Statewide perinatal quality improvement, teamwork and communication activities in Oklahoma and Texas

Abstract

Background and Objective:

The purpose of this study was to describe statewide perinatal quality improvement (QI) activities, specifically implementation of Alliance for Innovation on Maternal Health (AIM) patient safety bundles and use of teamwork and communication tools, in obstetric units in Oklahoma (OK) and Texas (TX).

Methods:

In January-February 2020, we conducted a survey of AIM-enrolled hospitals in OK (n=35) and TX (n=120) to gather data on obstetric unit organization and QI processes. Data were linked to hospital characteristics information from the 2019 American Hospital Association survey and hospitals’ maternity levels-of-care from state agencies. We generated descriptive statistics for each state and created an index to summarize adoption of QI processes. We fitted linear regression models to examine how this index varied by hospital characteristics and self-reported ratings for patient safety and AIM bundle implementation.

Results:

Most obstetric units had standardized clinical processes for obstetric hemorrhage (94% OK; 97% TX), massive transfusion (94% OK; 97% TX), and severe hypertension in pregnancy (97% OK; 80% TX); regularly conducted simulation drills for obstetric emergencies (89% OK; 92% TX); had multidisciplinary QI committees (61% OK; 83% TX); conducted debriefs after major obstetric complications (45% OK; 86% TX). Few obstetric units offered recent staff training on teamwork and communication to their staff (6% OK; 22% TX); those who did were more likely to employ specific strategies to facilitate communication, escalate concerns, and manage staff conflicts. Overall, adoption of QI processes was significantly higher in hospitals in urban than rural areas, teaching than non-teaching, offering higher levels of maternity care, with more staff per shift, and greater delivery volume (all p<0.05). QI adoption index scores were strongly associated with respondents’ ratings for patient safety and implementation of maternal safety bundles (both p<0.001).

Conclusions:

Adoption of QI processes vary across obstetric units in Oklahoma and Texas, with implications for implementing future perinatal QI initiatives. Notably, findings highlight the need to reinforce support for rural obstetric units, which often face greater barriers to implementing patient safety and quality improvement processes than urban units.

Introduction

Maternal mortality and severe maternal morbidity steadily increased in the United States over the past 30 years.¹ Pregnancy-related deaths more than doubled from 7.2 deaths per 100,000 live births in 1987 to 17.3 deaths in 2017,² and severe maternal morbidity increased from 49.5 to 144.0 cases per 10,000 delivery hospitalizations between 1993 and 2014, respectively.³ Reviews of adverse maternal events identify contributing factors at the community, health facility, provider, patient, and system levels.^4–6 Within the clinical setting, this includes missed diagnosis, inappropriate or delayed treatment, implicit racial/ethnic bias in care, poor communication and coordination between providers, and lack of guiding policies and protocols to ensure quality provision.^4–8 Most of these factors are modifiable, and approximately two-thirds of maternal deaths are preventable.^4–6

In recognition of these gaps, quality improvement (QI) initiatives have been developed and launched at national and state levels to improve the safety and quality of perinatal care. Nearly all states have or are currently developing perinatal quality collaboratives (PQC) to spread implementation of best practices and strategies to optimize perinatal care and outcomes.⁹ The Alliance for Innovation on Maternal Health (AIM)—a national data-driven maternal health initiative funded through a cooperative agreement between the American College of Obstetricians and Gynecologists (ACOG) and the Maternal and Child Health Bureau (MCHB)-Health Resource and Services Administration—is partnering with state and community-based teams to align national, state, and hospital level QI efforts to improve maternal health outcomes.¹⁰ As of June 2022, 45 states and the District of Columbia had enrolled in AIM.¹¹ A key AIM strategy is the implementation of patient safety bundles, a set of evidence-based practices that, when performed collectively and reliably, are known to improve patient outcomes.¹² The AIM patient safety bundles follow a 5R structure—readiness, recognition, response, reporting, and respectful care—to support best practices that make birth safer (Table 1).¹³ These best practices are based on available evidence and the consensus of major maternal healthcare professional organizations.^14–16 Implementation of patient safety bundles by PQCs in several states demonstrate this strategy can improve maternal care processes and outcomes at scale.^17–22

Table 1.

AIM Patient Safety Bundles & AHRQ Teamwork and Communication Tools

AIM patient safety bundles 13
5R Framework	Description
Readiness	Preparations that units, facilities, and communities can take in order to optimize care before a condition occurs. Examples: team-based simulation drills, development of standardized processes and protocols
Recognition & Prevention	Early awareness or avoiding a clinical condition or event. Examples: performance of screenings, standardized evaluation protocols
Response	Practices addressing how a care setting can optimize care when an event occurs for a best possible outcome. Examples: utilizing standardized protocols and escalation policies, transitioning to appropriate levels of care
Reporting & Systems Learning	Evaluations and structures put into place to support optimal care and improve practices. Examples: post-event debriefs, reviews of severe maternal morbidity and mortality cases
Respectful, Equitable, and Supportive Care†	Elements that center respectful, equitable, and supportive care. Examples: open, transparent, and empathetic communication with patients and their support network
AHRQ teamwork and communication tools 29
Tool	Description
SBAR	Technique for communicating critical information about a patient’s condition that requires immediate attention and action (Situation, Background, Assessment, Recommendation and Request)
Call out	Technique of loudly announcing critical information to inform all team members simultaneously during emergent situations
Check back	Closed-loop communication strategy to verify information is understood by the receiver as intended
Handoffs / ‘I Pass the Baton’	Strategy to enhance information exchange during transitions in care (Introduction, Patient, Assessment, Situation, Safety, Background, Actions, Timing, Ownership, Next)
Briefs	Semi-structured conversation among team members before the start of a case to share the care plan, assign roles and responsibilities, establish expectations and climate, anticipate outcomes and likely contingencies
Huddles	Ad hoc meeting among team members to reestablish situational awareness and assess the need to adjust the care plan in response to emergent needs
Debriefs	Semi-structured conversation among team members designed to improve patient care by reviewing team performance after a recent case or clinical event
Power words (a.k.a. ‘CUS’ words)	Assertive statements to communicate patient safety concerns using the CUS acronym (I am Concerned, I am Uncomfortable, This is a Safety issue)
Two challenge rule	Rule to assertively voice a concern at least twice to ensure it has been heard, and then escalate if the safety concern still is not addressed
DESCR script	Approach for managing and resolving conflict (Describe the specific situation/behavior, Express how the situation makes you feel, Suggest alternatives and seek agreement, state Consequences in terms of impact on team goals and seek Consensus)

^†Revisions to AIM patient safety bundles were completed in 2022 and incorporate Respectful Care elements as the 5th R.

Effective teamwork and communication are also critical to patient safety.^23–25 Another QI strategy builds on the Agency for Healthcare Research and Quality’s (AHRQ) Comprehensive Unit-based Safety Program and TeamSTEPPS, an evidence-based teamwork system proven to improve communication, teamwork, and patient safety in healthcare settings.^26,27 AHRQ’s Toolkit for Improving Perinatal Safety applies these evidence-based practices to address adverse events resulting from communication and system failures in obstetric units (Table 1).^28,29 An AHRQ-funded demonstration project is on-going in hospitals in Oklahoma (OK) and Texas (TX) to evaluate the feasibility and added value of overlaying AHRQ teamwork and communication tools with implementation of AIM patient safety bundles for obstetric hemorrhage and severe hypertension in pregnancy. The AHRQ tools are expected to complement QI processes developed by units already implementing AIM patient safety bundles. By reinforcing teamwork and helping units avoid the communication failures at the root of many adverse perinatal events, the theory is that obstetric units will maximize the impact of their ongoing QI efforts to improve perinatal outcomes.

This study was conducted to inform the AHRQ-funded demonstration project and provides an overview of statewide perinatal QI, including teamwork and communication activities, in OK and TX early in 2020, before the COVID-19 pandemic was declared. Specific goals of this study were: 1) to describe the extent that obstetric units have enacted standardized processes to promote perinatal safety and quality; 2) to identify characteristics of units more likely to enact such processes; and 3) to document baseline use of teamwork and communication tools in obstetric units prior to the demonstration project. Findings aim to identify successes and gaps in perinatal QI efforts in the two states, inform future implementation of the AIM program, and assess the need for further integration of teamwork and communication strategies in obstetric units in the two states.

Methods

We use data collected through a survey aiming to assess the implementation of AIM patient safety bundles and the use of teamwork and communication tools in obstetric units in OK and TX hospitals. At the time of the survey, the Oklahoma Perinatal Quality Improvement Collaborative (OPQIC) was supporting implementation of AIM patient safety bundles for obstetric hemorrhage and severe hypertension, and the Texas Department of State Health Services’ (DSHS) TexasAIM Program (TexasAIM) was supporting implementation of the bundle for obstetric hemorrhage.

In January 2020, we obtained a listing of designated AIM Team Leads in AIM-enrolled hospitals and their e-mail addresses from the AIM coordinating organizations: OPQIC and TexasAIM. AIM Team Leads are subject matter experts with experience and responsibilities in implementing QI and patient safety initiatives in obstetric units. These individuals coordinate the implementation of the AIM bundles in their hospitals and, hence, represented the target population for the survey. Between January and February 2020, we contacted all designated AIM Team Leads working in obstetric units at hospitals that were voluntarily enrolled in the AIM initiative in OK (n=44) and TX (n=217), and invited them to participate in our survey. If no response was received to the initial e-mail invitation and up to three follow-up emails, we invited AIM Team Leads attending training workshops held in conjunction with in-person state AIM meetings in February 2020 to complete the survey on paper. We informed potential participants that the survey was voluntary and that their name and the name of their hospital would not be identified in any reports or publications. Informed consent was obtained electronically for those completing surveys online; written informed consent was obtained for those completing surveys on paper. Participants received a $10 gift card as compensation for their time to complete the survey. Institutional Review Board (#IRB00009378) approval for conducting the survey was obtained from the Johns Hopkins University Bloomberg School of Public Health.

The self-administered survey collected information about unit staffing, policies, processes, and participation in QI activities (e.g., AIM bundles, TeamSTEPPS training). Respondents also gave their unit an overall “patient safety grade”, a self-reported outcome measure first developed and tested for AHRQ’s Hospital Survey on Patient Safety Culture.^30,31 Questions were either multiple-choice, rating using a 5-point scale, text entry, or numeric entry. The survey was created for email distribution and online data collection using Qualtrics Core XM survey software (Qualtrics, LLC, Provo, UT).

The American Hospital Association (AHA) 2019 annual survey provided additional information on hospital characteristics, including location (urban, rural); teaching status; hospital system membership; ownership (government, not-for-profit, for-profit); inpatient discharges insured by Medicaid/Medicare; and annual number of births.³² For OK, OPQIC (BO, EO) verified and corrected AHA information on annual delivery volumes and provided information about maternity levels of care offered in each hospital. For TX, information about hospitals’ maternity levels of care was available on the DSHS website for 97% of hospitals in the sample as of December 2021;³³ four hospitals no longer offer obstetric services and, therefore, were classified as “not designated”.

We created an index score to summarize the extent to which obstetric units adopted various QI policies and processes (“QI process index”). The index measured availability and use of standardized processes to improve readiness, recognition, response, and reporting of clinical conditions (e.g., obstetric hemorrhage, severe hypertension in pregnancy) consistent with the AIM bundles (prior to the revision and incorporation of the 5^th ‘R’ in 2022), and it measures adoption of a “stop the line” policy to empower team members to speak up when they sense a patient safety issue. The index is not intended to capture all QI processes in obstetric units, but it uses 13 indicators that reflect the strength of a unit’s QI processes in light of ongoing national efforts in this area. To create this index, we first coded survey questions as 1 (“yes”) or 0 (“no” or “missing”) and then divided the number of positive responses by the total number of questions (n=13) to produce scores ranging from 0 to 1, with 1 as the maximum score indicating the unit had all QI processes measured in this study.

Next, we examined differences in mean QI process scores by hospital characteristics (e.g., location, maternal level of care, ownership), AIM bundle implementation, and patient safety ratings. We hypothesized that robust implementation of AIM patient safety bundles would correlate with stronger QI processes, and stronger QI processes would reinforce a culture of patient safety. To test for significant differences in mean scores, we pooled the data from the two states and fitted resistant linear regression models with bootstrapped standard errors and adjusted for state. Postestimation commands were used to predict marginal QI process score means and corresponding 95% confidence intervals.

Last, we assessed use of teamwork and communication strategies by obstetric units in each state. We expected that recent TeamSTEPPS or similar training would increase use of these strategies, so we used pooled data from the two states to compare units’ use of teamwork and communication strategies, disaggregated by whether TeamSTEPPS or similar training was offered to unit staff in the past year.

All analyses were performed using Stata software (version 15.1; Stata Corporation, College Station, TX).

Results

Sample characteristics

A total of 35 of 44 (80%) birthing hospitals enrolled in the AIM initiative in Oklahoma and 120 of 217 (55%) birthing hospitals enrolled in the AIM initiative in TX as of January 2020 completed the survey, for an overall response rate of 59%. At the time of the survey, most but not all birthing hospitals in the two states were enrolled in the AIM initiative; hospitals responding to the survey represented approximately 73% (35 of 48) of active birthing hospitals in OK and 54% (120 of 224) of active birthing hospitals in TX. The majority of respondents were directors (37%) or clinical/nurse managers (34%) in the obstetric unit; other respondents included physicians, nurse educators, and registered nurses.

Birthing hospitals represented in the survey included government, not-for-profit, and for-profit hospitals providing different levels of maternal care in both urban and rural areas (Table 2). Every year, hospitals in OK had 771 median deliveries (interquartile range (IQR): 320–1,269), and hospitals in TX had 1,305 median deliveries (IQR: 546–2,369). At the time of the survey, nearly all respondents in OK reported their hospital was implementing AIM patient safety bundles to address obstetric hemorrhage (94%) and severe hypertension (97%) (Table 2). In TX, nearly all respondents reported their hospital was currently implementing the bundle for obstetric hemorrhage (97%), and more than half (56%) were also implementing the severe hypertension bundle.

Table 2.

Hospital Characteristics by State

	Oklahoma n=35	Texas n=120
Location1
Urban, %	54%	74%
Rural, %	46%	26%
Maternal level of care designation2
Level I, %	71%	21%
Level II, %	14%	43%
Level III, %	11%	16%
Level IV, %	3%	17%
Not designated, %	n/a	3%
Teaching status3
Teaching hospital, %	6%	8%
Non-teaching hospital, %	94%	92%
Hospital system membership4
Yes, %	74%	73%
No, %	26%	28%
Hospital ownership
Government, state or local, %	20%	23%
Government, federal5, %	6%	0%
Nongovernment, not-for-profit, %	54%	41%
Investor-owned, for-profit, %	20%	36%
Median (IQR) number annual births6	771 (320–1,269)	1,305 (546–2,369)
Mean (SD) percent of inpatient discharges insured by Medicaid or Medicare7	62.2 (18.0)	62.7 (11.5)
Mean (SD) number of full-time, hospital-employed providers per shift in obstetric unit8	6.3 (5.4)	7.6 (7.2)
Mean (SD) number of full-time, hospital-employed providers in obstetric unit8	48.7 (38.9)	57.7 (68.5)
Mean (SD) number of private providers with practicing privileges in obstetric unit8	6.1 (7.0)	13.3 (16.0)
AIM bundle practices implemented, %
Obstetric hemorrhage bundle	94%	97%
Severe hypertension in pregnancy bundle	97%	56%
Units where providers received TeamSTEPPS or similar training in the past year, %	6%	22%

Notes: IQR=interquartile range. n/a=not applicable. Do not know, not applicable, or missing values are excluded from calculation of summary statistics.

¹Urban hospitals are located inside a Metropolitan Statistical Area (MSA) as designated by the U.S. Office of Management and Budget. Rural hospitals are located outside an MSA.

²Maternal level of care designation for Texas facilities are as reported by the Texas Department of State Health Services at: https://dshs.texas.gov/emstraumasystems/maternalfacilities.aspx. Maternal level of care designation for Oklahoma facilities were provided by OPQIC leadership.

³Member of the Council of Teaching Hospitals of the Association of American Medical Colleges as reported in the 2019 American Hospital Association (AHA) annual survey.

⁴Hospital belongs to a corporate body that owns and/or manages health provider facilities or health-related subsidiaries as reported in the 2019 AHA annual survey.

⁵Category includes hospitals under the direction of the Public Health Service Indian Service.

⁶Annual births in 2019. Data sources are the Oklahoma State Department of Health for Oklahoma hospitals, and the 2019 AHA annual survey for Texas hospitals. Data were missing for 21 Texas hospitals; missing values were imputed using median births for Texas hospitals with the same maternal level of care.

⁷Percent of inpatient discharges insured by Medicaid or Medicare is calculated by dividing the total facility inpatient discharges covered by Medicaid or Medicare by the total hospital admissions as reported in the 2019 AHA annual survey.

⁸Providers in the obstetric unit include Obstetrics-Gynecology (ObGyn) Attending Physicians, ObGyn Residents, ObGyn Hospitalists, Registered Nurses, Certified Nurse-Midwives, Midwives, Physician Assistants, Nurse Practitioners, and other clinicians. Numbers as reported for the survey.

Perinatal safety and QI processes and policies

Nearly all units in both states had standardized clinical processes for obstetric hemorrhage and massive transfusion protocols (Table 3). In OK, where implementation of the AIM bundle for severe hypertension was widespread at the time of the survey, 97% of units also had standardized clinical processes for severe hypertension in pregnancy; these were in place in 80% of obstetric units in TX. Fewer units in TX and OK had standardized processes for recognizing and responding to maternal early warning signs (71% and 33%, respectively). Among units with standardized processes, 83% of respondents in OK and 77% in TX were “somewhat confident” or “very confident” that staff in the unit used existing standardized processes in an obstetric emergency. Simulation drills for obstetric emergencies were conducted in 89% of hospitals in OK and 92% in TX.

Table 3.

Perinatal Safety and Quality Improvement Processes and Policies by State

	Oklahoma n=35	Texas n=120
	%	%
READINESS, RECOGNITION & RESPONSE
Unit has standardized processes (e.g., order sets, unit policies, protocols) for:
Obstetric hemorrhage*	94%	97%
Massive transfusion*	94%	97%
Severe hypertension in pregnancy*	97%	80%
Maternal Early Warning Signs*	33%	71%
Among units with standardized processes,1 confidence that staff use existing standardized processes in an obstetric emergency
Not confident	6%	3%
Somewhat not confident	9%	14%
Neither confident nor not confident	3%	7%
Somewhat confident*	40%	51%
Very confident*	43%	25%
Unit conducts multidisciplinary in situ clinical scenario simulation drills for obstetric emergencies*	89%	92%
REPORTING & SYSTEMS LEARNING
Unit has a multidisciplinary quality improvement committee that meets regularly*	61%	83%
Unit has process for regular debriefs with unit staff after major obstetric complications*	45%	86%
Unit has standardized review processes for:
Review of Severe Maternal Morbidity cases*	27%	77%
Review of Maternal Deaths*	48%	87%
Source of data for tracking unit-based outcomes
Chart review	29%	22%
Computer-generated reports*	3%	3%
Both chart review and computer-generated reports*	69%	76%
TEAMWORK & COMMUNICATION
Unit has a “stop the line” policy*	55%	59%
Among units with “stop the line” policy,2 confidence that clinical staff at all levels are empowered to “stop the line”
Not confident	0%	0%
Somewhat not confident	6%	10%
Neither confident nor not confident	11%	8%
Somewhat confident*	78%	57%
Very confident*	6%	25%
Overall quality improvement process score,3 mean (SD)	0.63 (0.17)	0.74 (0.22)

Notes: Do not know, not applicable, or missing values are excluded from calculation of summary statistics.

¹Denominator restricted to those units with standardized process in place (n=152 total).

²Denominator restricted to those units with “stop the line” policy (n=83 total).

³Responses marked with an asterisk (*) are used to generate a quality improvement process index, with a score ranging from 0 to 1. The number of affirmative responses is summed and divided by the total number of items included in the index (n=13). Responses of either “somewhat confident” or “very confident” are considered affirmative. Only responses of either “computer-generated reports” or “both chart review and computer-generated reports” are included in the index given implied reporting simplicity and timeliness.

In terms of reporting and systems learning, regular debriefs with unit staff after major obstetric complications occurred in 45% of hospitals in OK and 86% in TX, and multidisciplinary QI committees that meet regularly were in place in 61% of hospitals in OK and 83% in TX. Units in TX and OK had structured processes for reviewing severe maternal morbidity cases (77% and 27%, respectively) and maternal deaths (87% and 48%, respectively). Over 70% of hospitals in both states used computer-generated reports to keep track of maternal outcomes.

More than half of units in both states had a “stop the line” policy that empowers all clinical team members, regardless of position, to speak up immediately if they see a risk to patient safety (55% in OK and 59% in TX). Over 80% of respondents from hospitals with such policy were “somewhat confident” or “very confident” that clinical staff at all levels were empowered to “stop the line”.

Overall, on a 0–1 scale, obstetric units in OK and TX had average QI process scores of 0.63 (SD: 0.17) and 0.74 (SD: 0.22), respectively. QI process scores were significantly higher in hospitals in urban than rural areas, teaching than non-teaching, those offering higher levels of maternity care, having larger numbers of hospital-employed obstetric providers per shift and larger numbers of annual births, as well as hospitals implementing both obstetric hemorrhage and severe hypertension AIM bundle than only one of the bundles (all p<0.05; Table 4).

Table 4.

Differences in Mean Quality Improvement Process Scores by Characteristics of Oklahoma and Texas Hospitals (n=155)

	QI process score Mean (95% CI) †	p-value
Location
Urban	0.78 (0.75, 0.82)	<0.01
Rural	0.67 (0.60, 0.74)
Maternal level of care designation1
Level I	0.68 (0.61, 0.75)	0.01
Level II	0.76 (0.72, 0.81)
Level III	0.78 (0.70, 0.86)
Level IV	0.84 (0.78, 0.89)
Teaching status
Teaching hospital	0.84 (0.76, 0.93)	0.03
Non-teaching hospital	0.74 (0.71, 0.77)
Hospital system membership
Yes	0.76 (0.72, 0.79)	0.41
No	0.72 (0.65, 0.80)
Hospital ownership
Government	0.68 (0.58, 0.77)	0.13
Nongovernment, not-for-profit	0.77 (0.73, 0.82)
Investor-owned, for-profit	0.75 (0.71, 0.80)
Average number of full-time, hospital-employed providers per shift in obstetric unit
<5 providers	0.70 (0.64, 0.77)	0.02
5–8 providers	0.76 (0.71, 0.81)
>8 providers	0.81 (0.77, 0.86)
Number annual births
<500 births	0.69 (0.62, 0.76)	<0.01
500–999 births	0.68 (0.61, 0.75)
1000–1999 births	0.82 (0.77, 0.87)
2000–2999 births	0.78 (0.71, 0.84)
>=3000 births	0.81 (0.75, 0.87)
Percent inpatient discharges insured by Medicaid or Medicare
Average change per percentage point increase in Medicaid/Medicare discharges	−0.002 (−0.005, 0.001)	0.16
Units where providers received TeamSTEPPS or similar training in the past year
Yes	0.76 (0.66, 0.85)	0.70
No	0.74 (0.70, 0.78)
AIM patient safety bundles implemented
Only one AIM bundle (i.e., obstetric hemorrhage OR severe hypertension)	0.69 (0.63, 0.75)	0.01
Both AIM bundles (i.e., obstetric hemorrhage AND severe hypertension)	0.78 (0.75, 0.81)
Self-reported grade on implementation of AIM patient safety bundles
Excellent	0.83 (0.74, 0.92)	<0.001
Very good	0.79 (0.75, 0.83)
Acceptable	0.70 (0.65, 0.74)
Poor or Failing	0.53 (0.38, 0.68)
Self-reported grade on patient safety
Excellent	0.77 (0.69, 0.84)	<0.001
Very good	0.79 (0.75, 0.82)
Acceptable	0.65 (0.58, 0.72)
Poor or Failing	0.46 (0.31, 0.61)

Notes:

^†Estimates derived using resistant linear regression to limit influence of extreme values with bootstrapped standard errors to account for skew and unequal variance. Sampling distributions of the test statistics are approximated by a normal distribution for a one-hypothesis test and a chi-squared distribution for a multiple-hypotheses test (Wald chi-squared test). All models are adjusted for state.

¹Results for hospitals classified as “not designated” are not shown due to small sample size (n<5).

Respondents in both states graded their hospital’s implementation of AIM bundles and patient safety highly (Figure 1). Specifically, 57% of respondents in OK and 73% in TX rated their hospitals’ work on AIM bundle(s) as “very good” or “excellent”. Related ratings of “very good” or “excellent” on patient safety were provided by 74% of respondents in OK and 81% in TX. We found strong associations between QI process scores and respondents’ ratings for patient safety and AIM implementation (both p<0.001).

Figure 1. Self-reported Grade on Patient Safety and Implementation of AIM Patient Safety Bundles by State.

In Texas, five hospitals were missing responses on the patient safety grade and six hospitals were missing responses on the AIM implementation grade.

AHRQ teamwork and communication tools

In the year before the survey, TeamSTEPPS or similar training had been offered to obstetric unit staff in 6% and 22% of hospitals reporting on this question in OK and TX, respectively (Table 2). Nearly all units employed the SBAR (Situation, Background, Assessment, Recommendation) technique to communicate critical information about a patient’s condition (91% in OK and 94% in TX) (Figure 2). Most units used huddles, i.e. ad hoc meetings to monitor and modify patient care plans as needed (80% in OK and 84% in TX), and debriefs, i.e. informal meetings to review team performance (71% in OK and 88% in TX). These techniques were common among units that both had and had not received TeamSTEPPS or similar training in the past year. Briefs, i.e. short sessions held at the start of a case to share the patient care plan, were less commonly used (41% in OK and 62% in TX). Additionally, while most units use Power Words to signal concerns about patient safety (64% in OK and 71% in TX), few units used the Two-Challenge Rule as a technique to escalate concerns (9% in OK and 15% in TX). Handoffs, checkbacks, callouts, the Two-Challenge Rule, and the DESCR (Describe, Express, Suggest, Consequence, Reach) script to constructively deliver feedback and manage conflict were more rarely employed, especially so in units where providers had not received TeamSTEPPS or similar training in the past year (all p≤0.02).

Figure 2. Use of TeamSTEPPS Tools and Strategies by State and Past Year TeamSTEPPS Training Status.

SBAR is a framework used to effectively communicate information about a patient’s condition to members of a healthcare team; the mnemonic stands for S=Situation; B=Background; A=Assessment; and R=Recommendation. DESCR script is a constructive approach for delivering feedback and managing conflict; the mnemonic stands for D=Describe the specific situation; E=Express your concerns; S=Suggest alternatives; C=state Consequences; and R=Reach consensus.

One Texas hospital was missing response on use of TeamSTEPPS tools.

Twenty-five hospitals (3 in Oklahoma and 22 in Texas) were missing responses on past year TeamSTEPPS training.

Discussion

Early in 2020, before the World Health Organization declared the Covid-19 pandemic, the majority of birthing hospitals enrolled in statewide QI initiatives in OK and TX were implementing one or more AIM maternal safety bundles—these are designed to facilitate standardized practices to improve readiness, recognition, response, and reporting/systems learning on maternal safety issues, as well as respectful care (i.e., the 5R framework). Overall, we found widespread adoption of standardized clinical processes associated with the AIM bundles. However, only half of units had a “stop the line” policy that empowers all clinical team members, regardless of position, to speak up immediately if they see a risk to patient safety. The concept of “stopping the line” is not explicitly mentioned as a best practice in AIM bundles, perhaps providing an explanation for its lower adoption relative to other QI policies and processes. “Stopping the line” is integral to AHRQ teamwork and communication strategies, but obstetric providers are less familiar with these strategies. Providing teamwork and communication training (e.g., TeamSTEPPS) and tools (e.g., Power Words, Two Challenge Rule) to obstetric units may add value to current QI efforts by systematizing speaking up about safety concerns.

There were significant differences between urban and rural hospitals in QI processes measured with our summative index. Rural hospitals, most of which are Level I maternity hospitals, with <1,000 births annually and fewer full-time providers per shift in the unit were significantly less likely than urban hospitals to have standardized clinical procedures, review processes, patient safety policies, and to conduct simulation drills for obstetric emergencies. Similarly, differences in index scores between OK and TX may, in part, arise from differences between the two states in rurality; nearly half of OK hospitals are rural and the majority have a Level I maternity care designation. Across the United States, half of rural counties lack hospital obstetric services,^34,35 and rural residents have higher rates of severe maternal morbidity and mortality than urban residents.³⁶ Indeed, obstetric units face unique challenges to offering high-quality perinatal services in rural communities, including workforce shortages and low birth volumes that make it difficult to maintain clinician competencies and ensure the unit’s financial viability.^34,35,37 Our results complement these findings, suggesting that, likely due to these challenges, rural obstetric units also face greater barriers to implementing patient safety and quality processes than urban units.

Several states have now demonstrated that QI initiatives implemented through state PQCs can improve care and reporting processes and maternal outcomes.^17–22 Not surprisingly, we found that units in OK and TX implementing both AIM bundles had higher QI process scores. Also, importantly, the documented associations between QI index scores and hospital ratings by AIM Team Leads suggest that they have a thorough understanding of how QI processes and adoption of evidence-based practices can indeed improve patient safety and, in turn, maternal outcomes. The strong collaboration with the AIM program by participating hospitals in both states is set to continue through future implementation of additional AIM patient safety bundles. OPQIC is implementing Opioid Use Disorder and Postpartum Support bundles and will launch a standardized communication model that is patient-centered, while TexasAIM focuses on the Severe Hypertension in Pregnancy and Opioid Use Disorder bundles and supporting provision of Obstetric Care during COVID-19. Providing safe, high quality and equitable perinatal care can be challenging for hospitals, especially those with limited resources and during or after a pandemic. Providing hospitals with recommended care guidelines, data systems, structures, and processes—such as those included in the AIM bundles—eases the burden for individual hospitals. Statewide QI initiatives can assist hospitals in bundle implementation and facilitate collaboration amongst hospitals working to improve outcomes at the hospital and state level. Moreover, uptake and standardization of recommended care practices also supports hospitals’ preparation for meeting accreditation and credentialing standards. Building on findings from this study, OPQIC and TexasAIM should develop initiatives and use resources to strengthen the organization and work of multidisciplinary QI committees, ensure that debriefs are conducted after major obstetric complications, and support the offering of training on teamwork and communication to staff in all obstetric units; such efforts are especially needed in rural hospitals and those offering lower levels of maternity care in both states.

The survey had limitations, including a moderate response rate, yet higher than that of similar surveys assessing QI efforts at the state level^18,20 or obstetric providers’ practices.³⁸ Although not representative of obstetric units in all states participating in the AIM program, our data are representative of obstetric units in OK and TX; we compared hospitals represented in our survey versus those not represented, and we found no statistically significant differences in the distribution of key hospital characteristics (e.g., location, maternal level of care, hospital ownership). In line with similar assessments, the information was provided by only one respondent in each hospital; yet respondents were clinical QI leaders usually participating in state QI trainings and coordination events and thus, were expected to be knowledgeable of hospital QI activities. Information about TeamSTEPPS or similar training was only assessed as having been offered or not in the year before the survey, without quantifying the actual proportion of staff trained in obstetric units. The QI process index used in our analyses includes items related to key processes needed for providing high quality obstetric care, but it is not a validated, comprehensive measure of QI implementation. Such measures are needed and items in our index can be considered for inclusion in a more comprehensive perinatal QI adoption measure given their performance in the current study. Finally, this study is intended to be descriptive in nature; it does not seek to make causal attributions between study variables. Instead, it is intended to inform planning and implementation of statewide perinatal QI initiatives by identifying characteristics of hospitals (e.g., rural, maternal level of care) that may benefit from targeted technical assistance and support.

Conclusions

In conclusion, findings from this study provide important contextual and baseline information for the on-going demonstration project in OK and TX, and for planning future QI initiatives after the COVID-19 pandemic. Many clinicians and hospital staff face continued challenges to providing care, which limit some hospitals’ ability to participate in QI initiatives now and potentially in the future. Notably, the study highlights the need for additional support to rural obstetric units, which face greater barriers to implementing patient safety and quality processes than urban units. Our survey findings can be used to inform future QI initiatives to improve maternity care processes and outcomes in the post-pandemic period. Maternal health remains a critical area where QI initiatives are greatly needed.

Supplementary Material

Supplemental Table 1

Supplemental Table 1. Characteristics of Birthing Hospitals Responding to Survey vs. Non-Respondents