Accountability for care through undoing racism and equity for moms: a study protocol for a cluster randomized trial of data accountability and community-based doula interventions in prenatal practices

Abstract

Background & aims

In the United States, Non-Hispanic Black (Black) people are 2–threefold more likely than Non-Hispanic White (White) people to have pregnancy complications, such as a baby with low birthweight. Multi-level, community-based interventions aimed at mitigating the impacts of interpersonal and structural racism may decrease bias, improve the quality of care, and improve pregnancy outcomes. The Accountability for Cancer Care through Undoing Racism and Equity and the Heart Health Now studies reduced disparities in cancer treatment and cardiovascular risk, respectively. The Accountability for Care through Undoing Racism and Equity for Moms Study (ACURE4Moms) was modified from these successful interventions and designed to test the impact of multi-level, community-engaged antiracism interventions delivered on disparities in low birth weight (LBW) (Aim 1), experiences with perinatal depression and experiences with discrimination in medical settings (Aim 2) and other patient-centered outcomes. We will also measure implementation barriers and facilitators (Aim 3).

Methods

ACURE4Moms is a 4-arm, cluster-randomized trial which has enrolled 39 prenatal practices randomized to implement the following interventions for 2 years: Arm 1–Standard Care; Arm 2—Data Accountability and Transparency; Arm 3—Community-Based Doula linkages; and Arm 4—Data and Doula interventions combined. Practice staff in Arms 2–4 will also receive quarterly Maternal Healthcare Equity Education and Training. A subgroup of 100 Black patients from each practice will participate in a longitudinal survey that measures mental health symptoms and experiences with discrimination during pregnancy and postpartum. A stakeholder advisory board including doulas, community members, and policymakers helps to make decisions regarding study design, implementation, and dissemination.

Multi-level mixed models will be used to evaluate outcomes using administrative, vital records and survey data. The primary outcome is a reduction in low birth weight for Black infants. Secondary outcomes include reductions in hospitalizations and emergency department use, mental health symptoms, and experiences with discrimination during pregnancy and postpartum. Intermediate outcomes include implementation barriers and facilitators.

Discussion

The findings of the ACURE4Moms study will inform policy makers, health systems, clinicians and communities about the effectiveness of multi-level, practice-based interventions to reduce maternal health disparities and provide information regarding scalability.

Trial Registration

This trial has been prospectively registered as (7/29/2022) with clinicaltrials.gov under the name Accountability for Care Through Undoing Racism & Equity for Moms (ACURE4Moms), Identifier: NCT05484804. URL: https://clinicaltrials.gov/ct2/show/NCT05484804.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13063-026-09514-9.

Introduction

Background and rationale {9a}

Non-Hispanic Black (Black) birthing people in the United States (US) have maternal morbidity and mortality ratios that are 2–threefold higher than that of Non-Hispanic White (White) birthing people [1, 2]. Multi-level, community-based interventions aimed at mitigating structural racism and improving maternal and infant disparities are urgently needed. However, maternal morbidity and mortality are relatively rare (1% and 0.04%, respectively), making it difficult to design a clinical trial with adequate statistical power to improve these outcomes. Infant low birth weight (LBW) affects more patients and is an important marker of maternal health [3, 4]. LBW deliveries, a combination of preterm births and/or fetal growth restriction, are associated with twice the long-term risk for maternal cardiovascular disease (CVD). LBW and preterm birth combined are the second most common cause of infant death in the US, with significant morbidity for surviving infants [5]. Black infants are also twice as likely than White infants to have a low birth weight (< 2,500 g, 11.4% versus 5.2% in 2016) [6]. In North Carolina (NC), LBW rates are higher than the national average (15.3% for Black infants versus 7.5% for White infants in 2020) [7].

Experiences with racism have been linked to LBW [8, 9]. One longitudinal study of 420 US women of color found that for every 1-point increase in the Everyday Discrimination Scale during the second trimester, birthweight decreased by approximately 50 g [8]. In the CARDIA study, those reporting high levels of discrimination around the time of a singleton birth had almost 5 times the odds of a LBW compared to those reporting no racial discrimination [9].

Although race is a social construct with no biological basis, experiences with racism can affect biological outcomes [10, 11]. Alhusen, et al. proposed an explanatory model (Fig. 1) [12], which distinguishes between two forms of racism: interpersonal and institutional, both of which can lead to differential treatment and access to resources and opportunities. This leads to interrelated negative health consequences including increased inflammation, uteroplacental dysfunction, and maladaptive health behaviors, which contribute to the risk for adverse pregnancy outcomes. By starting to dismantle the many sources of racism facing Black patients every day, it should be possible to reduce the biological impacts of institutional and interpersonal racism.

Fig. 1.

Conceptual Framework of the Potential Pathways Linking Racial Discrimination to Adverse Birth Outcomes. Used with Permission from the Journal of Midwifery & Women’s Health Published by Alhusen et al., 2017 [12]

Systematically using data accountability transparency interventions could improve racial inequities in outcomes [13, 14]. Inpatient race equity dashboards have reduced disparities in maternal morbidity when used during the birth hospitalization [15, 16]. Outpatient race equity dashboards and early warning systems have reduced disparities in cancer treatment [17] and cardiovascular risk [18] but have not been modified for use during prenatal care. Institutional racism and implicit biases from prenatal providers could be reduced through the transparency of race-stratified data dashboards of outcomes from their own practices, increasing awareness of inequitable race-specific outcomes, while concomitantly undergoing racial equity training. Early warning systems could reduce differential treatment by alerting practices about all patients who could benefit from preventive services. Patients with warnings could be linked to additional community resources as needed [19]. In the Accountability for Cancer Care Through Undoing Racism and Equity study, disparities dashboards, race equity training, an early warning system and care navigators at the level of the oncology practice eliminated the disparity between Black and White patients who completed therapy for treatable breast and lung cancers and improved treatment completion for patients of all races [17]. In the Heart Health Now Study, disparities dashboards, early warning systems, and quality improvement practice coaches reduced disparities in cardiovascular risk in a cluster randomized trial of over 200 primary care practices in NC [18]. See Tables 1 and 2.

Table 1.

Data Accountability Interventions from ACCURE and Heart Health Now Studies [17, 18]

Interventions	Causal Effect
Automated real-time registry with an Early Warning System for missed appointments and milestones	Decreased institutional racism by preventing differential treatment
Nurse Navigator specially trained in exploring and responding to patients’ social and belief-specific barriers and in using ACCURE’s real-time registry	Increased personal resources/support and decrease interpersonal racism to prevent differential treatment
Site-specific Clinical Feedback reports, according to race and co-morbidity status, delivered by ACCURE Physician Champion to clinicians	Decreased institutional racism by showing differential treatment by race and encouraging the providers to address it
Quality Improvement Coaches	Decreased structural bias by implementing system-based strategies to reduce bias in care
Racial Equity Training + quarterly booster sessions for providers	Reduced interpersonal racism + implicit bias among providers

Table 2.

Predicted Demographic Characteristics of the Population

Characteristic	The people we would expect to see included
Age	1% 15–18
	23% 18–24
	29% 25–29
	30% 30–34
	17% 35–44
	1% over 44
Sex	All females
Gender	Women
Race, ethnicity and ancestry	1% Non-Hispanic American Indian
	4% Non-Hispanic Asian/Pacific Islander
	22% Non-Hispanic Black
	19% Hispanic
	3% Multiracial
	51% Non-Hispanic White
Socioeconomic status	Education: 52% with Medicaid Insurance
	Education: 34% with Bachelor’s or higher; 28% with some college; 25% with high school diploma/GED; 12% with less than high school educatton
Geographic location	North Carolina, United States 32% rural and 68% urban per NC Urban/Rural Classifications
Other characteristics relevant to the trial

The health effects of experiencing institutional and interpersonal racism may be further mitigated via interventions that increase personal resources (Fig. 1: social support, coping strategies, and self-esteem), such as support from community-based doulas. Doulas are trained professionals who provide emotional, physical, and informational support during pregnancy and postpartum. Community-based doulas are a subset of doulas who are often racially or ethnically concordant, share lived experiences with their clients, are trusted members of their communities, and provide linkages to community resources to mitigate racism [20]. When relationships are created between community doulas and prenatal practices, they can work together to address: 1) institutional racism by providing patient-centered advocacy, 2) interpersonal racism by developing cross-cultural relationships with the health team, and 3) personal resources through offering social support and coping strategies to patients who are at high risk of delivering a LBW baby [21]. Community-based doula support has been associated with reduced LBW in several observational studies but has not been tested in a randomized controlled trial [22, 23].

Therefore, we have developed the Accountability for Care through Undoing Racism and Equity for Moms (ACURE4Moms) study, a cluster-randomized trial aimed at comparing the effectiveness of two practiced-based interventions (Data Accountability and Transparency, and Community-Based Doula Support) both separately and together, to reduce disparities in LBW (Aim 1). Secondary outcomes include emergency care and hospitalizations during pregnancy and postpartum. A subset of Black patients in each cluster will be surveyed longitudinally to measure experiences with discrimination and other patient-centered outcomes (Aim 2). We will also measure implementation barriers and facilitators (Aim 3).

Explanation for the choice of comparator {9b}

The comparator for both interventions is standard prenatal care which currently includes LBW risk factor assessment for NC Medicaid patients and intensive care management for those with the highest risk of LBW. This combination has previously been shown to reduce LBW, especially for Black women and is currently implemented across NC [24].

Objectives {10}

Our primary aim is to compare the impact of data accountability and doula support, both separately and in combination, on LBW (< 2,500 g) for Black patients. Secondarily, we will recruit a nested cohort study of up to 4,400 Black patients (100/practice) to complete a survey at 4 time points measuring mental health symptoms, experiences with discrimination, and other patient-centered outcomes at the following time points: 1) Between 14–23 weeks gestational age, 2) 24 weeks gestational age until birth, 3) 2–11 weeks postpartum, and 4) 12–18 weeks post-birth. See SPIRIT Figure (Fig. 2). Our secondary aim is to longitudinally assess self-reported discrimination and mood symptoms among Black patients during pregnancy and for up to 30 months postpartum and examine the mediating effect between the intervention and infant LBW using the Discrimination in Medical Settings scale and the Edinburgh Postpartum Depression Survey [28, 29]. Our third aim is to assess and identify intervention implementation factors that support or inhibit intervention uptake and sustainability.

Fig. 2.

ACURE4Moms Longitudinal Patient Survey schedule of enrolments and assessments (SPIRIT figure)

Methods: Patient and public involvement, and trial design

Patient and public involvement {11}

From the inception of our project, we determined it essential to involve representatives from among the populations most impacted by maternal health disparities. The academic Principal Investigators began collaborating with two community-based doula leaders who function as Community Principal Investigators in ACURE4Moms. These individuals identify as Black and have lived experience with adverse pregnancy outcomes themselves or within their immediate family. They are primarily responsible for designing the community-based doula intervention, as well as hiring, training, and managing the study doulas in their region. One was assigned to the Western NC practices and the other to the Eastern NC practices. We also work closely with members of the Greensboro Health Disparities Collaborative, an NC-based organization devoted to improving health disparities. This organization, together with academic researchers, designed the original Accountability for Cancer Care through Undoing Racism and Equity study [17], on which our data accountability and transparency interventions are based. Our research team has since built relationships with patient advocacy groups, community-based organizations, provider professional organizations, public health leadership, health insurance payers, patients, and doulas. These stakeholders meet quarterly with the research team as part of our Stakeholder Advisory Board Leading with Equity (STABLE, Fig. 3). The Board is co-chaired by a community-based doula and an academic health disparities researcher who are both Black, and it includes 22 members, the majority of whom are Black women with lived experience of adverse pregnancy outcomes. The Board has the authority to approve all aspects of study design and implementation. Each stakeholder is respected as an expert and their time is equally compensated, regardless of educational or occupational background.

Fig. 3.

ACURE4Moms Study Team and Stakeholder Advisory Board Structure

Trial design {12}

ACURE4Moms is a cluster randomized superiority trial.

Methods: Participants, interventions and outcomes

Trial setting {13}

The study is occurring at 39 prenatal practices in NC. NC is a diverse state on many.

levels: 78 out of 100 counties are rural, and 24%, 16%, and 1% of births are to Black, Hispanic, American Indian/Alaska Native people, respectively. NC Medicaid covers about half of all births in NC; of these, 20% are only covered for birth under “Emergency Medicaid” due to ineligibility for full Medicaid. As mentioned previously, Black North Carolinians experience large disparities in maternal and infant outcomes including morbidity, mortality, LBW, preterm birth, as well as disparities in other social determinants of health, such as sustained access to transportation, employment, and quality education [25, 26]. NC launched the Care Management for High Risk Pregnancies Program in 2011, which uses a screening tool to identify individuals at risk for LBW. A predictive model identifies those most likely to benefit from care management and offers intensive care management to them [24]. All of the approximately 40% of pregnant patients covered by NC Medicaid are screened for eligibility for this program.

Characteristics of the people who are needed for the trial

Eligibility criteria for participants {14a}

All births that occur to women 15 years and older at the included prenatal practices will be included the primary and secondary analyses for Aim 1. Based on historical birth volumes, we estimate that the included practices will perform an average of 30–60 deliveries per month, which will result in ~ 30,000–60,000 patients (8,400–16,800 Black) who initiate prenatal care during the 2 years of the study. All patients who receive care at each site during the study period will be included in the study. A waiver of informed consent was obtained given the fact that the interventions are provided at the level of the practice, are within the standard of care for pregnancy, require practices to be able to have reliable data about population to improve quality of care, and pose minimal risk to participants. Minors < 15 years were excluded from this waiver.

Recruitment for the Longitudinal Patient Survey (Aim 2) will occur across each of the practices with the goal of recruiting 100 Black mothers from each practice (N = 4,400).

For the implementation research, we will enroll up to 132 practice staff, 50 doulas, and 90 patients for in-depth interviews pre-, mid-, and post-implementation to help determine fit, feasibility, and fidelity of the interventions.

Eligibility criteria for sites and those delivering interventions {14b}

Our team divided NC into Eastern and Western halves with approximately half of all sites located in each region (Fig. 4). This division allows the two community-based doula organizations to supervise a similar number of practices and doulas. The study team identified potentially eligible practices based on the racial composition and numbers of births per county. Practices were recruited by emails, personal phone calls, and in-person visits.

Fig. 4.

Location of ACURE4Moms Prenatal Clinics by county across the state of NC. Green denotes areas served by Momma’s Village Fayetteville in Eastern NC and Red denotes areas served by Maame, Inc in Western NC

Inclusion criteria for prenatal practices are: 1) Care for at least 7–8 Black birthing people who deliver each month; 2) Full participation with the NC Health Information Exchange Authority; 3) Willingness to be randomized; 4) No current community-based doula partnership or outpatient race equity dashboards. New quality improvement measures are not prohibited. In most cases, the sites also provide labor and delivery care. In a few cases, the labor and delivery care is provided by a different group of clinicians. However, the interventions are designed to decrease LBW, so the focus is delivering interventions to health care professionals caring for people in the outpatient setting prior to labor. All interventions will be delivered to patients by the prenatal clinic staff who will receive support from our research team including clinical experts, community based doulas, practice facilitators, and the health information exchange authority.

Who will take informed consent? {32a}

For the Longitudinal Patient Survey (Aim 2), Community-Based Recruitment Consultants who are also community doulas have been contracted to recruit study participants from a list of potentially eligible patients provided by each practice via telephone and/or email. Interested patients then undergo the informed consent process by phone with the consultant and receive a copy of the consent by email. Community-Based Recruitment Consulants are not assigned to recruit patients from practices where they provide pregnancy and birth support. For the implementation research, research staff from the UNC Collaborative for Maternal and Infant Health recruit eligible participants and contact them by phone or secure video conferencing.

Additional consent provisions for collection and use of participant data and biological specimens {32b}

No Biological specimens will be collected as part of this study.

Intervention and comparator

Intervention and comparator description {15a}

Arm 1 is standard prenatal care including routine social determinants of health (SDoH) assessment for Medicaid patients and intensive care management for those with the highest risk of LBW. Arm 2 will include a suite of data accountability and transparency interventions modified from the ACCURE and HHN studies including disparities dashboards, early warning alert systems, quality improvement practice coaches, and quarterly maternal health equity education and training (MHEET) for staff. Arm 3 will include linkages with community-based doulas, referral workflows, funded doula support for up to 144 births (6 per month), and quarterly MHEET training for staff. Arm 4 will include both data accountability and community doula interventions as well as the MHEET training for staff. For complete detail about each of the arms, see Table 3.

Table 3.

ACURE4Moms Study Comparators, Interventions, and Clinical Strategies

Comparator Arm	Clinical Strategies Included in Arm
Arm 1 (Standard Care)	> 95% of NC OB practices use the Pregnancy Medical Home (PMH) care model for patients with Medicaid (~ 50% of deliveries)[28] • First prenatal visit risk screening for low birth weight in all Medicaid patients • Intensive care management to address SDOH in high risk
Arm 2 (Data Accountability and Transparency [Data])	1. Standard Care: PMH if Medicaid insured 2. Practice Facilitator (PF): These QI officers from NC Area Health Education Centers (AHEC) work with, 1) Practice teams to help implement our real-time Maternal Warning System (MWS, see below) using rapid QI cycles, and, 2) Practice Provider Champions to present Disparities Dashboard data (see description below) [29] 3. Maternal Warning System (MWS): Health Information Exchange Authority (HIEA) builds the MWS for each practice with HIEA data from their own and other health care facilities throughout NC. Data from nightly uploads from the HIEA are aggregated, analyzed, and shared with practices through a unique portal for each practice. Alerts include: elevated BP (≥ 140 systolic and/or ≥ 90 diastolic), critical BP (≥ 160 systolic and ≥ 110 diastolic), failure to follow-up within 2 weeks of elevated BP, delay with expected prenatal visit schedule (≥ 6 weeks until 32 weeks; ≥ 3 weeks 32–36 weeks; ≥ 2 weeks 36 weeks until birth), patients who qualify for aspirin to prevent preeclampsia who do not have it listed in the medication list. The Nurse and/or Admin Champions review the MWS alerts regularly and follow the pre-specified workflow for each patient’s alert [17, 29] 4. Disparities Dashboard: Disparities Dashboard reports will be generated from the practice EHR data, stratified by race/ethnicity. Dashboard outcomes, such as the proportion with low birth weight, recommended prenatal visits, controlled BP, and APOs, are defined using ICD and CPT codes and other fields that require direct entry into the EHR. The Provider Champion will present outcomes to practice staff and leadership [17, 18] 5. Maternal Health Equity Education and Training (MHEET): Nine quarterly, 1-h facilitated sessions modified from the ACCURE study, based on the People’s Institute for Survival and Beyond Undoing Racism™ principles[30] as a conceptual model for medical care for all practice staff (clinicians, nursing, front desk staff, etc.) Session topics include: Maternal Health Inequities Facts, Structural and Systemic Causes of Disparities, Identifying Structural and Systemic Racism in Patient Stories, Implicit Bias, and Communication. Training is grounded in sharing of race-stratified outcomes data from their own practice and interpreting them through a racial-equity lens. Provider Champions, Nurse Navigators, and an Administrative Lead from each practice are supported and encouraged to complete the 2-day Racial Equity Institute Phase 1 Training
Arm 3 (Community-Based Doula Support [Doula])	1. Standard Care: PMH if Medicaid insured 2. Maternal Health Equity Education and Training (MHEET): (as above) 3. Practice partnership with CBDs to support high-risk patients and improve accountability: Designed by Co-Investigators Angela Malloy and Cindy McMillan, NHB women who are CBD leaders at Momma’s Village Fayetteville and MAAME, Inc., respectively. Forty CBDs with training in birth support, breastfeeding, perinatal mood disorders, Afrocentric Care, postpartum care, and childbirth education have been recruited to provide services for the study, with a focus on serving patients at high-risk for LBW • Practices refer individuals to the CBD organizations before 28 weeks of gestation for any of these risk factors: 1) Self-identified Black/African American race; 2) medical comorbidity (e.g., Chronic Hypertension, Kidney Disease, Heart Disease); 3) prior history of LBW infant (< 5 lb 8 oz/2,500 g), preterm birth (< 37 weeks), infant death, and/or stillbirth (2 points); 4) limited social support (e.g., limited partner/family involvement) (2 points); 5) housing instability (2 points); 6) Substance Use Disorder (including tobacco) (2 points); 7) teen Pregnancy (< 19 years old at birth) (1 point); and/or 8) other clinician specified risk (e.g., Intimate Partner Violence) (1 point) • Each patient matched with a CBD receives the following services: 3–4 CBD visits during the antenatal period (including 1 with their OB clinician to help establish collaboration); up to 24 h of labor support; 1–2 h of immediate postpartum support; a home visit 7–10 days after delivery; ongoing telephone and/or online support throughout pregnancy and postpartum; and participation in CBD-organized peer support groups, which continue until 1 year post-birth • CBDs use online HIPPA-compliant EHR systems (IntakeQ and Maternity Neighborhood) to document notes from each patient interaction. Practice staff undergo a co-learning curriculum designed by the CBDs to educate them about scope of practice and communication between health care professionals and CBDs. CBD services are funded by several Medicaid payers and ACURE4Moms ($1000 per client)
Arm 4 (ALL)	Standard Care + MHEET + Data + Doula as above

Abbreviations: NC (North Carolina); OB (Obstetric); LBW (low birth weight); ACCURE (Accountability for Cancer Care through Undoing Racism and Equity); HHN (Heart Health Now); PF (Practice Facilitator); EHR (Electronic Health Record); HEIA (NC Health Information Exchange Authority); MWS (Maternal Early Warning System); EC (Emergency Care); PISAB (The People’s Institute for Survival and Beyond); CBD (Community Based Doula)

Both interventions were designed to be deployed at the practice level and improve outcomes for everyone in the population. The data accountability interventions and the MHEET training sessions were based on the Accountability for Cancer Care through Undoing Racism and Equity Study (ACCURE), which successfully reduced Black-White lung and breast cancer treatment disparities [17]. Regarding the doula support intervention (arms 3 and 4), this could not be feasibly offered to every person at risk for a LBW in each practice. However, we aimed 1) to support a clinically significant percentage of births to Black-identifying individuals at each practice understanding that not all Black-identifying patients would desire doula support and some who are not Black-identifying would receive doula support and 2) to create relationships and transparency between doulas and clinical providers that would ultimately lead to improvements in care for all patients, not only those who received doula support. The following calculations were used to plan the amount of doula support. Given that we targeted practices with an average of 60 births per month, and that approximately 25% of births in NC are to Non-Hispanic African American individuals, we estimated that this would be about 15 Black patients per month at each clinic. Doulas being able to support about 6 births per month (144 over the course of the 2 year study) would result in approximately 40% of births to Black individuals to be supported which we felt would be a significant dose.

Criteria for discontinuing or modifying allocated intervention/comparator {15b}

ACURE4Moms does not have a Data Safety Monitoring Board given that the study interventions present no more than minimal risk; however, the study investigators, Community Advisory Board (STABLE), and an Independent Safety Monitor will review study data quarterly throughout project implementation to ensure no unexpected risks or outcomes arise. Patients referred to doulas can withdraw consent for this support at any time.

Strategies to improve adherence to intervention/comparator {15c}

We will identify a provider, nurse, and administrative champion at each site to help improve uptake and adherence. Each practice is assigned a practice facilitator who is an experienced in helping practices to improve quality of care. to help with adherence to the data accountability interventions. One medical member of the research team and one member of the community doula team will travel in person to each of the sites to deliver Maternal Health Equity Education and Training (MHEET) sessions to the champion team and other staff. Ad hoc meetings will be used as needed to answer questions and provide implementation support. Community doulas will be paid by the study so that the expense will not be borne by clinics, patients, or the doulas themselves.

Concomitant care permitted or prohibited during the trial {15d}

This is a pragmatic clinical trial, so any concomitant care is permitted during the trial except that, as stated previously, practices with a current community-based doula partnership or outpatient race equity dashboards at baseline will be excluded from the study.

Ancillary and post-trial care {34}

No ancillary or post-trial care is planned given that the interventions should pose minimal harm to participants.

Outcomes {16}

Table 4 includes details about the outcomes and data sources. Data on health outcomes (i.e. LBW, emergency care use (emergency department and labor and delivery), hospitalizations, severe maternal morbidity, mode of delivery, hypertensive disorders) will be accessed from the NC Health Information Exchange Authority and NC Vital Records (birth and death certificates). The Health Information Exchange is a state agency responsible for promoting the access to, exchange of, and analysis of health information among providers [27]. All health care providers who are in network with NC Medicaid plans and/or the State Health Plan are required to share electronic health data. Some data will also be extracted from NC vital records (birth and death certificates). See Table 4.

Table 4.

ACURE4Moms Study Outcomes, Measures, and Timing of Assessment

Type of Outcome	Name of Outcome	Specific measure to be used	Timepoints	Powered? Yes or No
Aim 1: Maternal and Infant Health Outcomes
Primary	Low Birthweight	Infant birthweight < 2500 g from EHR and birth certificate data	Birth	Yes
Secondary	High utilization of EC during pregnancy	Proportion of pregnant women with ≥ 4 EC visits during pregnancy from EHR data	From time of 1st OB visit until birth	No
Secondary	# Maternal hospitalizations & EC Visits during 1st year post-birth	# Maternal hospitalizations & EC visits within 1 year post-birth from EHR data	From time of 1st OB visit to 1 year post-birth	No
Secondary	# Infant hospitalizations & EC Visits during 1st year post-birth	# Infant hospitalizations & EC visits within 1 year post-birth from EHR data	From time of birth to 1 year post-birth	No
ESecondary	Postpartum visit attendance	Postpartum visit documented in EHR data	14–60 days postpartum	No
Secondary	Severe Maternal Morbidity	ICD codes for 21 CDC indicators from EHR data[32] with and without transfusion	Birth Hospitalization	No
Secondary	Hypertensive disorders of pregnancy	ICD codes for eclampsia, preeclampsia, gestational hypertension from EHR data (See Appendix A)	Pregnancy until 6 weeks postpartum	No
Secondary	Nulliparous Singleton Term Vertex Cesarean Section	Birth Certificate Data	Birth	No
Aim 2: Longitudinal Patient Survey Patient-Centered Outcomes
Primary	Racism during OB care	Discrimination in Medical Settings Scale from LPS[28]	24–30 weeks of gestation	Yes
Secondary	Depression	Edinburgh Postnatal Depression Scale in from LPS[29]	24–30 weeks of gestation	Yes
Secondary	Satisfaction with Care	ICHOM Standard Set question from LPS[30]	24–30 weeks of gestation	No
Aim 3: Implementation Outcomes
Primary	Identification of intervention barriers and facilitators	Completion of key informant interviews and focus groups including Hexagon Tool and structured interview guides	Before/midpoint/endpoint of implementation period	No
Secondary	Racism knowledge	Knowledge score from Racial Equity Training pre-/post-tests	Before/after training	No

Abbreviations: EHR (Electronic Health Record); EC (Emergency Care); # (number); ICD (International Classification of Disease Codes Clinical Modification 10); OB (Obstetric); CPT (Current Procedure Technology); LPS (ACURE4Moms Longitudinal Patient Survey); ICHOM (International Consortium for Health Outcomes Measurement)

Patient-Centered Outcomes

Longitudinal patient surveys include survey responses from 4 time points. Surveys data will include the Birth Satisfaction Scale, the Patient-Reported Outcomes Measurement Information System-10 quality of life scale, the Edinburgh Postnatal Depression Scale, the Discrimination in Medical Settings Scale, and items from the International Consortium for Health Outcomes Measurement standard set for pregnancy and childbirth [28–30]. See supplementary information for more details on these items.

Maternal Health Equity and Education Training Outcomes

We will also evaluate 3 forms of knowledge about racism, bias, and disparities in healthcare among practice staff in Arms 2–4 through post-tests after each training session: 1) Factual knowledge (e.g. statistics on racial disparities in obstetric care), 2) Conceptual knowledge (e.g. institutional and interpersonal racism), and 3) Process knowledge from hearing the personal and professional experiences of facilitators, colleagues and the case presented during the trainings.

Implementation Data

Individual telephonic or video interviews will be conducted with practice staff including the provider, nurse, and administrative champions at 3 time points: prior to implementation, midway through implementation, and at the conclusion of the study. We use structured interview guides, and the interviews are recorded, transcribed, and analyzed using NVivo qualitative software.

Harms {17}

All the outcomes in Aims 1 and 2 could be harms or benefits depending on the direction of any effects reported (e.g. LBW, preterm birth, psychosocial effects). Sample size calculations incorporated two-sided alpha testing for this reason. We also plan to review all infant and maternal deaths and severe maternal morbidity that occur in our study population.

Participant timeline {18}

The study will take place over 5 years, with the first year (April 1, 2022-May 31, 2023) devoted to pre-implementation activities, including recruiting practices, developing protocols, building the Data Accountability Systems, and recruiting community-based doulas. Implementation at the first sites in June 2023, with sites starting the intervention in a rolling fashion between June 2023 and December 2024. Implementation will continue for 2 years (through December 2026 at the last site). See Fig. 5 for details about timeline of enrollment and implementation at each site.

Fig. 5.

ACURE4Moms Practice Enrolment and Completion Dates by Arm

Sample size {19}

To perform the sample size calculation, the design effect arising from the clustering inherent to the study’s nested design is estimated using the following formula: Deff = 1 + (m-1)ρ, where m = average cluster size and ρ = Intercluster Correlation Coefficient. The minimum number of study observations is approximately 8,400 Black mother-infant dyads in 40 clusters, assuming that up to 4 practices (clusters) will drop out after randomization. Using the power function from Stata [31] and assuming an Intercluster Corrrelation Coefficient of 0.015, a two-sided ɑ = 0.05 and 80% power, with 40 total clusters (10 per Arm) and within cluster sample size of 280, we will be able to detect a 38.6% difference in the LBW rate between arms. The coefficient of 0.015 is a conservative estimate chosen based on a literature review of similar studies [32–34]. The detectable effect size improves with smaller coefficients and higher cluster numbers (see Supplementary Information).

Recruitment {20}

To recruit prenatal practices, we are using the personal connections of the research team members including with leaders of large health systems, presentations about the project at state-wide and regional maternal health meetings, and advertisement in electronic newsletters for prenatal care practices in NC. Once practices express interest, the principal investigators conduct in-person site visits to discuss participation requirements and goals. For Aim 1, all patients 15 and older who receive prenatal care at the clinic over a two-year period will be included with a waiver of informed consent. We are enrolling prenatal practices that cared for at least 7–8 patients who identify as Black so that we have at least 8,500 Black women in the study. For the Longitudinal Patient Surveys, community doulas serve as community-based recruitment consultants (as above) to enrol patients.

Assignment of interventions: randomisation

Sequence generation: who will generate the sequence {21a}

Our statistician performed the initial randomization using the online tool (www.random.org) to generate all random numbers; it provides a “true” random number generator that derives its randomness from atmospheric noise.

Sequence generation: type of randomisation {21b}

To ensure similar numbers of randomized participants in each arm and similar numbers of doula assigned practices in each of 2 geographic areas, the practices were stratified by Eastern NC and Western NC and then by size prior to randomization. For each geographic stratum (East v. West), we ordered the clinics by number of deliveries per month from highest to lowest, then created groups of 4 clinics in that order and randomized to 1 of 4 arms.

Allocation concealment mechanism {22}

The first round of randomization for 42 clinics was performed on December 2, 2022 with 42 sites. Due to dropout prior to implementation, additional recruitment was needed. Therefore, 2 additional rounds of randomization were performed for an additional 7 sites. In these cases, the arms with the fewest sites were identified, and the practices were randomized into those arms. At each point, practices were notified of their allocation after the randomization was complete but before implementation.

Implementation {23}

There is no randomization sequence. See previous item.

Assignment of interventions: blinding

Who will be blinded {24a}

Given the design of offering differing interventions to practices, participating practices, their patients, and research staff could not be masked to the allocation of practices into each study arm.

How will be blinding be achieved {24b}

Not applicable.

Procedure for unblinding if needed {24c}

Not applicable.

Data collection and management

Plans for assessment and collection of outcomes {25a}

For Aim 1, data will be collected for the primary outcome (low birth weight) from vital records birth certificate data which has been well validated. Secondary outcomes will come from electronic health record information sent to the health information exchange. For Aim 2, the primary outcomes will be the validated Discrimination in Medical Settings (DMS) scale [28] and the Edinburgh Perinatal Depression Scale (EPDS) [29]. See Table 4 for additional details.

Plans to promote participant retention and complete follow-up {25b}

For Aim 1, the primary outcome will be assessed for all patients who begin prenatal care at one of the intervention clinics and deliver at 20 weeks gestation or greater, regardless of whether they continue to receive care at their original site. For Aim 2, the primary outcomes, the Discrimination in Medical Settings and Edinburgh Postpartum Depression Scale scores, will be assessed at the second survey, which is completed during the third trimester. Community-Based Recruitment Consultants will contact enrolled participants during their eligibility windows for all surveys to encourage them to complete the survey. An additional incentive will also be provided for the completion of survey 2.

Data management {26}

Prenatal practices will send new prenatal care patient panels, including race and ethnicity, to secure servers at the Cecil G. Sheps Center for Health Services Research. The Sheps Center is a “pan-university” research center within the University of NC which offers Data Management, Statistical Analyses, Secure Research Computing, and Integrated Research Solutions resources. The data sent to the Sheps Center will then be: 1) used by recruitment consultants to recruit participants for our Longitudinal Patient Survey, and 2) sent to the NC Health Information Exchange Authority, who will match patients with statewide health system encounter data and thereby populate the Race Equity Dashboards and the Maternal Warning System. De-identified s data from the Health Information Exchange will then be sent back to the Sheps Center team for analysis by the research team. Surveys will be sent electronically to enrolled participants from the secure Sheps research server, and their responses will be received and stored on the same server. A detailed data management plan has been created and is shared by all members of the study team.

Confidentiality {33}

All data will be kept on the Secure Server at UNC Sheps and is only accessible on the server via encrypted connection. Only those individuals who need to perform data management and analysis will have access to the data and will complete institutional privacy and human subjects protection training prior to receiving access. Data identifiers will be kept separately from analytic data sets, which will be de-identified. Only the Sheps teams will have access to the data identifiers.

Statistical methods

Statistical methods for primary and secondary outcomes {27a}

The primary Aim 1 analysis will compare the odds of LBW among Black birthing persons in Arm 1 (Standard Care Management) compared with Arm 4 (Combined Intervention) using Mixed Linear Logistic Regression Models. The Model will account for nesting of patients (Level 1), within practices (Level 2), adjusting for age, parity, education level, smoking, Body Mass Index category, census tract, and insurance status (variables known to be associated with LBW). Extensive informatics and practice-based quality improvement measures will be taken to limit missingness [18]. Mixed Linear Models will also be used to describe the effects of the intervention on the secondary outcomes.

Who will be included in each analysis {27b}

Analyses will be intention to treat with all patients who start prenatal care at each practice included in the arm the practice was originally randomized to and a planned sensitivity analysis only including patients who stayed at each practice through birth for Aim 1. Patients who transfer care will still have outcomes available in the NC Health Information Exchange and vital records. For Aim 2, a similar intention to treat analysis will be used but only the patients who complete surveys will contribute data to this analysis. Race-identification will be used in our study as a proxy for experiencing anti-Black racism. Race/ethnicity of included patients will be determined by the demographic data provided by practices. Practices will be encouraged and supported by our Quality Improvement experts (the North Carolina Area Health Education Centers Practice Support Program-www.ncahec.net/core_service/practice-support/) to use the best practice of asking patients to self-identify any race(s) and/or ethnicit(ies) with which they identify. However, this recommendation may not always be followed stringently by practices. Therefore, the subset of participants recruited to participate in the Longitudinal Patient Survey will be given an opportunity to self-identify their race(s) and/or ethnicit(ies) with trained recruitment consultants, and any discrepancies with electronic health data will be explored.

How missing data will be handled in the analysis {27c}

Data missingness will be described and associations between practice-level covariates, and missingness will be explored, with differences formally tested using logistic regression models for binary missingness variables. Inverse weighting based on missingness probabilities will be employed to adjust for the effects of practice level covariates on missingness, as needed.

Methods for additional analyses (e.g. subgroup analyses) {27d}

Similar methods will be used for planned subgroup analyses.

Interim analyses {28b}

No interim analyses will be performed.

Protocol and statistical analysis plan {5}

The statistical analysis plan will be submitted to the journal separately from this protocol as an update before data lock.

Oversight and monitoring

Composition of the coordinating centre and trial steering committee {3d}

The Stakeholder Advisory Board is co-chaired by a community-based doula and an academic health disparities researcher who are both Black, and it includes 22 members, the majority of whom are Black women with lived experience of adverse pregnancy outcomes. The Board has the authority to approve all aspects of study design and implementation and meets quarterly.

The UNC principal investigators and project management teams are responsible for the conduct of the trial and meet weekly. They also meet weekly with the study’s’ practice facilitators and the data accountability and transparency intervention team to ensure that the trial is being implemented correctly. The community doula team meets monthly with the research team, and the community doula leads also meet monthly with the doulas.

Composition of the data monitoring committee, its role and reporting structure {28a}

ACURE4Moms does not have a Data Safety Monitoring Board given that the study interventions present no more than minimal risk; however, the study investigators, Community Advisory Board (STABLE), and an Independent Safety Monitor at UNC will review study data quarterly throughout project implementation to ensure no unexpected risks or outcomes arise.

Frequency and plans for auditing trial conduct {29}

The study team will meet weekly to discuss implementation of the study interventions at each of the study practices to ensure that the trial is being conducted in accordance with the study protocol. In addition, periodic checks of recruitment phone calls, communication logs, and other recruitment and enrollment activities will be performed to ensure that the CBRCs are following the approved protocol.

Protocol amendments {31}

Protocol amendments are reported promptly to the UNC coordinating IRB and subsequently to other collaborating institutional IRBs.

Dissemination policy {8}

We are developing a dissemination plan, together with our stakeholders and community partners that will include peer reviewed manuscripts, scientific presentations at conferences, as well as community forums, meetings, and social media. Our funder will also disseminate the findings on their website, and we will report the findings on clinicaltrials.gov. A publication policy was developed at study start which includes pre-planned manuscripts for each study aim as well as a process to propose additional manuscripts by members of the research team or others.

Discussion

The ACURE4Moms study funded by the Patient Centered Outcomes Research Institute, the Duke Endowment, and the UNC Health Foundation will inform clinical practice and guide policy change in NC and beyond. Implementation began in June of 2023 and will continue through December of 2026. It will be the first randomized trial to test the impact of community-based doula support and prenatal clinic data accountability and transparency interventions on disparities in birth outcomes. It will generate a wealth of implementation data for health systems and policymakers, including costs of such interventions.

Though cluster randomized trials are subject to more bias and lower power than individual randomization [35], we chose this study design because both interventions are designed to be delivered at the level of the practice. For the community-based doula intervention, though it would be possible to randomize patients to receive doula support, the ACURE4Moms practices paired with doulas will develop bi-directional, educational relationships with the ACURE4Moms doulas from their communities. Practice staff can learn from the community-based doulas about improving trust with their patients and about community resources and concerns. The doulas can learn the rationale for certain medical recommendations and help explain this to their clients and communities. This integration of the community-based doulas into the care team will offer the opportunity not only to improve outcomes for specific clients, but also to improve the quality of care for all patients at that practice.

A major limitation of the ACURE4Moms interventions is that they started only at the beginning of pregnancy. The factors impacting LBW disparities for Black birthing people, particularly institutional and interpersonal racism, are present well before pregnancy occurs and cannot be completely mitigated during the relatively short duration of a pregnancy and without changing other systems with which these individuals interact. In addition, though we believe doulas may be part of the solution for disparities, we must avoid the possibility that health care professionals and systems will shift the burden of improving health disparities completely to the shoulders of community-based doulas. However, given the preliminary data that both community-based doula support and practice-based data accountability and transparency can improve health disparities, and our plan to combine these in Arm 4, we anticipate our approach to be impactful.

There are several potential barriers to successful implementation of the ACURE4Moms study. Perhaps the most important is the increasing time demands and workflow shortages for clinical staff in NC and nationwide that were exacerbated by the COVID-19 pandemic [36]. Despite these challenges, our team believes that the current disparities in maternal and infant health outcomes, also exacerbated by the pandemic [37], are unacceptable and preventable. The data accountability and transparency interventions will be the most time consuming for clinical staff but are modeled on two successful studies that reduced adult cancer disparities [17] and cardiovascular risk disparities [18]. One of these, the Heart Health Now (HHN) study, found that NC Area Health Education Centers (AHEC) Practice Facilitators (PFs) can effectively help practices establish risk-stratified quality improvement (QI) approaches that resolve racial disparities in CVD risk. HHN was a stepped-wedge, stratified cluster randomized controlled trial (RCT) of 219 primary care practices who participated in NC Medicaid’s medical home program. Where possible, the use of automated data feeds will reduce or eliminate the burden on clinical staff to report data. We will conduct implementation research with the practices to establish workflows that are flexible and workable in real-world clinics.

A second barrier is that the Race Equity Dashboards and Maternal Warning Systems will rely on the NC Health Information Exchange Authority to extract analyzable data from the electronic health records of included clinical sites which then is integrated into the race equity dashboards developed by the Sheps Center. Scaling to other states would require similar partnerships with local health information exchanges. However, the goal of United States Department of Health and Human Services is to achieve nationwide interoperability of health information systems by 2024 [38]. Building ACURE4Moms interventions within a state-wide exchange, rather than within an individual practice electronic health record, takes advantage of standardizations in interoperability and will allow other prenatal clinics to utilize the Dashboards and Warning Systems in the future. It will also allow us to potentially capture BP measurements and other risk factors from encounters for ACURE4Moms patients in NC health facilities not enrolled in the study.

A third barrier is engaging community-based doulas with practices given the history of institutional racism and many doulas’ desires to remain outside of these systems. Our research team is co-led by doula leads from two community-based organizations. Their leadership has been essential to designing an intervention that is acceptable to community members and community-based doulas. They are taking the lead in orienting practices and hospitals to the community-based doula scope of practice and in establishing relationships between the doulas and other members of the health care team.

Conclusion

The ACURE4Moms study is a cluster randomized trial of data accountability and transparency and community-based doula interventions implemented at diverse prenatal practices across the state of NC. The study of the implementation of these two innovative interventions, applied to maternal health disparities, offers the opportunity to inform maternal health policymakers, providers, and health systems about how to best use available resources to minimize racial disparities in maternal health.

Trial status

thirty-nine prenatal practices have been enrolled in the ACURE4Moms study including private practices, health-system owned clinics, academic medical centers with resident clinics, and community health centers. We have achieved approval from the Institutional Review Board for updated protocols, currently version 2.4. The study began recruitment at the first sites June 5, 2023 and the last site in December 15, 2024 and recruitment will be complete December 15, 2026. We have trained 35 doulas who are active in the program. 26 community-based recruitment consultants are actively recruiting participants to the Longitudinal Patient Survey. Recruitment began August 1, 2023 and should be complete January 31, 2027. We have built 3 maternal early warning alerts and (missed prenatal visits, elevated blood pressures, and eligible for aspirin to prevent preeclampsia). A sample disparities dashboard with 3 metrics (low birth weight, timely prenatal care, and preterm birth) has been built and is being tested with clinics. The full disparities dashboards are being built. The MHEET training has been developed for the first 7 sessions. All sites have received MHEET session 1 training, and some have received through session 7. Pre-implementation interviews have been conducted at every practice, and the midpoint interviews are in process. Nine practices dropped out prior to implementation due to either closure, changes in staffing, or inability to contract with the Health Information Exchange Authority. Two practices have dropped out after partial implementation because they were unable to continue providing prenatal care. See Fig. 5.

Abbreviations

ACURE4Moms

Accountability for Care through Undoing Racism and Equity For Moms

CVD

Cardiovascular disease

LBW

Low Birth Weight

NC

North Carolina

STABLE

Stakeholder Advisory Board Leading with Equity

UNC

University of North Carolina

Authors’ contributions {3a}

RPU is a principal investigator in the study who co-led conception of the study and led the proposal development along with JT. All authors read and approved the final manuscript. MB led the development of the statistical analysis plan and assisted with preparation of the manuscript. AM, CM, MJ, SV, SC, MB, WN, KM, CY, AM, DB, CY, BC, AA, KB, and RH all contributed to the design of the study and assisted with preparation of the manuscript.

Data availability {6}

De-identified longitudinal patient survey data and implementation data will be made available upon request to interested researchers. Due to contractual obligations with the North Carolina Health Exchange Authority, we are unable to share health outcomes data outside of the approved research team at UNC. All data will be kept on the Secure Server at UNC Sheps and only accessible on the server. Only those individuals who need to perform data management and analysis will have access to the data.

Declarations

Ethics approval and consent to participate {30}

ACURE4Moms has been approved by the University of North Carolina Institutional Review Board (IRB) as of August 30, 2022 (Review Number 22–1541). Written informed consent will be obtained from all participants in the longitudinal patient survey and for the participants in the implementation interviews. A waiver of informed consent was obtained for obtaining health information exchange authority and vital records data on pregnancy outcomes for participants given that the interventions are delivered at the level of the practice, focused on quality improvement, and pose minimal risk to participants. However, data security contracts were set up between each of the participating practices and UNC to and between UNC and the Health Information Exchange Authority to clarify how data would be stored and protected. Protocol deviations are reported to our sponsor, our independent safety monitor and the IRB.

Consent for publication

Not applicable.

Competing interests {7b}

AM is the founder/CEO of Momma’s Village Fayetteville, a non-profit organization with a mission to provide access to African-centered birth & breastfeeding support, postpartum care, parenting education and mental health resources for Black families in Fayetteville/Ft. Bragg and the surrounding Sandhills region. MJ is the founder/executive director of MAAME, Inc., a non-profit organization that provides compassionate support, education, and resources that empower birthing families to navigate systems of care and overcome barriers. The remaining authors report no conflicts of interest.