Introduction

The United States (U.S.) has the highest maternal mortality rate among high-income countries, with stark inequities in critical perinatal outcomes, including severe maternal morbidity, preterm birth, and infant death1. In 2023, the national maternal mortality rate was 18.6 deaths per 100,000 live births, but Black women and gender-diverse birthing people (we use ‘individuals’ for brevity) were nearly three times more likely to die from pregnancy-related causes than White individuals2. Infant mortality also reflects this inequity, as Black infants are twice as likely as White infants to die before age one3. Even in California—where the maternal mortality rate was 10.5 per 100,000 live births in 2022, less than half the national average—racial and structural disparities in outcomes persist2,4,5.

Racism is a key driver of inequities in maternal and neonatal health, shaping both access to and experiences of care6,7,8. For instance, Black and other racially or ethnically minoritized individuals are more likely to face economic barriers and challenges such as limited transportation, childcare, and time off work, even when covered by public insurance like Medicaid7,9. Discrimination and lack of person-centered care—defined as care that is responsive to and respectful of individuals’ needs, preferences, and values—are also common6,10,11. Black individuals often report being disrespected or dismissed, reflecting interpersonally mediated racism, which can delay or compromise care even without logistical barriers6,11,12,13,14,15. These inequities, compounded by the daily stress of discrimination, harm health and well-being. Interventions that address equitable access to care and racism and discrimination in care settings are therefore essential to improving pregnancy-related outcomes.

Several innovative perinatal care programs across the U.S. have sought to improve perinatal outcomes and address birth inequities among Black birthing people through culturally responsive, community-centered approaches. These include California’s Black Infant Health Program, which offers group-based care and social support (https://blackinfanthealth.org); Florida’s JJ Way® Clinic, a patient-centered care model that emphasizes equitable care access, provision of dignified and culturally responsive care, and wraparound services (https://commonsensechildbirth.org/the-jj-way); and Washington, D.C.’s Mamatoto Village, which integrates perinatal care with workforce development and community advocacy (https://www.mamatotovillage.org). However, to our knowledge, no cross-sector, community-centered, community-institutional co-led care delivery models exist to address birth inequities in San Francisco (SF) or elsewhere. This gap led our team to embark on a one-year Human-Centered Design (HCD) process in 2017, focused on SF residents facing the most severe inequities, particularly those insured by Medicaid and identifying as Black. Despite SF’s reputation as a progressive city with many equity-focused perinatal programs, participants described excessive practical barriers to care and rampant experiences of interpersonal racism. Referrals between agencies also caused stress because services were fragmented and disconnected16. Our results confirmed extensive evidence that U.S. healthcare delivery systems—built within a culture of white supremacy—often structure care that is difficult to access17,18, fails to meet individuals’ needs, and undermines the autonomy and dignity of Black pregnant individuals6,12,14.

Through the HCD process emerged the “Pregnancy Village” model: a community-institutional collaboration designed to provide a “one-stop shop” of cross-sector services (clinical, city/government, and community-based) in a safe, healing, and uplifting environment, with a focus on earning trust and comprehensively improving the health and wellness of individuals, families, and communities. Through a formal partnership, community-based organizations, institutions (e.g., health systems and city government agencies), and community members with lived experience co-designed and implemented the model together, ensuring it is culturally relevant, responsive to community needs, and driven by community priorities. Key model tenets include the integration of anti-racism principles across all aspects of model design and service delivery and leveraging community-institutional partnerships to ensure effectiveness and sustainability. Additional details on the development and implementation of PV can be found elsewhere16,19. After a three-year planning process, including a partnership with the “Pop-Up Village” event model founders, Designing Justice + Designing Spaces, the first iteration of the Pregnancy Village—the “SF Family & Pregnancy Pop-Up Village” event series (subsequently referred to as the “Pregnancy Village” (PV) for brevity)—was launched19. PV was piloted during 26 monthly events from July 2021 to June 2024.

Given that the sustainability of any health program is predicated on both its accessibility and acceptability, we evaluated these dimensions within the context of PV using the Theoretical Framework of Acceptability (TFA)20. We operationalized TFA constructs by exploring pregnant and postpartum participants and their families’ perceived affective attitude toward the intervention, the effort required to engage (accessibility and burden), their confidence in accessing both PV and external resources (self-efficacy), and their perceptions of intervention effectiveness. Although the TFA includes accessibility, we also examined it separately to capture aspects of accessibility not directly captured by the TFA constructs. To our knowledge, this study represents the first mixed-methods evaluation of accessibility and acceptability of a co-located, cross-sector perinatal care delivery model. This evaluation is situated within our team’s broader mission to engage community members in iterative model refinement to ensure responsiveness to their needs; thus, the present analysis focuses on the formative phase of implementation: the first nine monthly events (July 2021 to June 2022). We aimed to assess the perceived accessibility and acceptability of PV and their associated factors, and to explore barriers and facilitators to PV’s accessibility and acceptability.

Results

Participant characteristics

The analytic sample comprises 116 participants, after excluding four individuals who were determined to be ineligible based on a review of demographic data. Fifteen participants completed the survey more than once, resulting in 89 unique individuals represented in the study. Participants’ sociodemographic and reproductive health characteristics are summarized in Table 1. Fifty-three percent of participants were currently or recently pregnant (within the past year). About half of the participants were between 25–34 (32%) or 35–44 years of age (20%). Thirty-six percent of participants identified as Black or African American, and 38% identified as Latine. Forty-six percent had an educational level beyond high school, and 19% had not completed high school. About two-thirds (63%) of participants were unemployed, and nearly half (49%) were on income assistance. Two-thirds (65%) of participants had public insurance (e.g., Medi-Cal, Medicaid, etc.). Thirty-seven percent of participants were single, and 35% lived with a romantic partner. Twenty-four percent of participants had at least one preterm birth, and 29% had a prior pregnancy loss.

Table 1 Univariate distribution of sociodemographic characteristics, obstetric history, and care discrimination experiences among pregnant, postpartum, and family participants in the San Francisco Family & Pregnancy Pop-Up Village, San Francisco, CA, N = 89
Full size table

Accessibility of PV

The mean accessibility score was 76.0 (SD = 21.0) overall, 73.7 (SD = 24.4) for the pregnant and postpartum participants, and 78.5 (SD = 18.6) for family members (Table 2). The mean accessibility score for Black participants was 78.0 (SD = 20.9) compared to 74.6 (SD = 22.5) for participants from other racial and ethnic groups.

Table 2 Distribution of standardized Accessibility and Acceptability scale scores among pregnant, postpartum, and family participants in the San Francisco Family & Pregnancy Pop-Up Village, San Francisco, CA, N = 116
Full size table

Factors associated with accessibility

In bivariate analyses, several factors were associated with perceived accessibility of PV. Participants aged 25–34 years, those who did not disclose their English proficiency, and those working full-time scored on average 6.5, 17.2, and 16.1 points lower, respectively, than participants aged 15–24 years, proficient in English, and unemployed. Also, those who lived in other areas of San Francisco, reported only “somewhat” having social support, or had private or employer-sponsored insurance, scored on average 15.9, 10.7, and 12.9 points lower, respectively, than participants living in the Bayview, reporting strong social support (“Yes, definitely”), and having public insurance. Participants who did not disclose their food insecurity status and experienced discrimination during their usual prenatal care encounters on some occasions scored on average 16.2 and 15.8 points lower, respectively, than those who often experienced food insecurity and had no experiences of discrimination during prenatal care encounters. Participants with some level of college education, who owned a home or apartment, or were single, scored on average 11.9, 6.3, and 13.2 points higher, respectively, than participants who did not earn a high school diploma, lived in a homeless shelter, or were married or partnered and living together (see Supplementary Table 1 for full sample analyses and Supplementary Table 2 for subgroup analyses).

In the final multivariate model (Table 3), accessibility was significantly associated with English proficiency level, residence, medical insurance status, and experiences of discrimination during prenatal care encounters. Participants who preferred not to disclose their English proficiency level scored, on average, 23.8 points lower than participants who reported being proficient in English (95% CI −40.4, −7.2). Participants who lived in other areas of San Francisco, on average, scored 16.6 points lower than participants who lived in the Bayview (95% CI −24.7, −8.6). Participants who did not have any medical insurance scored, on average, 31.5 points lower than those who had public insurance (95% CI −52.0, −11.0). Participants who experienced discrimination during their usual prenatal care encounters on some occasions scored, on average, 11.0 points lower than those who did not experience discrimination (95% CI −20.8, −1.2).

Table 3 Multivariate mixed-effects linear regression model of select predictors on the Accessibility score, N = 116
Full size table

Participants’ responses from the in-depth interviews supported the quantitative data on accessibility. While most reported minimal burden of accessing PV services, some reported barriers, including lack of language-concordant navigation services for Spanish-speaking participants and provider follow-up.

Facilitators of accessibility

Participants cited several factors that made accessing services easier, including provider responsiveness, short wait times, the physical infrastructure of PV events, and PV’s central location and proximity to their residence. They also highlighted the importance of providers who met them where they were and made a genuine effort to understand and address their needs. As one Multiracial pregnant participant explained:

“Because you can go up to them, but 9 times out of 10, they are going to come up to you and ask you, ‘Do you need some of these services?’ or ‘Try this service out. ‘Because I never got any acupuncture before and that was a new experience for me. But she reached out to me. I didn’t – it wasn’t like I was going over there, so yes.” —Multiracial pregnant participant less than 30 years old

Other participants noted that services at PV were more accessible, citing notably shorter wait times; some reported waiting only a few minutes before being seen by a provider:

“And it wasn’t like a long line. It was just like if I walked to the table, I would be like the first person or the third. But it wasn’t like a long wait. So, it was like I got to visit the majority of all the programs.” —Multiracial pregnant participant, between 30 and 44 years old

Participants noted that PV’s central location and proximity to their residence made it convenient for them to walk or take public transportation to reach PV:

“You’re [Pregnancy Village] just right off 3rd and McKinnon. That’s just half a block I have to go. Usually, them blocks are long. So, 104 and up there – they [are] long. But it was just a little hop, skip, and a jump, and I was right there. It’s easier – you didn’t have to walk to catch the bus, or the T-train, the 24, the 54. All of it is right there. Even the 15 comes down that way. So, it’s very easy to get there.” —Black family member, 45 years old or above

Barriers to accessibility

While participants generally reported a low burden in accessing and receiving services at PV events, some identified challenges related to the physical layout, inadequate service navigation support, information overload, and lack of provider follow-up. Additionally, difficulties in learning about available services were attributed to the lack of designated navigators to guide them through the services:

“I thought it was going to be like a meeting with […] somebody who could give you details [like a prenatal group], and then I didn’t see anybody, so I was confused. And nobody explained it to us until someone told me, ‘You can go there and get some information.’” —Latine pregnant participant, between 30 and 44 years old

A few participants noted that lack of provider follow-up after the event made it difficult to access the services they needed, as this Multiracial pregnant participant recounted their experience:

“Well, with X organization, my experience was good up until they didn’t call me back. They didn’t follow up. So, they were not accountable for their word. So, I felt like [having connected with them] was pointless because getting that interaction with them and setting that all up with a doula was like the main focus of why I wanted to go. And that’s why my doctor encouraged me to go.” —Multiracial pregnant participant, less than 30 years old

Acceptability of the PV

The mean acceptability score was 91.9 (SD = 14.4) overall, 90.5 (SD = 15.7) for pregnant and postpartum participants, and 93.4 (SD = 12.9) for family member participants (Table 2). The mean acceptability score for Black participants was 94.2 (SD = 12.8) compared to 90.3 (SD = 15.3) for participants from other racial and ethnic groups.

Factors associated with acceptability

In bivariate analyses, several factors were associated with perceived acceptability of PV. Participants who spoke Spanish, had limited English proficiency, and had at least one preterm birth scored on average 8.1, 4.9, and 8.0 points lower, respectively, than participants who spoke English, were proficient in English, and did not report a history of preterm birth. Those who had no medical insurance, did not disclose their food insecurity status, and experienced discrimination during their usual prenatal care encounters on some occasions scored on average 18.0, 11.5, and 11.4 points lower, respectively, than participants who had public insurance, had no food insecurity, and had never been discriminated against during prenatal care encounters. Participants who worked part-time, owned a home or apartment, and did not experience food insecurity scored on average 7.3, 6.5, and 5.3 points higher, respectively, than those who were unemployed, lived in a homeless shelter, and often experienced food insecurity (see Supplementary Table 3 for full sample analyses and Supplementary Table 4 for subgroup analyses).

In the final multivariate model, acceptability was significantly associated with language, medical insurance, and experience of discrimination during prenatal care encounters (Table 4). Participants who spoke Spanish found PV less acceptable than those who spoke English, scoring, on average, 7.9 points lower (95% CI −15.6, −0.2). On average, participants without medical insurance scored, on average, 16.3 points (95% CI −28.0, −4.7) lower than those who had public insurance. Participants who experienced discrimination during their usual prenatal care encounters found PV less acceptable than those who did not experience any discrimination during care encounters, scoring on average 10.3 points lower (95% CI −15.5, −5.0).

Table 4 Multivariate mixed-effects linear regression model of select predictors on the Acceptability score, N = 116
Full size table

Participants’ responses from the in-depth interviews supported the quantitative data on acceptability. We summarized the data on the acceptability of PV using six deductive domains based on the TFA20. The definitions of the TFA domains and additional excerpts of the outlined domains are shown in Fig. 1.

Fig. 1: Theoretical Framework of Acceptability (TFA) domains and corresponding quotes.
Fig. 1: Theoretical Framework of Acceptability (TFA) domains and corresponding quotes.
Full size image

This figure displays the domains of the Theoretical Framework of Acceptability, paired with illustrative quotes that capture participants’ perspectives on each domain. Designed by Jorrín Abellán, I.M. Hopscotch Building: A Model for the Generation of Qualitative Research Designs (2016)52.

Anticipated affective attitude

When asked about their expectations of PV prior to attending, some participants anticipated a health fair-type setting focused on pregnancy-related resources and information. Upon reflection, many reported that these expectations were met, highlighting the inviting and uplifting atmosphere, as well as the abundance of resources, including mental health and wellness services such as massages and acupuncture, and prenatal offerings such as diapers, wipes, and clothing. Participants also noted gaining valuable awareness and knowledge of services designed to improve their pregnancy experience:

“[I expected PV to be] kind of similar to what it is. So, definitely, organizations and businesses that have resources for pregnant people. I didn’t expect there to be that many giveaways for pregnant people or services in terms of massages. So, more just like resources, knowing like, ‘Hey, this is where you can get XYZ. This is available to you here in San Francisco.’ And the education part, I kind of did expect, like the breastfeeding education.” —Pregnant participant, less than 30 years old, race/ethnicity not reported

A few participants, however, reported having no expectations prior to the event, as they were unfamiliar with community-centered events that offered both information and tangible resources. For example, one family member noted:

“Well, you know, I didn’t have any expectations. I never heard of anything like that, never been to a community gathering of that sort. So, I didn’t have a whole lot of expectations. But, as I read the flyer and read the different organizations that were going to be there, I kind of felt good about attending and felt like it was worthwhile.” —Black family member, 45 years old or above

Experienced affective attitude

Participants generally characterized the PV atmosphere as positive, vibrant, community-centered, down-to-earth, and healing. When asked how they felt upon arrival, many expressed feeling welcomed, excited, and filled with anticipation. For instance, one Black postpartum participant had this to say:

“I was happy. I was like, ‘Wow. Look at all of this.’[…] because I had my son there – he’s seven-and-a-half months. And I was like, “Look!” He really loved the color of the tents and the way the tents were kind of blowing in the wind. He loved the live music. It was just like, “Wow. Okay. Yes!” And it was like it was our first time being out in community ever. And so, I was pleased that it was such a warm, welcoming, fun environment for us to walk into. You know? You walk in and you see like the bags of groceries that, you know, were being handed out and things like that.” —Black postpartum participant, 45 years old or above

Participants also expressed their appreciation and surprise at the introduction of new services and resources at each event, noting these additions as an unexpected but welcome enhancement:

“I was excited to see what was new, because you guys don’t always have the same things; sometimes what you all give away and what you all do for us is unexpected, especially for the resources. I didn’t think the CalWORKS individuals were going to be there for the moms that don’t have any type of income.” —Multiracial pregnant participant, age not reported

Some participants described their initial impressions of PV as a community-centered, curated space where Black individuals could gather, connect, and access pregnancy-related services and resources. They shared that being surrounded by other Black individuals or expectant mothers made them feel at home—free to express themselves and engage in activities without fear of judgment:

“I felt kind of like at home, because I wasn’t the only one pregnant, or there were mothers who were there that were very comfortable breastfeeding and stuff. So, seeing that was like, oh, ‘this is a place for us to just be open and free without no judgment.’” —Multiracial Pregnant participant, 30–44 years old

“It was about seeing Black individuals coming together without fussing or arguing. We were there. We enjoyed each other. We laughed. That makes you feel good when Black individuals come together and everybody enjoys themselves […] I looked around. I’m like, ‘Wow, this is neat! This could help a lot of young girls.’ You know, teach them. That’s what we need to do is teach them the right way.” —Black family member participant, 45 years old or above

The community-centeredness of PV was especially salient in the context of the social isolation experienced during the COVID-19 pandemic, as illustrated by the account of this Black postpartum participant:

“I was pregnant from 2020 all the way through 2021. That was like the height of when everything was like shut down and isolated. We were having fires. I couldn’t go outside to walk even, hardly. It was extremely isolating. And so, one of the benefits that I don’t think that you all are necessarily picking up on is that, for individuals who have been isolated because you’re trying to follow COVID protocols and all of that, it was healing to just be out in the fresh air, in the sunshine, with other mothers. Like that, in and of itself, was the healing thing for me.” —Black postpartum participant, 45 years old or above

Participants reflected that attending PV had made them more aware of shared interests with others in their community, which, for some, had given rise to a desire to become more actively involved in their community:

“Going to the Pop-Up just opened my eyes to there’s so many individuals out there that I have common ground with. And we can all work on ourselves together. We could help each other. Some individuals don’t know where to start. So, that’s where I’m at. It kind of inspired me to definitely see what I could do for my community. But first, yeah, I felt very good.” —Latine postpartum participant, between 30 and 44 years old

Most participants said their expectations were met, with several noting that they were exceeded. This was largely attributed to the plethora of resources and services available, which addressed physical, psychological, and spiritual well-being:

“[My expectations] were met. I received a lot of information from different groups on childcare, doula services, even when the baby starts growing her first teeth, even a dentist that’s nearby my house. And chiropractic services for pregnant women – I didn’t even know you can get realigned as a pregnant woman.” —Black pregnant participant between 30 and 44 years old

Others noted that their expectations were met because they felt acknowledged and valued at PV, as one pregnant participant shared: “I really felt acknowledged and seen, and individuals were really trying to take care of me because they saw that I was pregnant.” —Pregnant participant, less than 30 years old, race/ethnicity not reported

However, a few participants reported that their expectations were unmet, citing insufficient guidance on the availability and use of relevant resources:

“Yeah, so once again I have, you know, totally different expectations. So when I went there, and I see just a bulletin and nobody, nobody is, like, welcoming you, you know. But, if someone doesn’t know anything, like where do you go? I just had to ask around, and they told me, like, ‘Okay, you can get some information here in that first booth.’ But I didn’t really get that information well.” —Latine postpartum participant, between 30 and 44 years old

Self-efficacy

Overall, participants expressed increased confidence in accessing services both within and beyond PV following their attendance. This enhanced confidence was attributed to PV’s strong referral system, exposure to new services, and the provision of clear and comprehensive information by service providers:

“I feel like it was a linkage sort of situation. So that if I needed something more extensive, then I would have the information to know like where to go, who to talk to – all of that – maybe even set up an appointment. So, when I did need to go and access that deeper level of services, I could go out into the community, and then do that.” —Black postpartum participant, 45 years or above

However, a few participants reported that participation in PV did not increase their confidence in accessing new services, primarily due to limited awareness of the full range of available services. In these cases, information was typically provided only by organizations with which they were already enrolled, thereby restricting exposure to new services:

“I don’t think so, I mean, the only, the only things that I saw that was convenient for me was the programs that I already know. But anything else that I don’t know, I don’t know. The Homeless Prenatal, like I’ve been a customer before. And then for, for WIC, you know, I’m a client already. So, anything else, I have no idea.” —Latine pregnant person between 30 and 44 years old

Appropriateness

Participants generally perceived the services as appropriate and aligned with their expectations. Several noted that the services effectively addressed the needs of pregnant individuals and their families, reinforcing their relevance and value:

“I feel like they’re appropriate because the need is so great right now. A lot of individuals are on hard times. A lot of individuals are struggling. And, with these organizations being implemented the way they have to help families, it just kind of lightens the load. It gives me just that great feeling of knowing – even though I’m on hard times right now – that I don’t have to worry because I can get the help that I need until I’m in a better place and can do better for myself. So, I feel like it’s totally appropriate.” —Black family member, 45 years old or above

Participants viewed the services as appropriate, particularly because they provided pregnancy-related education and addressed gaps in breastfeeding education:

“Even though I had already been through breastfeeding when I first started with my first child, I realized it is really something that women need support on. Even though it is natural, it doesn’t come naturally to a lot of people. And just knowing that there was this booth that really educated individuals on the latch, and how breast milk is established, and that it might take a while.” —Pregnant participant, less than 30 years old, race/ethnicity not reported

A few participants, however, perceived certain services as inappropriate, citing a lack of focus on pregnancy-specific support. As one Latine pregnant participant explained:

“I went for information for pregnancy, and when I went there, I didn’t find anything kind of related, at least, you know, with my eye at the beginning. And then I guess it was a couple of books for that, but I didn’t see anybody or was expecting to see more people, you know, kind of with a similar concern that I have, and I see nobody.” —Latine pregnant participant between 30 and 44 years old

Perceived effectiveness

Participants generally perceived PV to be effective, noting its uplifting environment and focus on centering Black women:

“I think they did a good job of providing that really positive atmosphere. Especially I feel like it was centering around women of color and Black women, and sometimes we will often feel like society looks down, like ‘oh you are pregnant.’ I think they did a good job of making it feel like a positive thing you’re pregnant.” —Black pregnant participant, between 30 and 44 years old

Other participants attributed PV’s effectiveness to its responsiveness to community needs, including proactive outreach and the delivery of resources in an accessible way:

“I like that you guys like to be part of the community, and you like to see what is going on in the community. It is really hard right now, especially to be homeless, and not have a lot of resources - that’s one thing I like - for you guys to go and give back to the community and say ‘Hey, we have resources. We want you guys to come out and have a good time as well.’ And on top of that, it also helps with depression.” —Multiracial pregnant participant, age not reported

Usefulness

Participants generally regarded the services and resources provided at PV as useful. Some participants cited pregnancy and postpartum-related education, such as newborn hygiene and nutrition, as well as mental health support, as key reasons for its perceived usefulness:

“Well, at the one table that was in the middle, I overheard them talking about how to give your baby a bath, how to do this with your baby, do that with your baby. That was so cool, because I saw individuals with babies and stuff. I was kind of listening a little bit. They were sitting there talking about their kids, what their kids do, what their kids eat, what they didn’t eat, what’s good for them, and what’s not good for them. And that’s what they need. They need to know more about baby nutrition.” —Black family member, 45 years old or above

Several participants found PV useful because it offered educational resources for children and child-related activities at events:

“I saw how they were with the kids. The little kids were up there drawing, doing all this stuff. You know, they were really, I think they were really supportive, helping all the little kids up there doing the thing.” —Black family member, 45 years old or above

Service recommendations

In the in-depth interviews, participants shared several recommendations to improve the PV model. These included adding or improving specific services, such as provider follow-up with interested individuals; provision of more food, particularly for pregnant individuals; expanded pregnancy education sessions; more support and education around mental health, including postpartum depression and managing the stress of newborn care; and additional child-focused activities and resources, particularly those that promote parent-child engagement:

“I think they did have a booth for, like, mental health support. But definitely more education around health and depression, and the fourth trimester, and postpartum essentials that are really helpful for a mother for recovery, and then also for a baby. Just that it’s not just about the pregnancy or when you give birth, but then the first few weeks or the month after that might be really intense on women.” —Black pregnant participant, less than 30 years old

“If you guys have like some sort of guided mommy and me activity like with music and stuff like that, where you could just kind of come over with your baby and be led in some like movement activities and stuff. That would be fun.” —Black postpartum participant, 45 years old and above

Others shared infrastructural recommendations, including the provision of more comfortable seating, the engagement of navigators to guide visitors and share information—particularly in their preferred language—and enhanced outreach to ensure broader community awareness of available services.

Sensitivity analyses

In sensitivity analyses in which missing and duplicate responses were excluded, we identified no differences in standardized acceptability or accessibility scores (see Supplementary Table 5). The acceptability score was 90.8 (SD = 15.2) overall (N = 85), 90.5 (SD = 15.9) for pregnant and postpartum participants (n = 45), and 91.2 (SD = 14.5) for family members (n = 40). The accessibility score was 74.5 (SD = 22.6) for the main sample (N = 87), 71.0 (SD = 24.2) for pregnant and postpartum participants (n = 46), and 78.5 (SD = 20.2) for family members (n = 41).

Discussion

Participants’ perceptions of PV’s accessibility and acceptability were high, suggesting the Pregnancy Village model is accessible and well-accepted by the community. Accessibility and acceptability were both patterned by language proficiency, socioeconomic status, and discrimination experiences within prenatal care. Quantitatively, lower scores were reported among Spanish speakers, those without medical insurance, and those who had experienced discrimination during prenatal care. Qualitatively, while most participants reported minimal burden in accessing services, some noted challenges related to language-concordant navigation and lack of provider follow-up. Nevertheless, participants described PV as inviting and resource-rich, felt confident accessing services during and after events, and perceived the offerings—ranging from wellness care to prenatal education—as useful and effective. These findings support the model as both accessible and acceptable, offering a pathway to improve the experiences of Black and other minoritized pregnant and postpartum individuals and their families.

Quantitative findings on the accessibility of PV suggest some room for improvement. One reason for this is that although PV was focused on SF residents, some participants came from other parts of the Bay Area. Thus, the time to reach events was longer for these individuals, who aren’t the target population for the program. The main barriers to accessibility cited in the qualitative interviews were, however, related to the burden of participation, as in the effort required by participants to participate in and access PV services once at the location. Although the resources provided were viewed as valuable and useful, some participants noted difficulty navigating and interacting with the resources, which suggests a need to optimize service linkage and delivery, even in community care settings such as PV. Some participants, particularly Spanish-speaking participants, had lower perceptions of the events’ accessibility due to the language barrier, making it more difficult for participants to obtain the services they needed. This is consistent with research indicating that a language barrier is an obstacle among migrant and diverse women who seek prenatal care21.

The high acceptability of PV was reflected in participants’ positive feelings about PV’s vibrant, welcoming, and healing atmosphere. Research on the built environment highlights how spatial features—such as openness, natural elements, and connectivity of spaces—can influence mood, engagement, and comfort22,23,24. Thus, PV’s outdoor setting, vibrant design, and diverse activities and services likely contributed to its perceived restorativeness22. Rather than merely providing services, PV cultivates a culturally affirming space where individuals feel seen, respected, and valued25. This environment may enhance affective attitudes by fostering hope, social connection, and a sense of personal growth26. Further, these affective dimensions are not peripheral but central to the intervention’s acceptability, suggesting that future community sociomedical care delivery models should explicitly integrate environmental and cultural design as core components of engagement.

Participants highlighted PV’s community-centeredness as a key driver of its acceptability, particularly relevant in the context of heightened social isolation during the COVID-19 pandemic. The opportunity to engage in group activities such as food demonstrations, yoga, and sharing circles offered more than health education; rather, it fostered connection, mutual support, and a shared sense of purpose27,28. This aligns with research showing that a strong sense of community enhances problem-focused coping strategies and facilitates engagement in health-promoting behaviors, especially in under-resourced settings27,29. These findings suggest that sociomedical interventions aiming to increase acceptability should center community-building as a structural feature, and not just a byproduct of health promotion.

Participants’ acceptance of PV was also linked to an increased sense of confidence in navigating and utilizing available resources both within and outside the PV events, suggesting that the intervention promoted self-efficacy. Access to clear, relevant information—delivered through PV’s resource-oriented and health literacy-focused activities—appears to have played a key role in strengthening participants’ belief in their capacity to take action. This aligns with existing evidence demonstrating that strategies promoting health literacy and community engagement can enhance self-efficacy, particularly among disadvantaged populations30,31. However, it is important to consider that individuals who attend events like PV may already possess a higher baseline self-efficacy, as prior research has shown that self-efficacious individuals are more likely to seek out and participate in health-promoting activities32,33.

Quantitative findings also revealed that participants who had experienced discrimination during prenatal care encounters reported significantly lower acceptability of PV. This suggests that prior negative encounters with the healthcare system—or with organizations affiliated with PV—may shape how individuals perceive the PV model. Such skepticism is consistent with prior research showing that experiences of discrimination can undermine trust and lead to the rejection of social care interventions34. These findings highlight the importance of embedding culturally responsive and trust-building practices into community-centered sociomedical interventions like PV, not only to increase engagement but also to begin repairing the harms caused by structural racism in health and social systems.

Participants offered several suggestions to enhance PV’s accessibility and acceptability, which align with existing research. First, consistent post-event follow-up is essential for maintaining trust and ensuring participants can access services introduced at PV. Therefore, a structured follow-up system would help bridge the gap between interest and utilization. Second, the presence of navigators to guide visitors to relevant services will enhance usability and reduce barriers to engagement. Third, the request to provide Spanish-speaking PV providers or translators aligns with research linking language-concordant care to improved communication, satisfaction, engagement, and outcomes35. Fourth, the provision of food can address meeting community food needs and food access36, foster community engagement37, and encourage engagement in health-promoting behaviors38. Fifth, the call for expanded perinatal education is essential as research has demonstrated its role in reducing childbirth anxiety39 and depression40 while improving self-efficacy41. Lastly, targeted outreach—especially to PV’s priority population of Black birthing and postpartum individuals and their families—is necessary to deepen PV’s reach, given the importance of sustained advocacy and tailored outreach in order to serve marginalized populations effectively42. Importantly, these recommendations emerged through PV’s embedded community feedback mechanisms and were incorporated into future events, underscoring the value of community co-design in sustaining culturally responsive, adaptive interventions.

This study has some limitations. First, because the evaluation was in the context of real-world implementation and the model’s dynamic co-creation approach allowed it to adapt to the expressed needs of the participants, the model being evaluated was inconsistent over time. Given that PV events were designed to respond to community needs, the factors cited as barriers to access or reasons for participants’ perceived low acceptance of the model became less prominent as the evaluation continued. Second, our purposive and convenience sample limits generalizability to other populations. To help minimize sampling bias, we aimed to maximize diversity within the sample by recruiting participants across multiple PV events to capture a broad range of experiences. Nevertheless, future studies should consider using sampling methods that enhance representativeness and support broader generalizability of findings. Third, the 3-item accessibility index showed relatively low internal consistency (α = 0.53) and item-rest correlations (0.22–0.36) as the items reflect distinct but related aspects of perceived accessibility. Given the importance of capturing those aspects of accessibility, we maintained the index and chose to report the summative score rather than the individual items to reduce the number of outcomes reported in the paper. Thus, the findings should be interpreted with appropriate caution. Fourth, although the response rate was high among eligible participants, the absence of data on non-responders limits our ability to assess potential selection bias. Fifth, completing the survey onsite by some participants may have introduced social desirability bias. To mitigate this, surveys were self-administered and anonymous, with assurances that responses would remain confidential and have no impact on participants’ involvement in PV events. In addition, given that participants could respond multiple times if they attended multiple events, the data may overrepresent the experiences of those who had positive experiences and returned to subsequent events. Sensitivity analysis, however, suggests this is unlikely since the results did not change significantly when restricted to only the responses from unique individuals.

The study has several strengths. First, is the use of a mixed-methods design: the in-depth interviews enabled us to gain a more nuanced understanding of participant perceptions across acceptability domains than was feasible with the quantitative assessment alone. Second, is the use of the Theoretical Framework of Acceptability as an orienting framework. We identified two additional acceptability domains in our study (i.e., perceived usefulness and appropriateness), which are not captured in the TFA domains. We, therefore, suggest modifications to the TFA to include these domains that may be useful for other researchers evaluating the acceptability of community sociomedical interventions. Further, by involving key partners in the development of data collection tools, including community members and the implementation team, we were able to identify questions that were relevant to both participants and those designing the events.

Reducing disparities in maternal and neonatal health requires expanding access to a broad range of perinatal supports—beyond what is traditionally provided in clinical settings—and delivering these in ways that are culturally responsive and acceptable. Our findings suggest that a “one-stop shop” model like PV, which co-locates diverse services in a community-centered and uplifting environment, is an acceptable and promising approach to meeting these broader needs. However, sustaining such programs will require supportive policies that prioritize a holistic, integrated, and community-centered approach to improving health43. One key policy strategy is to incentivize community-institutional collaboration and move away from siloed service delivery43. For example, California’s CalAIM Population Health Management program mandates that managed Medicaid plans conduct community needs assessments and invest in community health improvement projects that address upstream determinants of health through integration of healthcare, public health, and social services, while also fostering trust and engagement at the community level44. Longer-term meaningful collaborations may also entail redesigning health and social systems to overcome structural barriers and simplifying access to equitable care, which is the overarching goal of the “Pregnancy Village” vision.

Our study also underscores the importance of critical implementation practices that actively center community voice. Feedback loops—such as surveys, vision boards, and interactive activities embedded into PV events19—not only surfaced needs around navigation support, transportation access, food provision, and culturally responsive care, but also allowed for responsive changes to be made in real time. These challenges reflect persistent structural inequities disproportionately affecting people of color and low-income communities. Therefore, designing truly equitable programs requires not only responsive adaptation but a foundational anti-racism approach that interrogates and seeks to transform the systems that produce and perpetuate these inequities.

Future research should examine longitudinal changes in accessibility and acceptability of the PV model, while refining and validating accessibility measures to better capture participant experiences. Evaluating the model’s scalability across diverse settings and populations is also critical to understanding its broader applicability and effectiveness. Parallel to these efforts, rigorous evaluation of the model’s impact on key outcomes—including care access, participant experience, and mental well-being—is essential. Ultimately, this research aims to establish clear links between these intermediate outcomes and measurable improvements in clinical outcomes. Such a comprehensive evaluation will provide vital evidence to inform model refinement, policy support, and widespread implementation of the PV model and similar community-centered sociomedical interventions.

The first iteration of the Pregnancy Village model—the SF Family & Pregnancy Pop-Up Village—demonstrated high acceptability (92/100) and good accessibility (76/100), underscoring the model’s potential as a community-centered, cross-sector approach to addressing perinatal support needs. However, access barriers related to language proficiency and experiences of prenatal care discrimination underscore the importance of language-concordant staffing, navigation support, and culturally responsive care. Further, these findings highlight the need for ongoing community-institutional co-creation to ensure alignment with evolving community priorities. Future efforts to sustain and scale the model may benefit from policy support, long-term funding mechanisms, and shared leadership between community and institutional partners. Continued process and impact evaluations will be critical to understanding how this model can be adapted and scaled in other contexts.

Methods

Setting

San Francisco’s Bayview district was chosen as the PV site as it is home to 19% of Medicaid-insured and 36% Black-identifying birthing SF residents45. Most birthing individuals in the Bayview are Medicaid-insured (61%), and 93% identify as members of a racial or ethnic minority group46. Furthermore, Bayview residents have significantly lower rates of timely prenatal care (inequities in access) and higher rates of preterm birth (inequities in outcomes) than residents of other neighborhoods, even among Medicaid-insured individuals only45.

The Bayview combines industrial, commercial, and residential zoning areas, but industrial shifts have caused infrastructure disinvestment over time. Car traffic is higher than foot due to Bayview’s hilly terrain, few street trees, and accessible parks. PV events were initially centrally located near a major thoroughfare with public transportation and community institutions. Proximity to trusted community institutions was intentional to encourage community trust and engage these organizations. Vibrant colors, shade, ground treatment, and diverse seating were incorporated into the event design to transform the area into an active and welcoming space that represented PV values.

Intervention

The PV model of cross-sector collaboration aims to address perinatal inequities by providing a one-stop shop for comprehensive care and support in a dignified and empowering environment tailored to Black pregnant individuals and their families. Monthly events bring together various partners, including city government, healthcare, and community-based organizations. The range of services provided includes healthcare (e.g., Medicaid enrollment and consultations with medical professionals), as well as holistic wellness services (e.g., dance and cooking demonstrations, acupuncture, massage, and sharing circles). The model is rooted in anti-racism and person-centered care principles, prioritizes sustainable community-institutional partnerships, and includes a real-time community feedback mechanism to ensure continuous model improvement (Fig. 2). More information about its implementation can be found elsewhere19.

Fig. 2: Pregnancy Village model components and features.
Fig. 2: Pregnancy Village model components and features.
Full size image

This figure illustrates the core components of the Pregnancy Village model, designed to provide a one-stop shop of cross-sector services in a safe, healing, and uplifting environment for pregnant individuals and their families.

Study design

We employed a convergent mixed-methods, community-engaged process evaluation involving triangulation of qualitative and quantitative data to assess: 1) the feasibility and fidelity of the PV model; 2) accessibility and acceptability of PV among participants; 3) factors affecting the sustainability of provider participation; and 4) the preliminary impact of PV, such as perceptions of person-centeredness, comfort, and trust. This paper focuses on the findings related to accessibility and acceptability.

Sample

We recruited a convenience sample of pregnant and postpartum individuals and family members from the first nine PV events between July 2021 and June 2022. This period was selected because it signifies the early phase of PV, when both participants and providers were adapting to novel processes within a co-led community-institutional model, and several adaptations to the model were being made in response to emerging needs and feedback. Capturing and examining these early dynamics was therefore critical to understanding how the model functioned in its first iteration, how participants perceived the care environment, and how these early experiences might shape longer-term engagement and outcomes. Although the Pregnancy Village model and resulting SF Family & Pregnancy Pop-Up Village events focused on improving the perinatal experience and outcomes of Black individuals and were designed accordingly, the organizers believed that the events would also be appealing and beneficial to other pregnant individuals and families facing barriers to care, such as other minoritized groups and/or those living on low incomes. We, therefore, targeted Black and other minoritized pregnant or postpartum individuals and their families who were receiving services at PV for the evaluation. Inclusion criteria were: (1) at least 15 years old if pregnant or postpartum, or 18 years old if family; (2) participation in at least one PV event; and (3) able to speak English or Spanish. We aimed to recruit 120 individuals in total, targeting a feasibility number of 10–15 from each event. Eligible participants who attended multiple PV events were allowed to participate multiple times. We enrolled 120 individuals for surveys and purposively selected a subset of participants (n = 18) for semi-structured in-depth interviews, maintaining a 70–30 ratio between pregnant and postpartum individuals (n = 13) and family members (n = 5). We employed a purposive sampling strategy informed by the principle of maximum variation, with the goal of capturing a diverse range of experiences by considering factors such as race/ethnicity, pregnancy status, language, and age47. Interviews were conducted until thematic saturation was achieved, as no new themes emerged in the final interviews.

Recruitment and data collection

Quantitative

During registration at PV events, potentially eligible participants were invited to learn more about a study evaluating their experiences. Study team members provided information about the study, screened interested individuals for eligibility, and invited those who were eligible to participate. After obtaining verbal informed consent, participants were invited to complete a survey about their experiences onsite using a study tablet or later using their own devices via a QR code; most opted to do it on-site. The survey assessed various domains of the PV experience. Questions were adopted or adapted from validated scales or developed by the study team with input from a community advisory board (CAB), informed by cognitive interviews with three pregnant and postpartum individuals from the Bayview and the greater San Francisco Bay Area. Participants received a $20 gift card as compensation for completing the survey. Out of 104 eligible participants enrolled, 89 completed the survey, yielding a response rate of ~86%. However, 15 participants completed the survey more than once, yielding a total of 116 surveys.

Qualitative

Three to four survey participants from each PV event were invited to take part in in-depth interviews. The interviews were scheduled at times that suited the participants. The semi-structured interview guide was informed by the Theoretical Framework of Acceptability (TFA) and included additional questions on accessibility, person-centered care, community engagement, and trust in the healthcare system. Interviews lasted 30–60 min and were conducted via Zoom in either English or Spanish; participants could join through the Zoom app or by phone, ensuring access regardless of internet or device limitations. Interviews took place within four weeks of attending a PV event and were conducted by researchers with qualitative training (OJO and KV). In-depth interview participants were compensated an additional $20. In-depth interviews were recorded with participant permission and transcribed for analysis by a third-party transcriptionist. Transcripts were checked for accuracy and clarity by members of the study team (KV and JV). Field notes were taken during the interviews, and a templated summary outline was created afterward, facilitating rapid analysis for model iteration19.

Quantitative measures

Accessibility

Participants’ perceptions of PV accessibility were assessed through an accessibility index composed of three questions based on investigators’ prior experience conducting research with this target population (Supplementary Table 6): (1) “How long did it take you to get to PV from where you typically stay?” (2) “How do you feel about the time it took you to get to PV?” (3) “Compared to how you access your usual sources of care and resources, would you say the services at Pop-Up Village are…?” Responses to the three items were summed and then standardized (range 0–100), with higher scores indicating higher accessibility. Missing data (3.4%) were imputed as the mean of other items in the measure. Internal consistency reliability for accessibility within this sample was Cronbach’s α = 0.53. Item-rest correlations for the three accessibility items ranged from 0.22 to 0.36 (travel time = 0.25; emotional response to travel time = 0.36; comparison of service accessibility = 0.22), indicating moderate correlations. These results suggest that each item contributes uniquely to the overall scale and support the intended multidimensional structure of the accessibility construct.

Although the TFA encompasses aspects related to accessibility—such as burden—we chose to assess accessibility as a distinct construct in our quantitative analysis. This decision was driven by our goal to capture concrete logistical and structural barriers (e.g., physical access and time) that may significantly shape participant engagement but are not fully accounted for in the TFA’s conceptualization. Hence, treating accessibility separately allowed for a more nuanced evaluation of implementation challenges in this community-based setting.

Acceptability

Acceptability was assessed using a 7-item scale adapted from the acceptability of health apps among adolescents (AHAA) scale for the acceptability of health interventions (Supplementary Table 7)20,48. We assessed the following sub-domains of TFA: (1) affective attitude (i.e., how an individual feels about the intervention); (2) burden (i.e., the perceived amount of effort that is required to take part in the intervention; (3) intervention coherence (i.e., the extent to which the participant understands the intervention and how it works); (4) opportunity costs (i.e., the extent to which benefits, profits, or values must be given up to engage in the intervention); (5) perceived effectiveness (i.e., appropriateness or the extent to which the intervention is likely to achieve its purpose); and (6) self-efficacy (i.e., participants’ confidence that they can perform the behavior required to participate in the intervention) (2). Response options were structured using a 4-point Likert scale. Negatively worded questions were reverse-coded before generating summative scores, and any missing data (3.4%) were imputed as the mean of other items in the measure. The scores were standardized to range from 0 to 100, with higher scores indicating higher acceptability. Internal consistency reliability for acceptability within this sample was Cronbach’s α = 0.84.

Covariates

Participants self-reported sociodemographic characteristics, including age, gender, race and ethnicity, educational attainment, neighborhood of residence, employment status, receipt of public income assistance, preferred language, level of English proficiency, housing status, social support, medical insurance status, food insecurity49, and relationship status. They also shared their pregnancy status (pregnant or postpartum, or family member of pregnant or postpartum person), parity, history of preterm birth, history of pregnancy loss (e.g., miscarriage, induced abortion, or stillbirth), and any prenatal care attendance during current/recent pregnancy (if pregnant/postpartum participant). Experiences of discrimination were assessed during prenatal care encounters and daily activities, using questions adapted from the Discrimination in Medical Settings Scale and the Everyday Discrimination Scale, respectively50. (https://scholar.harvard.edu/davidrwilliams/node/32397).

Analyses

Quantitative

We performed descriptive statistics to examine participants’ sociodemographic characteristics and accessibility and acceptability scores. Prior to bivariate analysis, missing responses to the experience of everyday discrimination and experience of discrimination during prenatal care encounters questions were recoded to the middle response (“Sometimes”). We examined bivariate associations between accessibility and acceptability, as well as the sociodemographic characteristics, pregnancy, obstetric factors, and experiences of discrimination, using ordinary least squares (OLS) linear regression to identify statistically significant associations, accounting for clustering due to repeat survey responses. Statistically significant predictors (p < 0.05) were considered for inclusion in multivariate models. Multivariate models were refined by excluding closely related variables identified through collinearity tests and model fit assessment. Given that some participants answered more than once, we estimated linear mixed-effects models to account for clustering. A total of 15 participants completed the survey more than once across different PV events. This modeling approach allowed us to account for both within- and between-person variability, reducing bias due to repeated measures. To assess the robustness of the findings, we also performed sensitivity analyses using only the first survey response from each participant. Stata (version 14) was used to conduct all analyses (StataCorp, Stata Statistical Software: Release 14, 2017).

Qualitative

We used a framework analysis approach to analyze the qualitative data51. The qualitative lead (OJO) developed an initial deductive codebook based on the interview guide. Six analysts (OJO, PD, KV, EK, HS, and KM) independently coded transcripts using Dedoose software (https://www.dedoose.com), and two transcripts were reviewed collaboratively to ensure inter-rater reliability. We compared and discussed coding in detail, refining the codebook, and adding inductive codes. The remaining transcripts were coded individually by balanced pairs of analysts, and the qualitative lead reviewed the coding to ensure consistency. Codes related to the acceptability of the PV model and accessibility of PV services were then queried and evaluated. Analytic summaries for the codes were compiled in a framework table as follows: (1) anticipated affective attitude; (2) experienced affective attitude; (3) self-efficacy; (4) perceived effectiveness; (5) usefulness of PV services; (6) appropriateness of PV services; (7) barriers to access; (8) facilitators of access; and (9) service recommendations and suggestions for improvement.

Reflexivity

Reflexivity was an integral component of our methodology, particularly in light of co-authors’ varying degrees of involvement in the PV intervention, including development, implementation, and evaluation (MAN, GJ, and KC). Notably, some co-authors also served as community implementing partners (GJ and KC), contributing their critical lived and professional experiences that shaped the design and implementation of PV, as well as the dissemination of key findings. While these dual roles provide important contextual insights, they can also introduce the potential for bias. To address this, we engaged in ongoing reflexive practice to explore how our positionality and individual experiences could affect data collection, analysis, and interpretation. Further, quantitative data were self-administered, and qualitative data were collected and analyzed by team members who were not involved in the implementation, reducing the likelihood of social desirability bias. With respect to qualitative analysis, we employed a systematic approach, including collaborative coding, joint reviews of transcripts, and regular analytic discussions to ensure consistency and challenge our assumptions. While our proximity to the intervention may have shaped our analytical lens, we also consider it a strength when critical reflection and rigorous methodology are maintained.

Mixed-methods integration

Integration occurred during the design and sampling stages, as described above. We employed a convergent design, aligning questions in the in-depth interview guide with the same framework used in the survey questions to expand upon or explain the quantitative findings. Additionally, the qualitative sample was drawn from the quantitative sample. We also presented the quantitative and qualitative findings together in the results section using the weaving approach for joint interpretation.

Ethics approval

This study was conducted in accordance with all relevant ethical regulations, including the Declaration of Helsinki. We obtained ethics approval from the Institutional Review Board of the University of California, San Francisco (#20–32393). Verbal informed consent was obtained at PV events after being given information about the study and before survey administration.