum haemorrhage (PPH) rates and outcomes.


| I N TRODUC TION
Postpartum haemorrhage (PPH) complicates 6% of deliveries worldwide and remains a leading cause of treatable maternal morbidity and mortality in the USA. 1 Obstetric emergencies such as PPH are relatively rare and unpredictable.7][8] Prior studies assessing the impact of simulation training for PPH have shown decreased blood transfusion rates, 6 and have shown increased rates of invasive treatment for severe PPH after simulation team training. 9,10 recent systematic review of obstetric emergency training noted that the simulation effects on severe PPH were unclear, although in situ multiprofessional simulation-based training had the best impact on decreasing severe blood loss (≥1500 ml) and decreasing transfusions of ≥4 units. 10imilarly, a prospective study on multidisciplinary simulation for PPH showed improved responses for patients receiving transfusions who were diagnosed with uterine atony. 11A recent Cochrane Review noted that multiprofessional simulation-based team training in obstetric emergencies may improve team performance and contribute to improved outcomes, compared with no training; however, high-certainty evidence is lacking. 12he Military Health System (MHS) performs 40 000 annual deliveries with 49% of inpatient care related to pregnancy, 13 and PPH remains an important area for simulation training.PPH patient safety bundles were developed and implemented at all delivery sites in the MHS.These bundles require perinatal team members to participate in semi-annual simulation drills related to PPH. 14 However, these drills are not standardised in terms of scenario content, equipment or the expertise required to conduct training.Consequently, the true effects of MHS simulation drills on clinical outcomes are understudied.A new interprofessional in situ simulation training programme, Obstetric Simulation Training and Teamwork (OB-STaT), was developed to provide standardised training on the management of PPH, with appropriate recognition and interventions related to the underlying causes of haemorrhage.The curriculum included simulation scenarios with team debriefs and skills stations on massive transfusion protocols, rapid transfusion systems, uterine balloon tamponade devices, uterine compression sutures, uterine artery ligation, peripartum hysterectomy and neonatal resuscitation. 15The study objective was to determine the impact of the OB-StaT curriculum on PPH rates, outcomes (use of uterotonics, tranexamic acid, transfusion, hysterectomy, length of stay and composite maternal morbidity) and team performance in a sample of maternity care hospitals in the MHS.We hypothesised that OB-StaT, with its focus on appropriate recognition and response, would decrease PPH rates and improve PPH management and clinical outcomes.

| Overview
This was a multi-site before-and-after study conducted in the USA at eight maternity care hospitals in the MHS that provided inpatient obstetric care from February 2018 to November 2019.The MHS is a large network of military hospitals and clinics that provides medical care to active-duty service members and eligible beneficiaries, covering 40 000 annual deliveries.Care is provided to service members and their families at no charge, as part of their military service benefits.The hospitals in this study included tertiary care hospitals and smaller community hospitals, with annual delivery volumes ranging from 200 to 3000.The study was approved by the Institutional Review Board at the Naval Medical Center, Portsmouth, with a wavier for informed consent.The study conforms with the US Federal Policy for the Protection of Human Subjects.
A retrospective chart review was conducted to collect baseline PPH rates and clinical outcomes for the 6 months preceding simulation training at each site.A 4-h interprofessional education (IPE) simulation-based curriculum, OB-STaT, 15 was implemented at the sites to train obstetric team members that respond to obstetric emergencies, including obstetrics & gynaecology physicians, obstetrics & gynaecology residents, certified nurse midwives, nurses, anaesthesiologists, anaesthesia residents, certified and student nurse anaesthetists, paediatricians, paediatrics residents, family medicine physicians, family medicine residents, hospital corpsmen (medical assistants), blood bank staff, radiology technicians and operating room technicians.Following completion of OB-STaT training, a retrospective chart review was conducted of deliveries for the next 6 months at all sites.The primary outcome was PPH (defined as blood loss of ≥1000 ml at delivery).Secondary outcomes included use of uterotonic medications or tranexamic acid, transfusion of blood products, hysterectomy, length of stay and composite maternal morbidity.

| OB-STaT training intervention
The training intervention was implemented in a rolling fashion across the eight sites, from February 2018 to November 2019.Mobile training teams implemented OB-STaT training over 2-5 days at each site to maximise the number of participants and to minimise any impact on patient care.Each participating team completed a 4-h in situ interprofessional team-based PPH simulation training intervention, in a labour and delivery room and corresponding labour and delivery operating room, at each institution.The curriculum and its impact on team performance and communication have been reported previously. 15

| Data collection and patient-level outcomes
Prior to data collection, we reviewed the electronic medical record (EMR) at each site to ascertain where data elements were located because of variations in data fields between sites.Two of the sites were using an updated EMR that was unable to support electronic data retrieval; patient records from these sites were individually reviewed by clinical research coordinators who were trained as data abstractors.Deliveries were identified by ICD-10 (International Classification of Diseases 10th Revision) delivery codes and pulled from the EMR.Deliveries where blood loss was not recorded were excluded from the analysis.Clinical research coordinators reviewed the EMR of all deliveries.Data were abstracted from the admission history, physical examination, delivery summary, discharge summary and any additional operative reports.The first month of data following OB-STaT training at each site was excluded as a washout period.
The data collected included: PPH cases (estimated blood loss, EBL, or quantitative blood loss, QBL, of ≥1000 ml), type and cross, number of types of uterotonic medications used (oxytocin, carboprost tromethamine, methylergonovine and misoprostol), tranexamic acid use, hysterectomy prior to hospital discharge and postpartum length of stay (calculated from date of delivery to date of discharge).A composite maternal morbidity variable was calculated, which included one or more of the following: PPH, postpartum hysterectomy, transfusion of ≥4 units of blood products or intensive care unit admission for PPH.Data from the retrospective chart reviews were aggregated to calculate clinical outcomes of interest prior to the implementation of OB-STaT via the EMR.The change in PPH rates and other clinical outcomes before and after OB-STaT were compared to determine the impact of training on patient outcomes.
The following risk factors were used to assess participant PPH risk: (i) placental abruption; (ii) suspected or confirmed placenta accreta spectrum, placenta praevia or low-lying placenta; (iii) multifetal gestation; (iv) parity greater than four; (v) intrapartum use of magnesium sulphate; (vi) birthweight of >4000 g; and (vii) caesarean delivery or combined vaginal delivery and caesarean delivery.Haemorrhage risk factors were collected from the admission history, physical examination, delivery note and discharge summary using queries of free-text word fields and associated check boxes in the electronic medical record.Patients with two or more antepartum or intrapartum PPH risk factors were classified as being at increased risk of PPH.
The primary outcome was the PPH rate, calculated as the number of deliveries complicated by PPH divided by the total number of deliveries.Secondary outcomes included medical interventions for PPH (use of uterotonic medications, tranexamic acid or transfusion of blood products), surgical/ procedural interventions for PPH (uterine curettage, uterine balloon tamponade, ligation of uterine arteries or hypogastric artery, uterine compression sutures or hysterectomy), composite maternal morbidity, maternal mortality and postpartum length of stay, as shown in Table 1.
The planned primary metric for blood loss assessment was QBL. 16During data abstraction it was discovered that not all sites recorded QBL, as some were still using EBL and others recorded both.Where QBL was documented, QBL was used to determine the total blood loss, otherwise, the EBL documented in the EMR was used for blood loss assessment.No site switched from EBL to QBL during the duration of the study.

| Statistical analysis
Pre-and post-OB-STaT outcomes were compared at all sites to determine the impact of OB-STaT on these outcomes.Univariate analysis was performed using the chi-square test, Fisher's exact test, Student's t-test or Mann-Whitney U-test, as appropriate.A priori, the research team planned to control for increased PPH haemorrhage risk (defined as two or more antepartum or intrapartum PPH risk factors), mode of delivery and neonatal birthweight in the multivariate analysis.Additional variables included in the multivariate analysis that were determined to be significant on univariate analysis were: number of medications used to treat PPH; tranexamic acid use (for treatment and for prophylaxis for PPH); balloon tamponade; blood transfusion; total number of haemorrhage interventions; and length of stay in days.Variables that predicted failure perfectly were dropped from the final model.
The sample size for the study was calculated using the historical delivery volume for all participating hospitals: approximately 6000 deliveries over a 6-month period. 14The historical baseline PPH rate using EBL was 2%. 14A priori, it was determined that 2319 deliveries were needed per 6month period to have 80% power with a two-sided alpha of 5% to detect a decrease in PPH rate from 2% to 1%.Statistical analysis was performed using STATA 14 (StatCorp LLC, College Station, TX, USA) and SPSS (IBM, Armonk, NY, USA).

| R E SU LTS
The training teams conducted the OB-STaT curriculum at eight MHS hospitals during the study.A total of 54 teams were trained, including 721 staff (72% of the assigned staff). 15For the outcomes, a total of 9980 deliveries were included: 5059 deliveries in the 6 months before OB-STaT training and 4921 deliveries in the 6 months after OB-STaT training.There were no significant differences in mode of delivery, birthweight, rate of PPH, surgical interventions for PPH or transfusions before versus after OB-STaT training.
A greater number of subjects that delivered after OB-STaT training were documented to be at increased risk for PPH: 2047 (41.6%), compared with 438 (8.7%) before OB-STaT training (p < 0.001) (Table 1).The overall PPH rate did not decrease significantly following training (5.48% before vs 5.14% after, p = 0.46).
Composite maternal morbidity significantly decreased from 6.35% before to 5.28% after OB-STaT training (p = 0.03), and postpartum length of stay significantly decreased from 2.05 to 2.01 days following OB-STaT training in all deliveries (p = 0.04).The number of PPH interventions increased at participating sites following OB-STaT training (Table 1).On univariate analysis, the number of PPH medications used in all deliveries significantly increased after training (Table 1).Prior to the intervention, 82% of deliveries received no medications for the treatment of haemorrhage, and this decreased to 49% after training.Tranexamic acid (TXA) use in deliveries also significantly increased (2.72% vs 4.76%, p = 0.001) after OB-STaT training.Massive transfusions (≥4 units of blood products) decreased significantly by 57%, from 21 (0.42%) before to nine (0.18%) after (p = 0.043), and hysterectomy procedures significantly increased from zero before to five after training in all deliveries (p = 0.03).A review of hysterectomy cases revealed that one was a planned caesarean hysterectomy with a QBL of 934 ml, and this case was removed from the analysis.Two hysterectomies were for PPH at the time of caesarean (one for atony and the other for uncontrolled uterine bleeding from extensive adhesiolysis), and the other two were for PPH caused by refractory uterine atony after vaginal deliveries.Both hysterectomies occurred within 6 h of vaginal delivery.No significant changes were noted in operative interventions in all deliveries (including uterine curettage, uterine balloon tamponade, uterine compression sutures and uterine artery or hypogastric artery ligation).No significant changes were seen in total blood transfusions or intensive care unit admission rates in all deliveries (Table 1).T A B L E 1 (Continued) Using multivariate analysis, and controlling for delivery location, haemorrhage risk, delivery method and neonatal weight, the following variables demonstrated increased odds of PPH: increased PPH risk (OR 1.46, 95% CI 1.05-2.02),caesarean delivery (OR 6.28, 95% CI 4.65-8.48),increased number of types of uterotonics to treat PPH (OR 1.60, 95% CI 1.43-1.79),use of TXA for prophylaxis or treatment of PPH (OR 1.63, 95% CI 1.27-2.10)as well as TXA used to treat PPH (OR 2.52, 95% CI 1.81-3.52).Uterine balloon tamponade and transfusion use was also associated with an increased odds of PPH but had wide confidence intervals (OR 28.06, 95% CI 3.84-205.02and OR 10.97, 95% CI 1.69-71.04,respectively).Length of stay in days was also increased with PPH (OR 1.23, 95% CI 1.14-1.34)(Table 2).
When outcomes were analysed by PPH cases before compared with after OB-STaT training, PPH cases had higher percentages of type and cross for blood products during admission after OB-STaT (22.1%), compared with 10.5% before OB-STaT (p < 0.001).After OB-STaT, more of the PPH cases were identified as being at increased risk for PPH: 33.2%, compared with 27.4% before OB-STaT (p < 0.001).Significantly more deliveries complicated by PPH after OB-STaT used one or more medications to treat PPH (53.8% before vs 75.9% after, p < 0.001).The use of TXA in PPH deliveries also significantly increased from 34.4% before to 45.8% after OB-STaT training (p = 0.009).There was a trend towards a statistically significant increase in transfusions, from 4.7% before to 9.1% after (p = 0.06), and massive transfusions (≥4 units of blood products) had a non-significant decrease in PPH cases, from 10 (3.6%) before to seven (2.8%) after OB-STaT (p = 0.46).For PPH cases, no significant changes were noted in operative interventions (including uterine curettage, uterine balloon tamponade, uterine compression sutures and uterine artery or hypogastric artery ligation), and no significant changes were seen in maternal morbidity, length of stay or intensive care unit admission rates following OB-STaT (Table 3).
During the study period, approximately half of the sites used EBL for blood loss determination, whereas the remainder used QBL.For both EBL and QBL sites, there was an increase in the number of deliveries meeting the criteria for increased PPH risk following OB-STaT training (Table S1).PPH rates were similar before and after OB-STaT training at both EBL and QBL sites; however, the overall rate of PPH at QBL sites was higher than at EBL sites (6.4% vs 4.1%, p < 0.0001).This is likely to be associated with the improved accuracy of QBL compared with EBL, and not associated with differences in populations at each site or with true differences in PPH rates.The use of uterotonics, tranexamic acid and total number of PPH interventions increased similarly following OB-STaT training at both EBL and QBL sites.Additional comparisons of outcomes between EBL and QBL sites are shown in Table S1.

| Main findings
This study was conducted to evaluate changes in patient outcomes after an IPE PPH simulation training programme (OB-STaT) for the majority of staff involved in maternity care across a large healthcare system.Although overall rates of PPH did not significantly decrease following OB-STaT training, there was a significantly lower composite maternal morbidity rate despite an increased baseline risk of PPH.Following OB-STaT training there were lower rates of massive transfusions, and a slight decreased length of patient stay in PPH deliveries (although not clinically significant).There were also improvements in PPH management following OB-STaT, with an increased frequency of use of uterotonic medications, and an increase in the number of types of uterotonic medications, use of tranexamic acid and hysterectomy.This suggests that intensive training improves the identification of patients at increased risk for PPH, and improves team performance in treating PPH and in treating the underlying causes of PPH, such as uterine atony.More aggressive PPH management and earlier treatment of uterine atony in higher risk patients may have contributed to the decrease in composite maternal morbidity and decreased length of stay.OB-STaT training emphasises prompt recognition of uterine atony and appropriate response with uterotonic medications and interventions, such as uterine balloon tamponade.Although PPH is defined as blood loss of ≥1000 ml, it is important to identify causes of increased bleeding (such uterine atony or significant lacerations) early, and initiate treatment to prevent continued bleeding that can lead to PPH and worse outcomes.Patients with PPH were also more likely to receive more haemorrhage medications and TXA after OB-STaT training, and they had a trend towards increased transfusions and a nonsignificant decrease in massive transfusions.This suggests that more intensive treatment may have resulted in appropriate early transfusions with reduced need for massive transfusions.
Although the overall transfusion rates did not change significantly following OB-STaT training, the finding of decreased massive transfusions of ≥4 units of blood products by 57% (from 21 pre-training to nine post-training) is important.Early and appropriate interventions in PPH may have led to a decrease in the need for massive transfusions in all deliveries following OB-STaT.Whereas hysterectomy procedures significantly increased following OB-STaT training, from zero to five, this remains a relatively rare surgical procedure with a rate of 0.1% (five hysterectomies in 4921 deliveries).Of the hysterectomy procedures, one was a planned caesarean hysterectomy, two were completed during caesarean deliveries (one for refractory atony and one for uncontrolled uterine bleeding after extensive adhesiolysis) and the other two were performed for refractory uterine atony following vaginal delivery.Thus, it is difficult to determine whether the training led to a direct increase in hysterectomies or if this is linked to the small absolute numbers in the study and natural variation in the data.Additionally, the timely provision of hysterectomies in severe and refractory PPH cases is a life-saving measure that is necessary in some

| Strengths and limitations
The strengths of this study include the novel and standardised PPH simulation curriculum implemented in hospitals of varying sizes and locations in a large healthcare system, and the objective assessment of patient outcomes for 6 months following training.Despite using two different blood loss assessments across the sites, the study was adequately powered to look at PPH rates in those groups.The simulation curriculum trained a majority of labour and delivery staff, and the study evaluated almost 10 000 deliveries for clinical outcomes.Another strength of this study is that simulations were conducted in situ with team debriefs, which affords the opportunity to identify and correct systems processes, communication and clinical practices that may have further improved patient outcomes.Outcomes for this study were collected using data extracted from the electronic medical record rather than a review of coded data, permitting a detailed review of documented outcomes.Finally, this study provides a real-world, generalisable example of the implementation of a standardised interprofessional simulation education programme.
The limitations include the inability to determine whether the providers caring for patients during the 6 months following the training completed the OB-STaT training.Another limitation is the differences in delivery blood loss assessment between EBL and QBL, which may have affected the PPH rates.Although QBL was not used uniformly across sites, this was consistent during the course of the study, with no sites switching from EBL to QBL.Additionally, the mix of blood loss assessments provides a generalisable example of the potential impact of PPH training, regardless of blood loss assessment, as there was evidence of improved composite maternal morbidity and PPH management at all sites following OB-STaT training.Another limitation is that more subjects delivering after OB-STaT training were documented to be at increased risk for PPH: 41.6%, compared with 8.7% in subjects that delivered before OB-STaT training.It is unknown whether the magnitude of this increased risk reflects a true increase in haemorrhage risk, or rather the improved documentation of haemorrhage risk factors in the EMR.During the study period, system-wide modifications to the electronic medical record included checkboxes to document PPH risk factors on admission.These modifications along with an increased awareness of the importance of PPH risk assessment may have led to increased documentation of PPH risk factors and increased haemorrhage risk following OB-STaT training.The actual degree of haemorrhage risk (if truly increased) may explain some of the increased use of uterotonics, TXA and balloon tamponade interventions following OB-STaT.Moreover, when haemorrhage cases were analysed compared with non-haemorrhage cases, there were higher rates of documented increased PPH risk, use of uterotonics and TXA following OB-StaT training, indicating that appropriate interventions were undertaken in the setting of PPH.Finally, with the before-and-after study design, we cannot determine the causal influence of the OB-STaT training, or account for other changes in practice that may also have impacted outcomes.

| Interpretation
Although recognising that simulation training for labour and delivery units can be challenging to balance while also managing the clinical workload, the OB-STaT training was associated with improvements in care and outcomes.These findings support implementation at additional sites, and the importance of training as many interprofessional team members as possible.Dillon and colleagues demonstrated the faster administration of uterotonics and transfusions after implementing a simulation curriculum for PPH, although the overall PPH and transfusion rates did not decrease. 11heir study demonstrated improved PPH management, with increases in the number of uterotonic medications used and TXA after training.In the current study, although the PPH rates did not change, the outcomes were similar despite the increased risk for PPH after training.
The findings of this study suggest that a key component to improving PPH outcomes includes realistic in situ simulations and training large numbers of interprofessional teams.Meeting the requirements for simulation drills on labour and delivery units could include training the one team working that day; however, with relatively rare obstetric emergencies, improved outcomes occur when a greater percentage of the team has participated in PPH simulation training.Pattinson et al. demonstrated a 20% reduction in mortality as a result of severe haemorrhage and a 36% reduction in the case fatality rate for haemorrhage associated with caesarean delivery when more than 80% of the staff received training. 17ational organisations currently recommend regular training in obstetric emergencies, such as PPH and hypertensive emergencies, and this study adds to the evidence supporting these recommendations and emphasising the importance of training all team members.

| CONCLUSION
A standardised IPE simulation-based curriculum focused on teamwork and critical management actions and skills improves the composite maternal morbidity rate, decreases the number of massive transfusions and improves management, with a decreased length of stay across a large health system.OB-STaT simulation training contributed to improved PPH management by emphasising the identification of patients at increased risk and the use of medication procedures and surgical interventions to treat PPH.The OB-STaT curriculum provides realistic training and skills sustainment in highrisk, low-frequency events, and prepares interdisciplinary

T A B L E 3
Characteristics of PPH deliveries and PPH delivery outcomes before and after OB-STaT.Pre-OB-STaT (n = 277) Post-OB-STaT (n = 253) p Type and cross,

-OB-STaT (n = 5059) Post-OB-STaT (n = 4921) p
Characteristics of deliveries and delivery outcomes before and after OB-STaT.
T A B L E 1 a Number of PPH risk factors, n (%) a χ 2 or Fisher's exact tests.b Student's t-test.c Medications include: oxytocin, carboprost tromethamine, methylergonovine and misoprostol.d Mann-Whitney U-test.
Multivariate analysis of PPH rate after OB-STaT.
T A B L E 2 aVariables eliminated for a lack of statistical significance on univariate analysis: total blood loss; dilation and curettage; B-lynch; and hypogastric artery ligation.b Variables eliminated for predicting failure perfectly: combined mode of delivery; hysterectomy; admission to intensive care unit for PPH; massive transfusion; and maternal morbidity.
is of appropriate PPH management when other interventions have not been effective.