Evaluating the implementation of maternal safety bundles for obstetric hemorrhage and severe hypertension during pregnancy in Arkansas

Abstract

Background

Maternal mortality remains a critical public health concern in the U.S., with rising rates in recent years. Obstetric hemorrhage and hypertension in pregnancy are leading causes of preventable maternal mortality and morbidity. Maternal safety bundles, developed by the Alliance for Innovation on Maternal Health, provide guidelines to improve maternal health outcomes. This study examined the implementation of obstetric hemorrhage and severe hypertension safety bundles in Arkansas hospitals, identifying factors associated with successful adoption.

Methods

Safety bundle implementation data were collected from 37 hospitals in the third quarter of 2023. Hospital characteristics were obtained from the American Hospital Association database. Implementation index scores were calculated as the percentage of recommended elements adopted, both overall and within specific domains. Descriptive statistics summarized hospital characteristics and implementation status. Implementation differences across domains were assessed using Friedman’s test and Wilcoxon signed-rank test, while between-bundle comparisons were evaluated using the Wilcoxon-Mann-Whitney test.

Results

A total of 23 hospitals (62%) were in the urban region, while 14 hospitals (38%) were located in rural areas. In the obstetrical hemorrhage bundle, Readiness domain had significantly higher implementation than Recognition and Prevention (p = 0.0005) and Reporting and Systems Learning domains (p < 0.0001). In the severe hypertension bundle, Readiness (p < 0.0001), Recognition and Prevention (p = 0.0035), and Response (p < 0.0001) domains all had higher scores than Reporting and Systems Learning domain. The Recognition and Prevention domain had significantly higher implementation in the severe hypertension bundle than in the obstetrical hemorrhage bundle (p = 0.0351). Urban hospitals had significantly higher obstetrical hemorrhage bundle implementation scores than rural hospitals (p = 0.0121). Hospitals with more full-time facility personnel and registered nurses demonstrated better implementation of both obstetrical hemorrhage (facility personnel: p = 0.0454; registered nurses: p = 0.0126) and severe hypertension (facility personnel: p = 0.0180; registered nurses: p = 0.0093) bundles.

Conclusions

Readiness components were the most frequently implemented, while Reporting and Systems Learning elements had lower adoption rates. Urban hospitals had higher obstetric hemorrhage bundle implementation levels than rural hospitals. Hospitals with greater staffing resources, particularly full-time facility personnel and registered nurses, demonstrated better implementation of safety bundles. Findings highlight the need for targeted interventions to enhance maternal safety bundle adoption, especially in rural hospitals and under-implemented domains.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-025-07896-3.

Background

Improving prenatal and postpartum maternal health by preventing pregnancy complications and maternal deaths is a key objective of the Healthy People 2030 [1]. Despite this, deaths associated with childbirth are more common in women in the United States (U.S.) as compared to women living in many other developed countries: a global study reported that in 2015 the maternal mortality rates in the U.S. were 26.4 per 100,000 live births compared to 7.3 in Canada or 16.9 as the average across high-income countries [2]. In 2023, 676 pregnancy-related deaths occurred nationally, resulting in a maternal mortality ratio of 18.7 deaths per 100,000 live births [3–5]. Obstetric hemorrhage is a leading and preventable cause of maternal mortality in the U.S., responsible for 18.1% of maternal deaths nationally [5]. Interestingly, rates of obstetric hemorrhage in the U.S. which had declined from the 19th century till the early 2000 s have started climbing up again [6]. Pregnancy-related hypertension and its related complications are responsible for 6.4% of pregnancy-related deaths [5, 7]. Appropriate treatment received in a timely manner can help manage hypertension-related [8] and hemorrhage-related [9] complications.

To optimize perinatal care and enhance maternal health, various quality improvement initiatives have been formulated and implemented on both national and state levels [10]. The Alliance for Innovation on Maternal Health (AIM), a national and cross-sector commitment, is a key partner in developing patient safety bundles that are aimed towards promoting safe care for every U.S. birth [11]. Maternal safety bundles developed by AIM are designed towards improving patient outcomes through an evidence-based structure consisting of four domains: Readiness, Recognition and Prevention, Response, and Reporting and Systems Learning [12]. In 2022, these safety bundles were updated to incorporate a fifth domain: Respectful, equitable & supportive care [13, 14]. AIM’s maternal safety bundles provide guidance on best practices in preventing severe maternal morbidity and mortality in the United States through actionable steps [11].

Maternal safety bundles are designed to be versatile to ease their adaptation to a variety of facilities and resource levels [11]. Levels of maternal care vary [15] and depending upon their available resources, these facilities may have differential ability to adopt the key elements of patient safety bundles [12]. However, standardization of implementation within an institution is highly recommended. AIM also provides structure, process, and outcome metrics to support patient safety bundle implementation and data-driven quality improvement [11]. Moreover, the Centers for Medicare and Medicaid Services (CMS) have proposed a ‘Birthing Friendly’ designation for hospitals to promote better health outcomes among mothers [16]. The extent of implementation of patient safety bundles is a vital quality measure for the Hospital Inpatient Quality Reporting Program and key driver to obtaining ‘Birthing Friendly’ designation. Besides ensuring maternity care safety and quality, the ‘Birthing Friendly’ designation also is also expected to inform consumers about best choices for facilities to seek care from.

An increasing number of institutions are implementing maternal safety bundles and have reported their experiences, challenges, and outcomes [17–21]. While these studies mainly focus on the evaluation of outcomes after implementation of respective patient safety bundles, Duzyj et al. [18] have also evaluated the successes and failures of the bundle implementation process. In addition, elements of patient safety bundles are diverse and differential adoption of bundle elements may be associated with different outcomes [22]. Given that patient safety bundles provide a framework for facilities to adopt practice changes, and that there could be differential adoption and implementation of bundle elements, understanding the factors that may be associated with their adoption is essential for supporting effective implementation strategies. Knowledge about these factors would help better channel resources towards deficient areas and aid in successful adoption and implementation of patient safety bundles to improve maternal health outcomes.

Arkansas, the setting for this study, had one of the highest maternal mortality rates among U.S. states between 2018 and 2022, with a rate of 51.6 per 100,000 live births [3, 23]. Hemorrhage and pregnancy-related hypertension were among the top five causes of pregnancy related deaths reviewed by the state’s Maternal Mortality Review Committee [24]. The primary objective of this study was to describe the status of the implementation of the AIM maternal safety bundles for obstetrical hemorrhage and severe hypertension in pregnancy among birthing hospitals in Arkansas while examining the variability in implementation across the original four domains within each maternal safety bundle, as well as implementation differences between the two maternal safety bundles. Additionally, with a specific focus on exploring the potential relationship between facility characteristics and bundle implementation, this study aims to uncover potential factors that might play a role in the successful implementation of safety bundles.

Methods

Data sources and data collection

The High-Risk Pregnancy Program was established as a collaboration between the Institute for Digital Health & Innovation at the University of Arkansas for Medical Sciences and the Arkansas Department of Human Services with the support of the Arkansas Medical Society. The Perinatal Outcomes Workgroup through Education and Research (POWER) is a statewide outreach component of the High-Risk Pregnancy Program that provides outreach and education to support perinatal quality improvement at all birthing hospitals in Arkansas. As a part of POWER nurses with obstetric/perinatal experience provide education and quality improvement coaching to birthing hospitals in Arkansas, centered on maternal safety bundles from the Alliance for Innovation in Maternal Health (AIM). Each of the five nurses on the POWER team is assigned to hospitals within one of five public health regions in Arkansas, and meets quarterly with the nursing management of the labor & delivery units. The consistent contact between the POWER nurses and their assigned hospitals facilitates a strong, trusting relationship, and allows the POWER nurse to tailor QI coaching to the context of each hospital. The POWER team began providing education and coaching in 2017, and this coaching has continued each year, with a focus on AIM’s bundles addressing obstetrical hemorrhage and severe hypertension in pregnancy.

In addition to quality improvement outreach and coaching, the POWER team offers free staff to all 37 participating hospitals through in-person sessions and interactive video platforms. The team performs simulation drills, quantitative blood loss measurement, triage, and medication-specific education on obstetric topics. The obstetrics emergency simulation drills, one of the important educational opportunities offered by the POWER team, play a critical role in training labor and delivery staff statewide to prepare them for handling uncommon emergencies that may arise during childbirth and post-delivery periods. These simulation sessions emphasize the application of patient safety bundles, algorithms, and checklists to facilitate prompt recognition and intervention for hypertensive crises, severe postpartum hemorrhages, and instances of shoulder dystocia.

In 2022, the POWER team initiated a survey-based checklist assessment tool to evaluate the extent of the implementation of the AIM bundles for obstetrical hemorrhage [25] (Supplementary Table 1) and severe hypertension in pregnancy [14] (Supplementary Table 2) across Arkansas’ birthing hospitals. Five nurse facilitators perform quarterly data collection on the implementation of bundles from the 37 participating hospitals. During the POWER facilitation visit, the appointed facilitator conducts a survey-based assessment of the implementation of practices from both safety bundles using the checklist assessment tool. The authorized nurse manager or educator at each participating hospital completes the survey tool. These individuals lead AIM bundle implementation at their facility, and self-reported measures from nursing unit leaders have been utilized in other studies of maternal safety bundle implementation [22, 26]. The current study utilizes checklist data collected during the third quarter of 2023.

Study Measures

The survey-based checklist assessment tool includes practice elements from the four original domains of the maternal safety bundle including Readiness, Recognition and Prevention, Response, and Reporting and Systems Learning. The tool contains 18 discrete practice elements for the obstetrical hemorrhage bundle and 29 practice elements for the severe hypertension in pregnancy bundle (Additional File 1). To illustrate the types of practice elements assessed in the surveys, examples from the obstetrical hemorrhage bundle include: “Availability of a hemorrhage cart with supplies, checklist, and instruction cards for intrauterine balloons,” “Establishment of massive and emergency release transfusion protocols,” and “Assessment of hemorrhage risk prenatally.” For the severe hypertension bundle, example items include: “Rapid access to medications used for severe hypertension/eclampsia,” “Facility-wide education standards on signs and symptoms of hypertension and preeclampsia,” and “Facility-wide standard protocols with checklists and escalation policies for the management and treatment of severe hypertension.” Each of these elements is assessed dichotomously. If hospitals report adopting a specific bundle practice element, they receive a score of 1 for implementation; otherwise, they will receive a score of 0, indicating that they have not implemented that particular bundle element. A hospital’s overall bundle implementation index score is created by calculating the percentage of all possible/recommended bundle elements implemented by a hospital. To examine the implementation status in different domains of the safety bundle, domain implementation index scores are calculated as the percentage of adopted practice elements within each domain.

Hospital characteristics, drawn from the 2021 American Hospital Association hospital database, include hospital region (rural, urban), hospital teaching status (yes, no), critical access status (yes, no), hospital ownership (for-profit, nonprofit, state-run), hospital size (small, medium, large), provision of neonatal intensive care beds (yes, no), provision of neonatal intermediate care beds (yes, no), total full-time registered nurses, total full-time facility personnel, total obstetricians/gynecologists, annual number of live births, total bassinets in neonatal intensive and intermediate care units, total licensed beds, and obstetric unit care level ([1] uncomplicated maternity and newborn cases, [2] uncomplicated cases, the majority of complicated problems, and special neonatal services, [3] all serious illnesses and abnormalities, supervised by a full-time maternal/fetal specialist).

Statistical analysis

For this study, we calculated descriptive statistics for participating hospitals’ characteristics and implementation of the safety bundle practices, using checklists completed in the 3rd quarter of 2023. We also visually displayed the variations in the implementation of both individual practice elements and domain practice elements within each maternal safety bundle using Tableau (Version 2023.3). To further assess the implementation differences across the four domains within each maternal safety bundle, we used Friedman’s test to conduct within-bundle comparisons. Using Wilcoxon signed-rank test, we performed post hoc pairwise comparisons between the different domains within each safety bundle. To account for the 6 pairwise comparisons between the four domains, we followed the Bonferroni adjustment technique and used a significance level of (0.05/6 = 0.0083) for each individual pairwise comparison. To examine the implementation differences between the two safety bundles, the Wilcoxon-Mann-Whitney test was conducted. To investigate the association between hospital characteristics and bundle implementation, we utilized several statistical tests, including two-sample t-test and one-way analysis of variance (ANOVA) for categorical characteristics as well as non-parametric Wilcoxon-Mann-Whitney and Kruskal-Wallis H tests. To examine the relationship between continuous characteristics and bundle implementation, we used Spearman’s correlation. We conducted additional analyses focusing on practices adopted by fewer than 76% of hospitals. We compared adoption rates between rural and urban hospitals and between hospitals with ≤ 500 versus > 500 annual live births (a cutoff point that approximated the median in this sample and is consistent with other studies) [27]. We used Chi-Square tests and Fisher’s Exact Tests as appropriate based on expected cell sizes. All analyses were performed in SAS 9.4. Except in the case of the pairwise comparison, we used a two-sided p-value less than 0.05 to determine statistical significance.

Results

Hospital characteristics of the 37 participating hospitals are reported in Table 1. Among the participating hospitals, 23 (62.16%) were in urban areas, 24 (64.86%) were teaching hospitals, and most were not designated as critical care access hospitals (88.24%). Regarding obstetric unit care level, only 7 (21.21%) hospitals provide services for all serious illnesses and abnormalities and are supervised by a full-time maternal/fetal specialist. Six hospitals were for-profit (17.65%), while 25 (73.53%) hospitals were not-for-profit, and 3 (8.82%) hospitals were state-run. The majority of the hospitals did not offer neonatal intensive care beds (78.38%) or neonatal intermediate care beds (81.08%). Regarding hospital size, 18.92% were classified as small, 37.84% as medium, and 43.24% as large. The average number of full-time facility personnel was 884.14 (SD = 963.28), while the average number of full-time registered nurses and obstetricians/gynecologists on staff were 272.79 (SD = 350.70) and 8.78 (SD = 9.12), respectively. Hospitals, on average, had 212.05 (SD = 183.33) licensed beds and 20.50 (SD = 22.62) bassinets in neonatal intensive and intermediate care units. On average, there were 916.79 (SD = 874.42) annual live births at the participating hospitals.

Table 1.

Hospital characteristics (N = 37)

Characteristic		Percent or Mean (SD)		Total n with characteristic reported
Region				37
Rural		37.84		14
Urban		62.16		23
Teaching Hospital				37
No		35.14		13
Yes		64.86		24
Critical Access Hospital				34
No		88.24		30
Yes		11.76		4
Obstetric Unit Care Level				33
Services for uncomplicated maternity and newborn cases		45.45		15
Services for uncomplicated cases, the majority of complicated problems, and special neonatal services		33.33		11
Services for all serious illnesses and abnormalities and is supervised by a full-time maternal/fetal specialist		21.21		7
Ownership				34
For-Profit		17.65		6
Nonprofit		73.53		25
State-run		8.82		3
Neonatal intensive care beds				33
No		78.38		25
Yes		21.62		8
Neonatal intermediate care beds				33
No		81.08		26
Yes		18.92		7
Hospital size				37
Small		18.92		7
Medium		37.84		14
Large		43.24		16
Number of full-time facility personnel		884.14 (963.28)		29
Number of full-time registered nurses		272.79 (350.70)		29
Number of obstetricians/gynecologists		8.78 (9.12)		32

Bundle practices across individual practice elements and across the four bundle domains exhibited some variability in the number of hospitals implementing them (Figs. 1 and 2). The percentage of the hospitals that successfully implemented each element of the Readiness domain for obstetric hemorrhage bundle ranged from 73% to100% (Fig. 1), while 70% (n = 26) of all hospitals fully implemented all elements of the hemorrhage bundle’s Readiness domain (Fig. 2). There were 3 (16.67%) out of 18 obstetrical hemorrhage bundle practice elements and 7 (24.14%) of 29 severe hypertension in pregnancy practice elements were implemented by all 37 hospitals.

Fig. 1.

Percentage of hospitals implemented each practice element within obstetrical hemorrhage and severe hypertension safety bundles

Fig. 2.

Number of hospitals fully implemented the domains of obstetrical hemorrhage and severe hypertension safety bundles

As shown in Table 2, the results indicated variations in implementation across domains within each bundle and between the two bundles. For the obstetrical hemorrhage safety bundle, the mean number of practice elements implemented across all domains was 15.30 (SD = 1.10), with a median of 16 (range: 8–18) and a mean bundle implementation index score of 0.85 (SD = 0.15). There were significant implementation differences among the four domains of the obstetrical hemorrhage safety bundle (p = 0.0079). The pairwise comparisons showed that the implementation score of the Readiness domain was significantly higher than the implementation score of the Recognition and Prevention domain (p = 0.0005, which is < 0.0083). In addition, the implementation score of the Readiness domain was significantly higher than the implementation score of the Reporting and Systems Learning domain (p < 0.0001, which is < 0.0083). For the severe hypertension in pregnancy safety bundle, hospitals implemented an average of 26.08 (SD = 3.31) practice elements, with a median of 27 (range: 17–29) practice elements and a mean bundle implementation index score of 0.90 (SD = 0.11). Significant differences in implementation levels were found across all domains (p < 0.0001). The Wilcoxon signed-rank tests showed that the implementation score of the Readiness domain was significantly higher than the implementation score of the Reporting and Systems Learning domain (p < 0.0001, which is < 0.0083). The implementation score of the Recognition and Prevention domain was significantly higher than the implementation score of the Reporting and Systems Learning domain (p = 0.0035, which is < 0.0083). In addition, the implementation score of the Response domain was significantly higher than the implementation score of the Reporting and Systems Learning domain (p < 0.0001, which is < 0.0083). Furthermore, when implementation differences were compared between the two safety bundles, we observed that while there was no significant difference in the overall bundle implementation index score (p = 0.1566), the implementation index score for the Recognition and Prevention domain was significantly higher in the severe hypertension safety bundle compared to the obstetrical hemorrhage safety bundle (p = 0.0351).

Table 2.

Hospital implementation of obstetrical hemorrhage and severe hypertension bundle practices (N = 37)

Bundle Name	Bundle Domain	Number of Practice Implemented Mean (SD)	Number of Practice Implemented Median (Range)	Index Score Mean (SD)	p-value a	p-value b
Obstetrical Hemorrhage	All Domains	15.30 (1.10)	16 (8–18)	0.85 (0.15)		0.1566
	Domains				0.0079
	Readiness	6.49 (0.96)	7 (3–7)	0.93 (0.14)		0.4965
	Recognition and Prevention	3.16 (0.93)	3 (1–4)	0.79 (0.23)		0.0351
	Response	2.59 (0.55)	3 (1–3)	0.86 (0.18)		0.1725
	Reporting and Systems Learning	3.05 (1.10)	3 (0–4)	0.76 (0.28)		0.9925
Severe Hypertension	All Domains	26.08 (3.31)	27 (17–29)	0.90 (0.11)
	Domains				< 0.0001
	Readiness	12.97 (1.44)	14 (10–14)	0.93 (0.10)
	Recognition and Prevention	5.41 (0.96)	6 (3–6)	0.90 (0.16)
	Response	4.62 (0.72)	5 (2–5)	0.92 (0.14)
	Reporting and Systems Learning	3.08 (1.04)	3 (0–4)	0.77 (0.26)

^ap-Values calculated using Friedman’s test

^bp-Values calculated using Wilcoxon-Mann-Whitney test

Table 3 presents the association between hospital characteristics and bundle implementation index scores. The bundle implementation index scores for obstetrical hemorrhage and severe hypertension in pregnancy safety bundles varied across some hospital characteristics. Specifically, urban hospitals showed significantly higher bundle implementation index scores for the obstetrical hemorrhage safety bundle compared to rural hospitals (p = 0.0121). The mean bundle implementation index score of the severe hypertension in pregnancy safety bundle for urban hospitals was higher compared to rural hospitals; however, the difference was marginally significant (p = 0.0643). The number of full-time facility personnel exhibited a significant positive correlation with the bundle implementation index score for both the obstetrical hemorrhage (correlation coefficient = 0.37; 95% CI = 0.003–0.648; p = 0.0454) and the severe hypertension in pregnancy (correlation coefficient = 0.44; 95% CI = 0.075–0.688; p = 0.0180) safety bundles. Similarly, an increase in the number of full-time registered nurses was associated with higher bundle implementation index scores for both the obstetrical hemorrhage safety bundle (correlation coefficient = 0.46; 95% CI = 0.101–0.702; p = 0.0126), and the severe hypertension safety bundle (correlation coefficient = 0.48; 95% CI = 0.122–0.712; p = 0.0093). Additionally, the number of obstetricians or gynecologists showed a marginally significant positive correlation with the bundle implementation index scores for both the obstetrical hemorrhage safety bundle (correlation coefficient = 0.35; 95% CI = −0.009−0.617; p = 0.0523) and the severe hypertension safety bundle (correlation coefficient = 0.35; 95% CI = −0.006−0.618; p = 0.0508). The annual number of live births also showed a positive correlation with the bundle implementation index score for the obstetrical hemorrhage safety bundle, with a marginal significance (correlation coefficient = 0.32; 95% CI = −0.027−0.590; p = 0.0664). As shown in Supplementary Table 3 (Additional File 2), rural hospitals were significantly less likely than urban hospitals to provide unit education on drills with post-drill debriefs (R1.7, p = 0.0230) within the obstetric hemorrhage bundle, and to establish a culture of huddles for high-risk patients within the severe hypertension bundle (R4.1, p = 0.0006). No statistically significant differences were observed by live birth volume.

Table 3.

Safety bundles implementation index score by hospital characteristics

Hospital Characteristic (N = 37)	Obstetrical Hemorrhage		Severe Hypertension in Pregnancy
	Bundle Implementation Index Score Statistics*	p-value	Bundle Implementation Index Score Statistics*	p-value
Region		0.0121a		0.0643 a
Rural	0.78 (0.14)		0.85 (0.13)
Urban	0.89 (0.14)		0.83 (0.10)
Teaching Hospital		0.3173 a		0.3850 a
No	0.81 (0.17)		0.88 (0.13)
Yes	0.87 (0.14)		0.91 (0.11)
Critical Access Hospital		0.4096 a		0.9421 a
No	0.85 (0.15)		0.90 (0.12)
Yes	0.82 (0.07)		0.92 (0.06)
Obstetric Unit Care Level		0.2955 b		0.7102 d
Services for uncomplicated maternity and newborn cases	0.86 (0.09)		0.90 (0.12)
Services for uncomplicated cases, the majority of complicated problems, and special neonatal services	0.79 (0.21)		0.90 (0.10)
Services for all serious illnesses and abnormalities and is supervised by a full-time maternal/fetal specialist	0.92 (0.10)		0.94 (0.09)
Ownership		0.4374 d		0.3312 b
For-Profit	0.81 (0.20)		0.88 (0.11)
Nonprofit	0.86 (0.13)		0.92 (0.09)
State-run	0.76 (0.21)		0.76 (0.22)
Neonatal intensive care beds		0.1454 a		0.0733 a
No	0.83 (0.15)		0.89 (0.11)
Yes	0.92 (0.10)		0.96 (0.06)
Neonatal intermediate care beds		0.7681 c		0.3994 a
No	0.85 (0.16)		0.90 (0.11)
Yes	0.87 (0.08)		0.93 (0.10)
Hospital size		0.2545 b		0.3609 d
Small	0.80 (0.18)		0.92 (0.08)
Medium	0.82 (0.16)		0.86 (0.13)
Large	0.90 (0.12)		0.92 (0.11)
Number of full-time facility personnel	0.37 (0.003–0.648)	0.0454 e	0.44 (0.075–0.688)	0.0180 e
Number of full-time registered nurses	0.46 (0.101–0.702)	0.0126 e	0.48 (0.122–0.712)	0.0093 e
Number of obstetricians/gynecologists	0.35 (−0.009−0.617)	0.0523 e	0.35 (−0.006−0.618)	0.0508 e
Number of licensed beds	0.21 -(0.129–0.495)	0.2195 e	0.13 -(0.203–0.436)	0.4367 e
Number of bassinets in neonatal intensive and intermediate care units	0.24 (−0.115−0.529)	0.1799 e	0.30 (−0.058−0.569)	0.0963 e
Annual number of live births	0.32 (−0.027−0.590)	0.0664 e	0.29 (−0.062−0.566)	0.1013 e

*Mean (Standard Deviation) reported for categorical characteristics; Spearman correlation coefficient (Confidence Interval) reported for continuous characteristics

^ap-values calculated using Wilcoxon-Mann-Whitney test

^bp-values calculated using Kruskal-Wallis H test

^cp-values calculated using Two-sample t-test.

^dp-values calculated using One-way analysis of variance

^ep-values calculated for Spearman correlation coefficient

Discussion

We studied the implementation status of two maternal safety bundles (obstetric hemorrhage and severe hypertension) across 37 hospitals in Arkansas and also examined hospital-level characteristics for their association with higher level of implementation. We found substantial differences in the implementation of individual elements of the two safety bundles and also across domains within each safety bundle. For example, hospitals were more likely to implement elements of the Readiness domain for both safety bundles, compared to elements of the Reporting and Systems Learning domain. Similarly, hospitals were more likely to implement all elements of the Readiness domain for the obstetrical hemorrhage safety bundle, compared to that of the Recognition and Prevention domain. The number of full-time facility personnel and the number of registered nurses, two hospital-level factors, were positively associated with the implementation of both safety bundles.

The implementation of the two obstetric safety bundles in Arkansas shows some similarities with and differences from the implementation at other sites as reported in published studies. An evaluation of the postpartum hemorrhage safety bundle implemented at Massachusetts General Hospital reported that Readiness components were the most successfully implemented while components of Reporting and Systems Learning required sustained efforts for successful implementation, which is in line with what we found in our study [18]. Similarly, the monitoring and evaluation component of the hypertension bundle was fully implemented by only 59% of the hospitals in New York [28]. The practices recommended in the Reporting and Systems Learning domain include care huddles, case reviews, and monitoring of quality metrics. Obtaining and analyzing quality metrics can be a challenge, requiring modifications to the Electronic Health Record system or substantial staff time to complete record abstraction. Care huddles may be more successfully implemented on units with a strong teamwork culture [29], and perceptions about teamwork vary between labor & delivery units [30].

Additional studies have also reported considerable variation in implementation of individual elements of maternal safety bundles. In California’s initiative to reduce severe maternal morbidity from hemorrhage, roughly half of hospitals implemented a benchmark of 14 (out of 17) practice changes, with 65% reporting regular post-hemorrhage debriefs [19]. In New York, where 63% of hospitals fully implemented the severe hypertension safety bundle [28] most hospitals (81%) had implemented algorithms for 1 st line treatment of emergency, while roughly one-third implemented checklists for management in emergency department [28]. The adoption patterns in New York are lower on average than what we observed in Arkansas, which suggests that the outreach of POWER has been effective at promoting practice change. Across the two bundles included in this study, quantified blood loss was the least frequently implemented practice. This practice has been noted as complex to implement and/or sustain with fidelity [18]. Given the importance of quantification for accurate assessment of blood loss, [31, 32] local tests of change with approaches that simplify and reduce staff burden for quantitative assessment, [33] such as automatic calculations in the electronic medical record, may improve local adoption of this important practice change.

Several studies have identified barriers and facilitators to implementation of the hemorrhage and hypertension safety bundles. In a public hospital in Atlanta, successful implementation of the hemorrhage bundle was considered feasible given proper supports, such as continuous education, regular monitoring and evaluation, are put in place [34]. A pilot study conducted for determining effective approaches for implementing the hypertension safety bundle determined that training and simulation activities are instrumental in successful implementation of the safety bundle [35]. A qualitative study investigating barriers and facilitators for obstetric safety bundle implementation highlighted inadequate staff and lack of institutional support as the key barriers [36]. Like other previous studies, the key facilitators were training and education and involvement of a multi-disciplinary team. In this study, we found that hospitals with higher number of full-time employees including registered nurses had higher implementation scores for both safety bundles, which is consistent with the qualitative study. Managers at hospitals with more labor & delivery staff likely have more ongoing, dedicated administrative time to devote to bundle implementation tasks, particularly the case reviews and quality monitoring recommended in the 4th domain of both the hemorrhage and hypertension bundles. Lack of staff time has been noted as a barrier to healthcare team huddles and case reviews in other studies [37–39].

While this study provides valuable insights into the implementation of maternal safety bundles for obstetric hemorrhage and severe hypertension in Arkansas hospitals, several limitations should be considered. The reliance on self-reported implementation data, collected from the designated nurse manager or educator at each hospital, introduces the possibility of reporting bias and inconsistencies in hospital-level assessments. Additionally, the cross-sectional design captures implementation status at a single point in time, preventing an evaluation of changes in adoption over time or the impact of interventions. The study is also limited by its focus on hospitals in Arkansas, which may reduce the generalizability of findings to other states or healthcare settings with different facility characteristics, resource availability, staffing structures, or regulatory environments. Although the inclusion of all birthing hospitals in the state is a strength, the relatively small number constrains the ability to detect smaller but potentially meaningful differences in bundle adoption across hospital characteristics. Another limitation is the restricted availability of hospital characteristic data, which were sourced solely from the American Hospital Association (AHA) hospital database. While this dataset provides useful insights into structural and operational attributes, it does not account for other influential factors such as institutional culture, teamwork dynamics, leadership support, and staff engagement. These unmeasured variables may play a critical role in shaping the adoption and sustainability of maternal safety practices. Furthermore, the study does not include patient health outcomes following bundle implementation, making it difficult to determine whether increased adoption of safety bundles translates into improved maternal health metrics. Future research should integrate a broader range of hospital-level factors and assess patient-centered outcomes to develop a more comprehensive understanding of the impact of maternal safety bundle implementation.

Conclusion

This study provides insights into the implementation of maternal safety bundles for obstetric hemorrhage and severe hypertension across 37 Arkansas hospitals. Findings indicate that, for both safety bundles, Readiness components were the most frequently implemented, while Reporting and Systems Learning elements had lower adoption rates. Urban hospitals had significantly higher implementation levels for the obstetric hemorrhage bundle than rural hospitals. Hospitals with greater staffing resources, particularly more full-time facility personnel and registered nurses, demonstrated better implementation of safety bundle elements. These findings highlight the need for targeted interventions to enhance maternal safety bundle adoption, especially in rural hospitals and domains with lower implementation rates. Strategies such as additional training and education, institutional support and resource allocation, and policy-driven standardization may help improve adherence to these evidence-based practices.

Abbreviations

AIM

The Alliance for Innovation on Maternal Health

CMS

The Centers for Medicare and Medicaid Services

POWER

The Perinatal Outcomes Workgroup through Education and Research

ANOVA

One-way analysis of variance

Authors’ contributions

Authors’ contributions CP: Writing– original draft, Conceptualization, Data acquisition, Statistical analysis, Data visualization. MA: Writing– review & editing, Writing– original draft, Data visualization. BK: Writing– review & editing, Writing– original draft. JCK: Writing– review & editing; Methodology. DB: Writing– review & editing, Data acquisition. RP: Writing– review & editing. HE: Writing– review & editing, Funding acquisition.

Funding

The Perinatal Outcomes Workgroup through Education and Research (POWER) is part of the outreach efforts of the High-Risk Pregnancy Program (HRPP) at UAMS. The UAMS HRPP and associated research are funded by the Arkansas Medicaid program. The content of this research is solely the responsibility of the authors and does not necessarily represent the official views of the U.S. Department of Health and Human Services Medicaid office.

Data availability

The datasets analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was reviewed by the Institutional Review Board (IRB) at the University of Arkansas for Medical Sciences (UAMS) in Little Rock, Arkansas, and was granted a Human Subject Waiver (UAMS IRB #31075), as the evaluation involved only de-identified data and posed minimal risk to participants.

The requirement for informed consent was waived by the UAMS IRB in accordance with institutional policy and federal regulations governing the use of de-identified data in research. No identifiable private information or biospecimens were collected or used.

This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.