A systematic review of community-based interventions to address perinatal mental health

Abstract

Little is known about the scope and effectiveness of community-based interventions to address maternal perinatal mental health in the US. We searched PubMed, CINAHL, and PsychINFO in January 2024 to conduct a systematic review of studies using community-based interventions for maternal mental health from pregnancy to 1 year postpartum in the US. We reviewed 22 quantitative studies, and assessed methodological quality and effectiveness of interventions. Most were randomized trials (n=16) with strong or good methodological quality. The majority of the studies included racially and ethnically diverse participants (n=14), delivered interventions through community health workers, nurses, midwives, and doulas (n=18), and had mixed effectiveness of interventions (n=14). Limitations included small sample sizes, interventions not specifically developed for mental health, limited community involvement in designing interventions, and focus on participants with no mental health issues. Community partners augment this review with lived experience and recommendations for research and clinical practice.

1. INTRODUCTION

Community-based interventions and programs are designed to improve the health of individuals and families while leveraging the resources, knowledge, and networks available within the community.¹ These interventions are particularly important in health care, where tailored services are required to effectively serve specific patient populations with unique cultural and socioeconomic needs.² There is considerable variation in the definition and design of community-based interventions, including by setting (e.g., home-based, clinic-based), type of health care worker providing care (e.g., community health worker, nurse), and content (e.g., mobile health, group-based therapy).³ Globally, community-based interventions have played a critical role in health promotion for the perinatal population, given diverse informational, emotional, and tangible needs.⁴ Existing literature reviews indicate that community-based interventions have had mixed effects on perinatal health outcomes in the global context.⁵ In the United States (US), community-based interventions in the perinatal period range from home visiting programs by clinical professionals, such as nurses and doulas, to peer support programs with experienced birthing individuals.

A key component of perinatal health promotion is supporting mental health. Perinatal mental health conditions are a major contributor to pregnancy-related deaths in the US,⁶ and many of these deaths are preventable. Given that the US has the highest maternal mortality rates among all high-income nations,⁷ there is a need to identify best practices to optimize mental health in the perinatal period. Though community-based interventions represent a promising mechanism to improve perinatal mental health, it is unclear if they are effective in the US context in supporting the mental and psychological health of birthing individuals during pregnancy and into the first postpartum year. Therefore, the purpose of this review was to systematically review and synthesize evidence on community-based interventions to address mental and psychological health among birthing individuals in the perinatal period in the US. We aimed to (1) describe existing community-based interventions focused on mental and psychological health outcomes in the perinatal period, (2) examine the effectiveness of community-based interventions used to address mental and psychological outcomes in the perinatal period, and (3) describe the unpublished perspectives and interventions addressing perinatal mental health from local community leaders in New York City and their recommendations for future research.

2. METHODS

2.1. Design

We conducted a systematic review with a narrative synthesis. We followed published standards for systematic reviews from the Institute of Medicine,⁸ and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to report our results.⁹ A protocol was developed prior to conducting the review to ensure its rigor and reduce bias.

2.2. Search methods

We searched three electronic databases: PubMed, CINAHL, and PsycINFO. The search was developed and conducted by a PhD-prepared nurse researcher and PhD nursing student with clinical and research expertise in perinatal health research. We used a mix of keywords and subject headings representing perinatal populations, community-based interventions, and mental and psychological health outcomes. The search was conducted on January 22, 2024. Complete reproducible search strategies are detailed in supplementary materials.

2.3. Eligibility criteria

We included peer-reviewed, quantitative, data-based US studies of community-based interventions or programs to address mental and psychological health outcomes in the perinatal period, from pregnancy to 12 months postpartum. We included articles written in English and published from 2000 to 2024. Informed by the typologies of community-based interventions by McLeroy et al.,¹⁰ we operationally defined community-based interventions as education, groups, or programs delivered in the community (e.g., home, clinic, or community health centers). We included interventions or programs led and/or delivered by “community actors or agents of change” (i.e., doulas, midwives, nurses, and community health workers). Detailed eligibility criteria can be seen in Table 1.

Table 1.

Eligibility criteria

Inclusion criteria

Exclusion criteria

• Include articles published in and after 2000. • Include peer-reviewed articles published in English. • Include studies conducted in the United States. • Include data-based research studies. • Include studies with outcome data collection time points of up to 1 year postpartum. • Include quantitative studies with observational (descriptive), quasi-experimental, or experimental designs using primary or secondary data. • Include mixed methods studies that provide quantitative findings. • Include non-acute and non-clinical settings, community health centers, and other community settings. • Include studies with adolescents (< 18 years) but primarily focused on adults (> 50% of the sample are adults or the mean/median ≥ 18 years). • If multiple studies used the same data, include those that provide unique information to answer the research questions.

• Exclude studies published before 2000. • Exclude non-peer-reviewed articles written in a language other than English. • Exclude studies conducted in countries other than the United States. • Exclude articles without full text. • Exclude articles that are not data-based research (e.g., editorials, letters, commentary, dissertations, case studies/reports, conference abstracts, study protocols, guidelines, practice standards) or review articles. • Exclude interventions or programs without mental and psychological health outcomes. • Exclude outcome data collection time points beyond 1 year postpartum or unknown/unclear time points. • Exclude qualitative studies. • Exclude interventions provided in non-community settings only (e.g., acute care, hospital). • Exclude online community. • Exclude studies primarily focused on adolescents (< 18 years) or fathers. • Exclude animal studies.

2.4. Search selections

The search results were imported into Covidence, an online program to support systematic screening (Veritas Health Innovation, Melbourne, Australia). Two reviewers conducted title and abstract and full-text screening independently. Any disagreement between the two reviewers was resolved through discussion. Full-text studies that did not meet the eligibility criteria were excluded and reasons for exclusion were recorded. We made efforts to locate full-text articles through searches on multiple databases and interlibrary loans. After full-text screening, we manually searched the reference lists of included articles to further locate eligible studies.

2.5. Quality appraisal

We used the Standard Quality Assessment Criteria to evaluate the methodological quality of included studies.¹¹ This appraisal tool can be used to evaluate both experimental and quasi-experimental studies and was suitable for this review. This appraisal tool consists of 14 items that can be scored on a 3-point scale (0=no, 1=partial, 2=yes). A total score is calculated by dividing the total sum by the total possible sum. A higher total score indicates a higher methodological quality. We reported both total scores and their qualitative interpretations (i.e., limited, adequate, good, strong), a method that has been used in previous systematic reviews.¹² Two reviewers independently scored the included articles. Any disagreement was resolved through discussion. Detailed scoring results of all included studies can be seen in supplementary materials.

2.6. Data extraction and synthesis

Two researchers with experience in conducting systematic reviews extracted the key information into the evidence table. Two researchers cross-checked the extracted data for its accuracy and completeness. We followed the recommended standards for a qualitative narrative synthesis of evidence to provide both factual descriptions of studies and the patterns of clinical and methodological characteristics across the studies.⁸ We synthesized the main findings and reported results in terms of (a) study characteristics, (b) intervention description, and (c) effectiveness of intervention. We used the same language referring to participant race, ethnicity, and gender (e.g., American Indian) as they were defined in each of the reviewed studies. Of note, only one study used a gender-neutral term “pregnant people” to refer to their participants.¹³ All other studies either used mothers or women.

3. RESULTS

The search selection process is shown in Figure 1. Two reviewers conducted a title and abstract screening of 10,117 articles after duplicates were removed, which resulted in 182 articles that met the eligibility criteria. Of these, we could not locate full-text articles of five studies. Two reviewers conducted a full-text screening of 177 articles. Full-text screening resulted in 21 articles that met the eligibility criteria. We manually searched reference lists of these articles and reviewed 23 additional full-text articles, which resulted in one additional article that met the eligibility criteria. Therefore, we included a total of 22 articles in this review.

Figure 1.

PRISMA flow diagram

3.1. Description of studies

A detailed description of all included studies can be seen in Table 2 (N=22). Most studies were either randomized controlled trials (n=16)^14–29 or quasi-experimental (n=6),^13,30–34 and had either strong (n=6)^{14–17,28,34} or good (n=10)^{18,20,22–25,27,29,30,32} methodological quality. Studies were conducted in at least 16 different US states, excluding studies that did not mention specific geographic locations. Sample sizes varied from 39 to 1,229 participants. Recruitment occurred in clinical (e.g., obstetric clinics, community health centers) or sub-clinical settings (e.g., WIC offices) within the community for most studies (n=18).^{13–15,19,20,22–34}

Table 2.

Description of included studies (N=22)

Author (year)	Study design	Settings	Samples and sample size	Sample characteristics	Community involvement	Quality
Allen (2023)	RCT	Colorado, US	425 pregnant and postpartum women/mothers who are primarily under-resourced Intervention=235Control=190 Analysis sample=393	• 18-43 years old. • 73% Hispanic or Latina, 15% non-Hispanic White, 12% Black or African American. • 73% high school, 27% < high school, 10% college degree. • 58% unemployed, 18% temporary work, 13% part-time, 11% full-time. • 76% Medicaid, 72% used governmental financial assistance in the past month. • 47% partnered, 33% married, 13% no partners. • 40% first child, average gestational age 24 weeks at enrollment. • 33% history of depression, 26% screened positive for prenatal depression.	• Intervention (non-clinical setting) and recruitment in a community. • Recruitment in exam rooms at obstetrics and pediatrics clinics. Flyers and referrals from clinics were used. • Intervention delivered by other professionals (e.g., social worker, case manager).	Strong
Ammerman (2013)	RCT	Southwestern Ohio and northern Kentucky covering urban, suburban, and rural areas	93 low-income postpartum mothers with a diagnosis of major depressive disorders Intervention=47 Control=46	• Mean age 22 (SD=5, range=16-37). • 62% Caucasian, 32% African American, 93% Non-Hispanic. • 76% annual household income < $20,000. • 87% single or never married, 13% married. • 28% mild depression severity, 48% moderate, 24% severe. • 74% history of recurrent depression, mean number of episodes 3 (SD=3), mean onset age 15 (SD=5), 76% history of other psychiatric comorbidities.	• Intervention (home) and recruitment in a community. • Recruitment in clinics, hospitals, and social service agencies. • Intervention delivered by nurses, other professionals, and paraprofessionals.	Strong
Dodge (2019)	RCT	Durham, North Carolina	316 mothers Intervention=158Control=158	• Mean age 29 (SD=6), 7% adolescents. • 56% White, 38% Black, 29% Hispanic, 8% other. • 65% Medicaid or no insurance. • 28% Cesarean births, 5% multiple gestation births, 51% female infants.	• Intervention in a community (home). • The agency offering the intervention is a local nonprofit organization (“community own”), funded by the county government, Medicaid reimbursement, and philanthropy. • Intervention delivered by nurses.	Strong
Dodge (2014)	RCT	A mid-sized community with a high rate of poverty in Durham, North Carolina	531 mothers Intervention=260Control=271	• Mean age 28, 9% adolescents. • 39% non-Hispanic Black, 27% Non-Hispanic White, 25% Hispanic, 9% other. • 66% Medicaid or no insurance. • 32% Cesarean births, 3% multiple gestation births, 54% female infants. • 9% low birth weight, 6% gestation < 37 weeks, 6% any birth complications.	Same as Dodge (2019)	Strong
Duggan (2004)	RCT	Oahu, Hawaii	648 mothers from at-risk families for child abuse Intervention=378Control=270	Intervention group characteristics are reported. Only partner violence was statistically significantly different. • Mean age 24 (SD=6), 55% < 20 years old. • 34% Native Hawaiian, Pacific Islander, 28% Asian or Filipino, 10% Caucasian, 27% no primary ethnicity or unknown. • 52% worked in the year before birth. • 63% household income below poverty level. • 26% married, 29% living together, 35% friends or going together, 11% no partner. • 43% first birth. • 43% poor general mental health, 19% maternal substance use. • 43% partner violence (52% in control group).	• Intervention in a community (home). • Intervention is operated by three community-based organizations or agencies. • Intervention delivered by community paraprofessionals.	Good
Duncan (2023)	Quasi-experimental, non-randomized	San Francisco Bay Area, California	49 mostly low-income, primiparous and multiparous pregnant people Intervention=25 Control=24	• 18-35 years old. • 65% Latino(a)/Latine American/Hispanic, 10% non-Hispanic White/European American, 8% Indigenous/Native American/Alaskan Native/American Indian, 8% Black/African American, 8% multiracial. • 63% Spanish speaking, 65% born outside of the US, Mean years living in the US 7 (SD=5, range=1-22). • 96% heterosexual. • 43% 5-11 years of formal education. • 33% annual family income < $10,000, 29% $10,001-$30,000. • 92% Medicaid, 29% received EBT/food stamps. • 63% living in an apartment or house, 37% homelessness or unstable living situations.	• Intervention (non-clinical setting) and recruitment in a community. • Recruitment in safety net public hospital midwifery clinic in a community. • Intervention delivered by nurse midwives and other professionals. • A pilot study was conducted with healthcare providers in a community to inform curriculum and approach. • A partnership with the safety net public hospitals and community-based organizations.	Adequate
Gjerdingen (2013)	RCT	Minneapolis and St. Paul metropolitan area, Minnesota	39 women/mothers with depressive symptoms Doula group=12 Peer telephone group=13 Control=14	• Mean age 30 (SD=6). • 95% White. • 74% 4-year degree or higher. • 87% employed during pregnancy, 81% employed at 3 months, 78% employed at 6 months follow-up. • 66% family income $20,000-$79,999; 26% > $80,000. • 84% married. • 44% one child.	• Intervention (home and telephone) and recruitment in a community. • Recruitment in three community hospitals, local practices, websites, and early childhood family education programs. • Intervention delivered by doulas and peer supporters.	Adequate
Hans (2018)	RCT	High-poverty communities in a large city and two smaller urban areas in Illinois	312 young, low-income women/mothers Intervention=156Control=156	Intervention group characteristics are reported. Only co-habitation status was statistically significantly different. • Mean age 19 (SD=2). • 44% Black or African American, 39% Latina or Hispanic, 8% White, 9% multiracial or other. • Mean 11 years of education (SD=2). • 55% attend school, 20% employed. • 72% partnered, 25% cohabitating (31% cohabitating in control group). • 92% received public insurance, 84% WIC. • 97% first child, mean gestational weeks 26 (SD=6), 99% received prenatal care. • Mean CES-D score 14 (SD=9).	• Intervention (home) and recruitment in a community. • Recruitment in doula home visiting agencies, which use public health departments, WIC programs, clinics, and schools for recruitment. • Intervention delivered by doulas and home visitors. • Partnership with existing agencies offering doula home visiting services for high-poverty areas within Illinois communities.	Good
Heberlein (2016)	Quasi-experimental, non-randomized	Large prenatal care provider system in the southeastern US	218 pregnant women Intervention=117 Control=101	Intervention group characteristics are reported. Age and other children variables and timing of survey completion were statistically significantly different. • Mean age=24 (SD=5). • 48% Black, 52% White/other. • 24% less than high school, 66% high school diploma, 10% associate degree or higher. • 41% < $10,000 annual household income; 34% $10,000-$20,000; 25% > $20,000. • 18% married. • 61% no other children (37% in control group). • 25% trying to get pregnant. • 18% First feelings about pregnancy somewhat/very unhappy, 27% not sure, 55% somewhat/very happy. • Mean CES-D score=13 (SD=9), mean pregnancy-related distress 12 (SD=7), mean perceived stress 18 (SD=6).	• Intervention and recruitment in a prenatal clinic. • Intervention delivered by nurse midwives and nurse practitioners.	Good
Horowitz (2013)	RCT, repeated measures	Boston, Massachusetts	134 depressed postpartum women/mothers Intervention=66 Control=68 Analysis sample=125	• Mean age 31 (SD=6). • 54% Caucasian, 22% Hispanic, 12% African American, 8% Asian (8%), 5% other. • 73% English speaking, 10% Spanish, 18% other. • Mean years of education 16 (SD=3). • 24% employed. • Mean family income $80,132 (SD=$49,070). • 75% married, 24% single, 2% other, 83% cohabitating with a partner. • 56% first baby. • 49% history of depression, 1% history of substance abuse.	• Intervention (home) in a community. • Intervention delivered by nurses.	Adequate
Jesse (2015)	RCT, repeated measures	Southeastern region	146 rural minority, low-income women at risk for antepartum depression Intervention=72 Control=74	• Mean age 25 (SD=6). • 68% African American, 32% Caucasian, 9% English-speaking Hispanic, 3% non-English-speaking Hispanic. • 17% < high school, 28% high school or equivalent, 45% some college, 10% college graduate or above. • 38% employed. • 82% Medicaid, 11% no insurance, 5% Medicare, 1% private insurance, 95% WIC. • 58% live alone, 29% single living with a partner, 14% married living with a partner. • 34% primigravida, 76% multigravida, 43% nulliparous, 67% multiparous. • 27% history of depression. • Mean EPDS score 11 (SD=5), mean BDI-II score 17 (SD=9), 45% low-moderate risk for depression, 55% high risk based on EPDS.	• Intervention (clinic and telephone) and recruitment in a community. • Recruitment in local health department prenatal clinical and affiliated regional perinatal center. • Intervention delivered by community health workers (“resource mom”) and other professionals. • A focus group study with minority low-income women in the community was conducted before RCT to tailor the program.	Good
Kieffer (2013)	RCT	Community Health and Social Services Centers in southwest Detroit, Michigan	275 predominantly Mexican or Mexican American origin Latinas/Latino women with low income Intervention=138Control=137 Subgroup analysis sample of non-English-speaking women=218 (Intervention=117, Control=101)	Intervention group characteristics are reported. Only language was statistically significantly different. • 35% age 18-24 years, 30% 25-29 years, 35% ≥ 30 years. • 93% born in Mexico, 2% in the US, 5% other. • 85% don’t speak English (75% in control group). • 13% lived in the US < 2 years, 39% 2-5 years, 23% 6-9 years, 25% ≥ 10 years. • 31% ≤ 6 years of education, 38% 7-11 years, 30% ≥ 12 years. • 91% homemaker. • 61% married and living with a spouse. • 28% parity 0, 57% 1-2, 16% ≥ 3. • 67% received prenatal care at the Community Health and Social Services Center. • 40% at risk for depression based on CES-D.	• Intervention (home and likely local clinics) and recruitment in a community. • Recruitment in community partner organizations (e.g., WIC clinics). • Intervention delivered by community health workers. • Community-based participatory research design. Community stakeholders (e.g. women residents of childbearing age, community representatives) were interviewed and participated in focused group discussions to guide the study design and development.	Good
Lutenbacher (2018)	RCT	A large metropolitan area in Tennessee	188 pregnant Hispanic women Intervention=94 Control=94 Analysis sample=178	• Mean age 30 (SD=7). • 67% from Mexico, 16% Honduras, 10% El Salvador, 7% Guatemala, 1% Costa Rica, 1% Peru. • Median months in the US 108 (IQR=36-156). • 40% completed ≥ 8th-grade education, 40% 9-12th grade without a diploma, 19% high school diploma/GED. • 10% full-time employed, 16% part-time, 1% unemployed looking, 74% unemployed not looking. • 69% annual family income < $10,000; 28% < $15,000; 3% < $40,000. • 39% married, 4% separated/divorced/widowed, 57% never married. • Median number of children at home 2 (IQR=1-3), adults and children in home 4 (IQR=3-6).	• Intervention (home and community sites) and recruitment in a community. • Recruitment in clinics, markets, apartment complexes, and churches with high volumes of Hispanic population. • Intervention delivered by community health workers.	Good
McHale (2023)	RCT	Lower-income neighborhoods in an urban, southeastern US city	138 low-income, unmarried, mother and father dyads Intervention=70 Control=68	• 77% Black, 19% Caucasian, 4% mixed-race. • 86% employed at some point in the past year. Intervention group: • 13% junior high school, 47% high school, 6% GED, 10% college, 17% some college, 7 trade/vocational school. • 24% individual income < $5,000; 16% $5,000-$9,999; 11% $10,000-$14,999, 6% $15,000-$19,999; 3% $20,000-$24,999; 3% $25,000-$34,999; 6% $35,000-$49,999; 1% $50,000-$74.999; 16% don’t know; 14% no regular work. • 51% co-residential. • 20% history of mental illness, 1% drug/alcohol use, 50% intimate partner violence.	• Intervention (community sites) and recruitment in a community. • Recruitment in WIC offices and through flyers distributed in a community. • Intervention delivered by community mentors. • Followed a “community partnership paradigm” to guide the study. Community leaders collaborated to co-create the intervention • A proof-of-concept study was conducted with community leaders (e.g., Black elders, mentors, pastors) before RCT to inform curriculum development.	Good
McKee (2006)	RCT	South Bronx, New York City	187 low-income Black and Hispanic women/mothers Intervention=57 Usual care=43 Non-depressed comparator group=87	• Mean age 25 (SD=6). • 43% Black, 57% Hispanic. • 77% born in the US. • Mean 12 years of education (SD=2). • 27% working, 49% unemployed, 24% student/homemaker • 54% single, 19% married, 22% living with a partner, 3% separated, 2% divorced. • Mean gestational age 28 weeks at baseline (SD=3), at first prenatal visit 12 weeks (SD=5). • 44% nulliparous.	• Intervention (home, community health centers, and/or telephone) and recruitment in a community. • Recruitment in community health centers. • Intervention delivered by other professionals (social worker students).	Adequate
Mersky (2022)	RCT	Large, urban public health department in the Midwest US	237 low-income pregnant women Healthy Families America group=72 Enhanced prenatal care coordination group= 65 Control=100 Analysis sample=210	• Mean age 27 (SD=6). • 42% Hispanic, 45% non-Hispanic Black, 13% other (non-Hispanic White, American Indian, multi-racial, unknown). • 24% postsecondary education. • Mean EPDS baseline score 8 (SD=6). • Mean Social Support baseline score 4 (SD=1).	• Intervention (home) and recruitment in an urban health center. • Intervention delivered by paraprofessional family support workers and public health nurses.	Good
Mundorf (2018)	Quasi-experimental, non-randomized	Southeast regions in Louisiana with high African-American population	141 pregnant women at risk for postpartum depression Intervention=102 (remained enrolled through 6 months postpartum) Comparison=58	• Stated that demographic information around race and ethnicity, marital status, family income, age, region was collected at baseline. The results were not reported in the paper.	• Intervention (home and mobile devices) and recruitment in a community. • Recruitment in WIC clinics. • Intervention delivered by community health workers. • Community-based participatory research design. Project developed in response to recommendations from the Transdisciplinary Research Consortium for Gulf Resilience on Women’s Health.	Adequate
Posmontier (2016)	Quasi-experimental, non-randomized	US	61 postpartum women with postpartum depression Intervention=41 Control=20	Intervention group characteristics are reported. Only race was statistically significantly different. • Mean age 29 (SD=9). • 63% White, 17% Hispanic, 15% Black, 5% other (30% White, 5% Hispanic, 55% Black, 10% other in control group). • 5% completed 9-12th grade, 20% high school, 20% some college, 32% 4-year college, 20% graduate school, 5% other. • 61% unemployed. • 37% < annual income $50,000; 39% $50,000-100,000; 15% > $100,000; 10% declined to answer. • 78% live with a partner. • 42% primiparous • 27% complication in pregnancy, 39% during labor and birth. • 15% depression during pregnancy, 29% currently taking antidepressants. • 20% chronic medical illness. • 71% negative life stress.	• Recruitment in obstetric clinics. • Intervention delivered by nurse midwives via telephone. • A pilot study conducted with “eligible women” to inform non-randomized study design and nurse midwife-delivered intervention	Good
Tandon (2021)	RCT, cluster-randomized	7 Midwest states	874 low-income pregnant women MHP=310 HVP=405 Control=159 Analysis sample=824 (MHP=293, HVP=382, Control=149)	• Mean age 26 (SD=6). • 70% racial and/or ethnic minority. • 12% born outside the US, Median years in the US for non-natives 12 (IQR=7-15). • 13% Spanish intervention receipt. • 40% at least some college. • 63% unemployed, 20% employed part-time, 15% full-time. • 71% income < $25,000; 18% $25,000-$49,999; 4% $50,000-$74,999; 2% &75,000-$99,999; 1% ≥ $100,000. • 36% first-time mother, 64% unplanned pregnancy. • Median gestation weeks 23 (Range=4-39). • Median number of current children 1 (Range 0-9).	• Intervention (home, visiting program centers, community locations) and recruitment in a community. • Recruitment in home visiting programs and in the community. • Intervention delivered by mental health professionals and home visiting paraprofessionals.	Strong
Tandon (2022)	Quasi-experimental, non-randomized	Florida Healthy Start Home Visiting network covering 67 counties in Florida	1,229 low-income pregnant women Intervention=446 Control=571 Analysis sample=1,017	Intervention group characteristics are reported. • Mean age 28 (SD=6). • 30% African American, 53% White, 17% multi-race or other, 24% Latina ethnicity. • 39% employed full- or part-time. • 32% income < $30,000; 9% $30,001-$50,000; 5% $50,001-$74,999; 21% prefer not to answer. • 44% first-time mothers. • Mean weeks pregnant at enrollment 22 (SD=8).	• Intervention (home) and recruitment in a community. • Recruitment by community service providers. • Intervention delivered by community health workers.	Adequate
Upshur (2016)	Quasi-experimental, random selection of an intervention site	Two federally qualified community health centers in a Northeast city	149 pregnant women with posttraumatic stress disorder symptoms Intervention=89 Control=60	• Mean age 27 (SD=6). • 66% Hispanic, 14% non-Hispanic Black/African American, 13% non-Hispanic White, 8% other (Asian, Biracial, or other). • 54% born in the US or Puerto Rico. • 62% English language, 32% Spanish, 4% Portuguese, 2% other. • 15% currently dealing with immigration issues. • 42% < high school, 29% high school or GED, 24% some college or trade school, 5% college degree, 1% master’s degree. • 34% employed. • Mean monthly income $762 (SD=660). • 18% married, 4% separated, 3% divorced, 62% never married, 13% living with a partner. • Mean number of children 2 (SD=2), mean number of pregnancies 4 (SD=2), 19% primigravida. • Mean gestation weeks at first obstetrics visit 8 (SD=5). • 72% history of any birth problems in a prior pregnancy. • 68% any chronic health problems. • Mean PTSD severity score 22 (SD=12), mean depression score 12 (SD=6), mean positive coping= 3 (SD=1), mean negative coping=2 (SD=1).	• Intervention (prenatal clinic) and recruitment in a community. • Recruitment in community health centers. • Intervention delivered by paraprofessional prenatal advocates.	Strong
Walkup (2009)	RCT	Navajo and Apache reservations in New Mexico and Arizona	167 young, reservation-based American Indian women/mothers Intervention=81 Control=86	• 47% age 4-17, 53% age 18-22 (Range=14-22). • 100% American Indian, 65% Navajo, 18% White Mountain Apache, 17% mixed tribes. • 39% high school/GED/some college. • 12% currently employed. • 8% currently married. • 68% live with a partner, 72% live with parents. • 48% male partners enrolled in the program. • 10% parity ≥ 1. • 55% gestational age ≤ 20 weeks, 29% 21-28 weeks, 16% ≥ 29 weeks (Range=3-35).	• Intervention (home) and recruitment in a community. • Recruitment in prenatal and school-based community clinics. • Intervention delivered by paraprofessional American Indian women from the community. • Authors acknowledged Navajo and White Mountain Apache leaders and community stakeholders for their input to inform study design and the intervention	Good

Notes. Baseline characteristics of the whole sample are reported otherwise specified. All values were rounded to the nearest integer.

Abbreviations: BDI-II=Beck Depression Inventory; CES-D=Center for Epidemiologic Studies Depression Scale; EBT=Electronic Benefit Transfer; EPDS= Edinburgh Postnatal Depression Scale; GED=General Educational Development; IQR= Interquartile Range; PTSD=Post-traumatic Stress Disorder; RCT=Randomized Clinical Trials; SD=Standard Deviation; US=United States; WIC=Special Supplemental Nutrition Program for Women, Infants, and Children.

The majority of studies provided information about at least five or more demographic participant characteristics (n=16).^{13,14,18–26,28–30,32,34} Across studies, mean participant ages ranged from 19 to 31 years. Participant socio-economic characteristics were largely homogenous, as half of the studies focused enrollment on low-income participants (n=11).^{13–15,20,22,23,25–28,33} In contrast, three studies included participants with mostly higher education or income levels.^19,21,32). Two studies focused on young mothers (aged ≤ 25) with a mean participant age of 19.^20,29

Most studies enrolled racially and ethnically diverse participants, defined as >50% of study sample being from a historically minoritized group (n=14).^{13,14,17,18,20,22–29,34} Among these, four focused on specific racial and ethnic groups (e.g., Hispanic, American Indian).^23,24,26,29 The majority of studies (n=17) provided information about at least one or more perinatal health characteristics, such as parity, birth methods, gestational age at recruitment, and history of pregnancy complications.^{14,16–24,26,28–30,32–34} Seven studies provided information about at least three or more perinatal health characteristics .^{16,17,20,26,28,30,34} Among the studies that reported perinatal health characteristics, most studies included both first-time and experienced mothers (n=15).^{14,18–24,26,28–30,32–34}

3.2. Community-based interventions used in studies

We evaluated community involvement in the reviewed studies in terms of locations of intervention and participant recruitment, who administered the intervention, and study or intervention design (Table 2). Most studies conducted both recruitment and intervention in a community setting (n=17).^{13–15,19,20,22–31,33,34} Some studies included descriptions of preliminary studies or projects involving community members that informed the intervention study and/or design (n=5).^{13,22,23,25,32} Fewer studies specifically reported having a community partnership or using a community partnership paradigm (n=2)^13,25 and using a community-based participatory research design (n=2).^23,31

A detailed description of study interventions is presented in Table 3. Three studies had three-arm designs (2 interventions and 1 control).^19,27,28 Thus, a total of 25 interventions were provided across the 22 included studies. About one-third of interventions were delivered by paraprofessionals or community health workers (n=10).^{18,19,23–25,28,29,31,33,34} About two-thirds of interventions were delivered by professionals including nurses, certified nurse midwives, doulas, and/or nurse practitioners (n=7),^{13,16,17,19,21,27,30} as well as social workers and case managers (n=5).^{14,15,22,26,28} In three studies, a mix of paraprofessionals and professionals delivered interventions.^20,22,27

Table 3.

Community-based interventions and their effects on perinatal mental and psychological health (N=22)

Author (year)	Community-based intervention	Intervention focused on mental and psychological health?	Preventive intervention?	Outcome data collection time points	Outcome measures	Effectiveness of intervention on outcome
Home visiting only (n=8)
Ammerman (2013)	• Where: Home. • When: Postpartum. • Who: Regular home visiting provided by nurses, social workers, related professionals, and paraprofessionals. Cognitive-based therapy provided by social workers. • What: Home visiting focused on infant health development, maternal-infant relationships, maternal health, and linkage to community resources. Cognitive-based therapy entailed behavioral activation, identification of automatic thoughts, schemas, thought restructuring, and relapse prevention. It also focused on stress management, parenting challenges, and transition to adult roles for young mothers. • Intensity: A weekly 1-hour session for 15 sessions in total. An additional “booster session” 1-month post-intervention. • Control: Regular home visiting program without the cognitive-based therapy where mothers were permitted to obtain treatments.	Yes Cognitive-based therapy developed to reduce depression	No Developed for mothers with major depressive disorders (a positive screen on EPDS)	T1: 3 months postpartum (baseline) T2: 8 months postpartu m	Hamilton Depression Rating Scale35 EPDS36 BDI-II37 Structured Clinical Interview for DSM-IV Axis I Disorders38	• Effective: Yes. • Intervention group had significantly lower HDRS, EPDS, BDI-II scores than the control group. Intervention group were significantly less likely to receive major depression disorder diagnosis than the control group (29.3% vs. 69.8%). • Effect sizes range from 0.65 to 0.90 (medium to large). Those who completed the entire sessions had larger effect sizes than those who partially completed the sessions.
Dodge (2014)	• Where: Home. • When: Between 3-12 postpartum weeks. • Who: Nurse. • What: Nurses provided a brief intervention, assess long-term needs, and connect a family with community resources. An assessment of family risks in 12 domains and provision of brief education interventions occurs in the first visit. Education is organized as 20 “teaching moments”. Extended education is offered based on parental needs (e.g., lactation). The follow-up session ensured the community connection with care was continued and if other resources or problem solving is needed. • Intensity: 1-3 home visits with additional telephone or home follow-up 1 month later. • Control: Usual care.	No Developed to assess family needs and connect them with community resources	Yes Offered to all eligible births universally	6 months postpartum	EPDS,36 cut-off > 10 GAD,39 cut-off > 5 8-item CAGE and CAGE-AID questionnaire,40 substance use cut-off > 1	• Effective: Mixed. Effective in one outcome variable. • The percentage of possible maternal depression (7.69% vs. 11.81%), anxiety disorder (21.15% vs. 29.52%), and substance use problems (4.63% vs. 6.27%) were low in intervention groups. But the differences were only significant for the anxiety disorder.
Duggan (2004)	• Where: Home. • When: Healthy Start Programs are provided up to 3-5 postpartum years. • Who: Paraprofessionals • What: A long-term home visiting program for at-risk families. Trained paraprofessionals built trusting relationships with families and provided help to address risks and crisis (e.g., domestic violence, substance use, mental health). They also provided access for services (e.g.., income, nutrition) and education around parenting. • Intensity: Level 1 program required weekly visits, level 2 biweekly, level 3 monthly, and level 4 quarterly. Healthy family functioning and achievement of milestones were monitored to be promoted in levels. • Control: No Healthy Start Program.	No Developed to prevent child abuse and neglect for at-risk families	Yes Offered as a preventive measure for adverse child health outcomes	T1: At birth (baseline) T2: 1 year postpartum	CES-D,41 cut-off ≥ 24 Mental Health Index,42 cut-off < 67 Parenting Stress Index43 CAGE questions,44 substance use cut-off ≥ 2	• Effective: Mixed. Effective in one outcome variable and in subgroup analysis. • The percentage of depression (23% for both groups), severe parenting stress (11% for intervention, 10% for control), illicit drug use (9% for both), problem with alcohol use (8% for intervention, 10% for control) were not significantly different. • Intervention group had significantly lower percentage of poor general mental health (36% for intervention, 46% control), adjusted odds ratio of 0.52, 95% CI=0.33-0.81 in agency b, but not in agencies a or c. • In sub analysis of including high dose service families only, problem with alcohol use was significantly smaller in intervention group (AOR 0.41, 95% CI=0.18-0.94).
Hans (2018)	• Where: Home. • When: Prenatally to 6 weeks postpartum. • Who: Doula and home visitors (known as family support worker or parent educator). • What: Mothers were visited by a doula, home visitor, or both. Doula support was provided more intensively during pregnancy and first weeks postpartum. Home visitors were the primary support by 6 weeks postpartum. Doulas focused on issues around pregnancy, health, childbirth, breastfeeding, newborn care, and early bonding. Doulas attended births at hospital and sometimes accompanied to prenatal and postpartum clinic visits. Home visitors focused on mother-infant relationship, infant development, safety, and screening of family basic needs. • Intensity: Weekly. • Control: Case management services including a minimum of two meetings (one each prenatally and postpartum) at home, agency offices, or via telephone.	No Developed to improve general maternal child health outcomes	Yes Offered to all eligible women universally	T1: 37 gestational weeks T2: 3 weeks postpartum T3: 3 months postpartum	CES-D,41 cut-off ≥ 16	• Effective: No. • The percentage of high depressive symptoms did not significantly different by groups at all 3 time points. • The percentage was lower at 3 months in the intervention group (13% vs 15.1% in control), but not significantly lower.
Horowitz (2013)	• Where: Home. • When: 6 weeks to 9 months postpartum. • Who: Nurse. • What: Communicating And Relating Effectively is relationship focused behavioral intervention. Nurses offered coaching and education on infant cues, how to response to those cues sensitively, effective communication, and changing negative behaviors (e.g., withdrawal, intrusive or irritable interactions). • Intensity: 1-hour visit at 6 weeks and 3, 6, and 9 months postpartum. Intervention group received 30-to-40-minute visits at 2 and 4 months postpartum. • Control: Home visiting services without the two additional visits.	No Developed to improve responsive interaction between mothers and infants	No Developed for mothers with depression (a positive screen on EPDS)	T1: 6 weeks postpartum (baseline) T2: 3 months postpartum T3: 6 months postpartum T4: 9 months postpartum	EPDS36 Postpartum Depression Screening Scale45	• Effective: Mixed. Effective in lowering scores but not in comparison with control. • There were no significant differences in EPDS and PDSS scores at each time point cross sectionally between intervention and control groups. • There were no significant differences in PDSS scores over time between intervention and control groups. • In both groups PDSS scores significantly decreased over time.
Mersky (2022)	Healthy Families America • Where: Home. • When: 2nd or 3rd trimester up to 3 year postpartum. • Who: Paraprofessional (as opposed to professionals such as nurses and social workers) family support worker and a public health nurse. • What: Services included screening and assessment, family goal planning, parenting guidance, social support, and referrals to community-based resources. Typically led by family support workers with auxiliary support from nurses. • Intensity: 1-hour weekly visit up to 6 months postpartum at minimum. Enhanced Prenatal Care Coordination •Where: Home. • When: Prenatally to 60 days postpartum. • Who: Public health nurse. • What: Services included screening and assessment, health and nutrition education, and care coordination. • Intensity: Bi-weekly with tailored frequency based on individual needs. • Control: No Healthy Family America or Enhanced Prenatal Care Coordination.	No Healthy Family American program developed to prevent child abuse and neglect. Prenatal Care Coordination program developed to promote positive birth outcomes and infant development	Yes Both programs offered to all eligible families and women	T1: prenatally (baseline) T2: 14-60 days postpartum T3: 6 months postpartum T4: 12 months postpartum	Parenting Stress Index-Short Form43	• Effective: Mixed. Effective in one time point and in moderation analysis. • Parental distress scores show a significant increase over time in both groups. Healthy Families America group had significantly lower parental distress at time 2 compared to time 4. Prenatal Care Coordination group had lower distress at time 2 compared to time 3 or time 4. • At times 2 and 3, Prenatal Care Coordination group had lower parental distress scores than Healthy Families America or control groups, but only time 2 was significantly lower. • Prenatal Care Coordination intervention had a significantly greater impact on participants with higher baseline depression scores.
Tandon (2022)	• Where: Home. • When: Unclear, likely prenatal to postpartum. • Who: Lay home visitors (community health workers). • What: Regular Healthy Start home visiting services included prenatal and parenting education, inter-conception education, screening for psychosocial risks, risk appropriate referrals, care coordination, and anticipatory guidance for new parents. Intervention group received the additional Mothers and Babies Courses, a cognitive behavioral intervention that provided education to increase thoughts and behaviors that can lead to positive mood, to manage stress, and to reduce depressive symptoms. • Intensity: Mothers and Babies Course included 15-20 minute, weekly or bi-weekly for 12 sessions in total. Entire intervention lasted 3-6 months. • Control: Regular Healthy Start home visiting program without the Mothers and Babies Course education intervention.	Yes Mothers and Babies Courses developed to promote healthy mood management	No Developed for women with mild-to-moderate depressive symptoms (a positive screening on EPDS) or at risk for postpartum depression based on home visitor’s clinical assessment	T1: Mean 22 gestational weeks for intervention group (SD=8, baseline) T2: 6 months post-intervention	BDI-II37 PSS, 4-item version46 Negative Mood Regulation Scale47 Behavioral Activation Depression Scale48	• Effective: Mixed. Effective in as-treated analysis and in subgroup analysis. • Intervention group had significantly lower BDI-II scores than control (1.1 point difference) in as-treated analysis. Not significant in intent-to-treat analysis. • In the intervention group, “low dosage” (attended 1-5 sessions) participants had a non-statistically significant reduction in BDI-II scores. “High dosage” (attended 6-11 sessions) and “full dosage” (attended all 12 sessions) had significant reductions in BDI-II scores by 2.1 points and 2.4 points respectively. • Intervention group had significantly lower PSS scores in both intent-to-treat (0.4 points) and as-treated analysis (0.6 points) than control. • Significantly greater reduction in PSS scores if had high or full dosage intervention. • No significant differences in NMRS or BADS.
Walkup (2009)	• Where: Home. • When: 28 gestational weeks to 6 months postpartum. • Who: Paraprofessional American Indian women. • What: Family Spirit Intervention provided prenatal and infant care education, family planning, substance use prevention, problem-solving, and coping skills and strategies. The curriculum was developed to reflect local native practices. • Intensity: 25 visits, 1-hour per visit. • Control: Paraprofessional home visiting for breastfeeding and nutrition support and education (23 visits).	No Developed to improve parenting knowledge and involvement	Yes Offered to all eligible women	T1: 28 gestational weeks (baseline) T2: 2 months postpartum T3: 6 months postpartum T4: 12 months postpartum	CES-D41 Substance Use49 Parenting Stress Index43	• Effective: No. • No significant differences between groups for reduction in depressive symptoms, parenting stress, or substance use.
Home visiting combined with additional locations/modalities (n=4)
Dodge (2019)	• Where: Birthing hospital and home. • When: Postpartum. • Who: Nurse. • What: Nurses assessed the family for 12 key domains (parent and infant health, medical home, child care planning, parent-infant relationship, infant crying management, material supports, family violence, past experience of maltreatment, maternal depression and anxiety, parental substance use, social support) on a 4-point scale. Nurses provided intervention based on the scores (1=no risk, 2=brief intervention, 3=connect to community resource, 4=emergency requiring crisis intervention). • Intensity: Initial intervention at birthing hospital followed by 1-3 home visits. • Control: Usual care.	No Developed to assess family needs and connect them with community resources	Yes Offered to all eligible births universally	6 months postpartum (Range=4-8)	EPDS36, cut-off > 10 GAD39, cut-off > 5	• Effective: No • Cases of possible maternal depression or anxiety were lower in intervention group (18.2% vs. 25.9% in control), but it was not statistically significant.
Kieffer (2013)	• Where: Home and likely local clinics. • When: Prenatally to 6 weeks postpartum. • Who: Community health workers. • What: Curriculum offered information, discussion, and activities about knowledge and skills to reduce social and environmental barriers of healthy eating and exercise. Led by Spanish-speaking Latina community health workers, known as women’s health advocates, who provided education, social support, facilitated group meetings, and performed home visits. Offered childcare, transportation. • Intensity: A weekly,14- session curriculum with 2 home visits and 9 group meetings prenatally. 2 additional home visits and one group meeting 2-6 weeks postpartum. • Control: Standard educational material about eating and exercise as well as four group meetings (3 prenatally and another at 6 weeks postpartum). Delivered by trained staff from a community mental health agency.	No Developed to reduce risk factors for obesity and type 2 diabetes by adopting healthy lifestyles	Yes Offered to all eligible women	T1: Mean 17 gestational weeks (baseline) T2: Mean 28 gestational weeks T3: Mean 8 postpartum weeks	CES-D, 11-item abbreviated version,50 cut-off ≥ 16	• Effective: Mixed. Effective in one time point. • Before and after adjusting for covariates, a significant decrease in CES-D scores from time 1 to time 2 (13.2 to 11.2). Not significant decrease from time 2 to time 3 (11.2 to 10.5). Overall significant decrease from time 1 to time 3. • Significantly more decrease in intervention group compared to control group from time 1 to time 2 and time 1 to time 3. After adjusting for covariate, CES-D mean score was only significantly different from time 1 to time 3 between groups. • For categorical CES-D, percentage of women at risk for depression at time 2 was significantly less for intervention group (18-19% vs 34-36% in control Cohen’s h=0.34-0.42), adjusting for baseline and covariates. This was not significant at time 3. • No significant decrease in CES-D scores among those who did not speak English from T1 to T3 in intervention group only or in comparison with control.
Lutenbacher (2018)	• Where: Home and community sites. • When: Prenatally to 6 months postpartum. • Who: Community health workers recruited within the local Hispanic community. • What: Monthly home visits and periodic group meetings for listening maternal concerns, educating about pregnancy, developmental milestones, healthy eating, and breastfeeding, and referral to medical social resources. Materials were available in both Spanish and English. • Intensity: 1-hour monthly home visits. • Control: Minimal educational intervention with printed materials only.	No Developed to improve general maternal child health outcomes	Yes Offered to all eligible women	T1: ≤ 26 gestational weeks (baseline) T2: 35 weeks gestational weeks. T3: 2 weeks postpartum T4: 2 months postpartum T5: 6 months postpartum	EPDS36 Parenting Stress Index-Short Form,43 only assessed postpartum	• Effective: Yes. • Intervention group had significantly greater decrease in EPDS scores from baseline to postpartum. • Median scores at each time point for control group 7, 4, 5, 3, and 0. For intervention group 7 at baseline and 0 at all prenatal and postpartum follow ups (Cohen’s d=0.57). • Intervention group had significantly lower PSI scores in the postpartum period. Median scores at each time point for control group 76, 76, and 77. For intervention group 74 at all time points (Cohen’s d=0.43).
Mundorf (2018)	•Where: Home and mobile devices. • When: First trimester to 6 months postpartum. • Who: Community health workers trained and certified in public health or reproductive health. • What: Interaction between community health workers and participants via home visits and mobile devices (texts and phone calls) to build supportive relationships. • Intensity: Unclear, “throughout the study period.” • Control: Comparison group data obtained from another project in the same region with similar populations.	No Developed to improve general maternal child health outcomes	Yes Offered to all eligible women	T1: 1-13 gestational weeks T2: 14-27 gestational weeks T3: 28-40 gestational weeks T4: 6 weeks postpartum T5: 6 months postpartum	EPDS36	• Effective: Mixed. Effective in unadjusted analysis and subgroup analysis. • Intervention group had significantly lower EPDS scores at time 5 (M=4.2, SD=3.7) than the comparison group (M=6.3, SD=6.2). Not significant after adjusting for income level and relationship status. • In three separate subgroup analysis of women who were single, non-Hispanic White, or rural analysis, the intervention group had significantly lower scores than comparison group.
Primary care or community health services (n=2)
Heberlein (2016)	• Where: Prenatal clinic. • When: Prenatally. • Who: Certified nurse midwives and nurse practitioners. • What: CenteringPregnancy is a group prenatal care of 8-12 pregnant women with due dates in the same month. Group prenatal care participants received individual care for a few minutes at the beginning of the appointment (~3 min each) followed by 90 min of group discussion regarding nutrition, exercise, relaxation techniques, infant care, self-esteem, healthy relationships, preparation for birth, etc. Women were allowed to bring partners to sessions. • Intensity: 10 sessions occurring monthly until 28 weeks, every 2-3 weeks until 36 weeks, then weekly until birth. 90 minutes to 2 hours per session. • Control: Individual prenatal care occurring at same intervals. 15-20 minutes per visit.	No Developed to address shortcomings of traditional prenatal care for better birth outcomes	Yes Offered to all eligible women	T1: Mean 13 gestational week (SD=2) T2: Mean 33 gestational weeks (SD=1) T3: Mean 7 weeks postpartum (SD=3)	Prenatal Distress Questionnaire51 PSS46 CES-D41 Positive and Negative Affect Schedule52	• Effective: Mixed. Effective in moderation analysis. • No significant differences in pregnancy-related distress, perceived stress, positive or negative affect, or depressive symptoms. • Among those with inadequate social support, intervention group had significantly greater decrease in pregnancy-related distress (2.9-point difference) than those in control group at T2. • Among women with high pregnancy-related distress at T1, intervention group had significantly greater decrease in depressive symptoms (4.9-point difference) than those in control group at T3.
Upshur (2016)	• Where: At the federally qualified health center prenatal clinic. • When: Prenatally. • Who: Paraprofessional prenatal advocates who were bilingual and bicultural. Had limited knowledge of mental health education. • What: Provided education on eight topics (introduction, safety, taking care of yourself, taking back your power, grounding, setting boundaries, respecting your time, healing from anger) adapted from “Seeking Safety”, which is a psychosocial education program based on cognitive-behavioral therapy to improve coping skills for people with posttraumatic stress and substance use disorders. Offered both in English and Spanish. • Intensity: Eight sessions delivered from prenatally to 1 month before the due date. • Control: Usual care entailing routine prenatal care, prenatal advocate services, and referral to mental health services for women with positive screenings.	Yes Developed to improve posttraumatic stress disorder during pregnancy.	No Developed for women with high risk for posttraumatic stress disorders (a positive screening on Primary Care PTSD Screen)	T1: prenatally (baseline) T2: prenatally (immediately after completion of the last session)	Posttraumatic Stress Scale53	• Effective: No. • No significant effect of the intervention on reduction of PTSD symptom severity. Cohen’s d=0.21, small.
Non-clinical settings (n=2)
Allen (2023)	• Where: A designated workshop space in the community. • When: Prenatally. • Who: Group facilitator with experience in case management and in case management and relationship skills and relevant education (e.g., social work or psychology) and a case manager. Facilitators and case managers often shared similar lived experiences. • What: A group-based workshop with evidence-based curriculum on partner communication, problem solving, conflict management, co-parenting, infant care and bonding, self-care, and postpartum depression. Offered meal, childcare, and transportation for workshops. Offered both English and Spanish. Each participant is assigned to a case manager who facilitates applying skills in real life and to connect with community resources (food assistance, housing, employment). • Intensity: 4 hours per workshop for 6 weeks. • Control: Usual care.	No Developed to provide education on infant care and parenting and to offer case management services	Yes Offered to all eligible women	Mean 7 weeks postpartum (SD=2, range=2-12)	EPDS, 36 cut-off ≥ 10	• Effective: Mixed. Effective in moderation analysis. • Intervention group less likely to screen positive for postpartum depression (16%) than control group (21%), but not significantly different. • Intervention had significant effect only for those who identified Black or African American. In intervention group 15% of Black or African American screened positive, compared to 39% in control group. Cox index effect size=1.53. • Intervention had a significant effect for those who did not have a history of depression. In intervention group 10% had history of depression, compared to 19% in control group. Effect size=0.51.
Duncan (2023)	• Where: A space from a community-based organization, near the public safety net hospital. • When: Prenatally. • Who: Clinicians (Certified nurse midwives or obstetrician) and group facilitators (e.g., medical assistant, social workers). The mindfulness component was delivered by certified nurse midwives only. • What: CenteringPregnancy is a group-based prenatal program with a provider and group facilitator (e.g., medical assistant, social workers). CenteringPregnancy + provided additional mindfulness skills training, adapted from Mindfulness-based childbirth and parenting program (e.g., breathing, yoga, meditation, emotions). Offered both in English and Spanish. Transportation vouchers and snacks were offered for each visit. • Intensity: 10 sessions, 3 hours per session for 9 weeks. • Control: CenteringPregnancy without the mindfulness program.	Yes Mindfulness component developed to reduce stress and promote mental health and well-being	Yes Offered to all eligible women	T1: Mean 20 gestational weeks (SD=4, range 14-27, baseline) T2: Mean 32 gestational weeks (SD=3, range 28-39) T3: Mean 13 postpartum weeks (SD=5, range 6-23)	EPDS36 STAI54	• Effective: Yes. • Intervention group had significantly lower EPDS scores adjusting for baseline CES-D scores (intervention M=4.9, SD=4.1; control M=8.7, SD=5.4). Cohen’s d=0.8, large effect. • Intervention group had significantly lower STAI scores adjusting for baseline anxiety, family income (intervention M=35.8, SD=7.8; control M=40.8, SD=9.1). Cohen’s d=0.6, moderate effect.
Multiple community locations and modalities (n=4)
Jesse (2015)	• Where: Prenatal clinic and telephone. • When: Prenatally. • Who: Social workers, mental health professionals, counselors, and community health workers (“resource mom” who were Black). • What: Cognitive-behavioral intervention included group meetings of 2-6 women around topics such as depressive symptoms in pregnancy, stress reduction, relationship/domestic violence. Meetings were facilitated by professionals and resource moms. Resource moms also offered weekly booster session via phone, and provided case management services. Education materials were written at a fourth grade reading level, offered in English and Spanish, and culturally tailored (e.g., section addressing depression for women of color). Transportation, childcare, and snacks were provided for attendance. For technology enhancement, each participant was given MP3 players with preprogramed list of weekly homework assignments, stress reducing guided visualization, review of thoughts/feelings/beh aviors, positive affirmations, and motivational and inspiring music. • Intensity: A weekly 2-hour session for 6 weeks. • Control: Usual care entailing routine social services, prenatal care, and case management.	Yes Developed to reduce risks for antepartum depression.	No Developed for women with low and high risk for antepartum depression (a positive screen on EPDS)	T1: 6-30 gestational weeks (baseline) T2: 6-weeks from T1 T3: 1-month from T2	EPDS,36 cut-off ≥ 10 high-risk, 4-9 low-moderate risk BDI-II37	• Effective: Mixed. Effective in subgroup analysis. • For the low-moderate risk group, the change in BDI-II scores from T1 to T2 and T1 to T3 were significantly greater in intervention group (significant improvemen t in scores in intervention group mean change of 4.9, SE=1.5 vs. control 0.6, SE=0.8). • The changes were also greater in the high risk group, but not significantly different. • Among American women only subgroup analysis, the change in BDI-II scores from T1 to T2 was significantly greater for the intervention group in the low-moderate risk group (mean change of 5.2. SE=1.2 vs, control 0.7. SE=1.1), but not in high risk group. • The changes in mean EPDS scores were statistically significantly greater from T1 to T2 in high risk group (mean change of 5.6, SE=0.9 vs. control 2.2, SE=1).
McHale (2023)	• Where: Community sites of the participants’ choosing (e.g., participants’ houses, neighborhood centers, parks, churches). • When: Prenatally. • Who: Gender-concordant, community mentors (e.g., pastors, respected community members). • What: Provided interactive, skills-based educational sessions focused on: increasing consciousness about co-parenting, informational and skill-building competencies to support co-parenting, and guided enactments. Community mentors support parents to explore their parenting philosophies and respond to actual or potential barriers to co-parenting. All sessions maintain a strict focus on the well-being of the child. • Intensity: 6 sessions. • Control: Usual care entailing navigational support to access existing community resources.	No Developed to improve co-parenting and child health outcomes	Yes Offered to all eligible women	T1: 3rd trimester (baseline) T2: 3 months postpartum T3: 12 months postpartum	EPDS36	• Effective: Mixed. Effective in lowering scores but not in comparison with control. • Mothers in both the intervention and control group reported significantly lower depression at T2 and T3, compared to baseline. • Depression among mothers decreased over time in both study groups from T1 to T3. • EPDS scores were lower for mothers in the intervention compared to control, but not significantly lower.
McKee (2006)	• Where: Home or community health centers of participants’ choosing and telephone. • When: Third trimester to early postpartum. • Who: Primarily graduate social worker students. • What: Three components included (1) an eight-session cognitive behavioral depression prevention course around relaxation technique, cognitive reframing, and goal setting; (2) four modules detailing infant development and “responsive mothering”; and (3) social support building telephone sessions, which provided opportunities for mothers to discuss needs and identify formal and informal supports. Participants had individual sessions that include one to three intervention components. • Intensity: The social support building telephone sessions occurred at least twice monthly. • Control: Usual care with psychoeducational workshops.	Yes Developed to reduce depression	Yes Offered to all eligible women	T1: 3rd trimester (baseline) T2: 3 months postpartum	BDI-II37	• Effective: Mixed. Effective in lowering scores but not in comparison with control. • All randomized groups and the comparator group had significantly reduced depressive symptoms at T2 compared to T1. Change in scores of −8.4 for intervention, and −9.8 for control, −2.5 for non-depressed comparison group. • There were no significant differences in depression improvement for participants in the intervention compared to control group.
Tandon (2021)	• Where: Home visiting program center or other community location. • When: Unclear, likely prenatal to postpartum. • Who: Mental health professionals or home visiting paraprofessionals. • What: Mothers and Babies is a cognitive-behavioral, group-based session taught to participants to recognize negative thoughts and learn strategies for interpersonal interactions. Session topics included engaging pleasant activities, strategies to overcome obstacles, and identifying negative and positive thought patterns. Two intervention groups were provided with Mothers and Babies by mental health professionals or home visiting paraprofessionals respectively. • Intensity: 6, weekly sessions for Mothers and Babies. • Control: Usual home visiting program focused on preparing childbirth, infant care, and social support without Mothers and Babies education.	Yes Mothers and Babies Courses developed to promote healthy mood management	Yes Offered to all eligible women	T1: Median 23 gestational weeks (Range 4-29, baseline) T2: 1 week post-intervention or 8 weeks from T1 for control group T3: 12 weeks postpartum T4: 24 weeks postpartum	Quick Inventory of Depressive Symptomolo gy-Self Report55 Maternal Mood Screener,56 cut-off ≥ 5 Behavioral Activation Depression Scale48 Negative Mood Regulation Scale47	• Effective: Mixed. Effective in subgroup analysis at one time point. • All study groups showed an overall mean decrease in QIDS scores (no statistical test). • Intervention groups did not have significantly greater reduction in depression compared to control. • In mild depressive symptom subgroup, QIDS scores at T4 significantly different for the intervention delivered by mental health professionals then control group, but not significant between home visiting paraprofessionals and control group. • All other outcome variables not significant.
	Others (n=2; telephone only, home visiting vs. telephone)
Gjerdingen (2013)	Postpartum Doula Group • Where: Home. • When: Up to 6 weeks postpartum. • Who: Doula • What: Education on infant care/feeding, practical support (e.g., help with childcare, light housekeeping, assistance with errands), and emotional support provided. • Intensity: A total of 24 hours of doula services (about 4 hours per week). Peer Telephone Group • Where: Telephone. • When: Up to 3 months postpartum. • Who: Peer volunteers with experience of postpartum depression and recovery. • What: Educational, emotional, and comparison support (by sharing similar experience) provided. Calls were usually initiated by the volunteers. • Intensity: Average of 6 calls, totaling a 1-hour long conversation per mother. • Control: Two intervention and control groups received usual care entailing mailed postpartum depression brochure and resource list (e.g., contact information for support groups, classes, providers).	Yes Developed to improve postpartum depressive symptoms	No Developed for mothers with depressive symptoms (a positive screening on PHQ-9).	T1: 0-6 months postpartum (baseline) T2: 6-12 months postpartum	PHQ-9,57 cut-off >10 CES-D41	• Effective: No. • CES-D scores at T1 and T2 were not significantly different by study groups. • The changes of CES-D scores had the greater improvement in doula group at T2 (−16.3 vs. −14.5 for telephone support group and −10.1 for control group), but not significantly greater.
Posmontier (2016)	• Where: Telephone. • When: Postpartum. • Who: Certified nurse midwives. • What: Provided interpersonal psychotherapy. The first few sessions established the depression diagnosis and discussed causes and social context of depression. Middle sessions supported women to learn how to communicate their needs and resolve interpersonal conflicts. The final sessions reinforced competencies and terminated the therapeutic relationship. • Intensity: 8 sessions over 12 weeks. Up to 50 minutes per session. • Control: Usual care entailing referral to mental health professionals.	Yes Developed to reduce postpartum depression symptoms	No Developed for women with postpartum depression (a positive screen on EPDS)	T1: 6 weeks-6 months postpartum (baseline) T2: 4 weeks from T1 T3: 8 weeks from T1 T4: 12 weeks from T1	EPDS36 Hamilton Depression Rating Scale35	• Effective: Mixed. Effective in one outcome variable. • The Hamilton Depression Scale scores at T3 and T4, were significantly lower in intervention group than control group. T3 M=7.9, SD=1.2 in intervention vs. T3 M=12.3, SD=1.7 in control. T4 M=7.5, SD=1.3 in intervention vs. T4 M=12.4, SD=1.7 in control. • No significant differences found for EPDS scores between the two groups.

Notes. Outcome assessment time points were rounded to the nearest integer. Preventative intervention was operationally defined as interventions offered to all participants without eligibility criterion in mental and psychological health outcomes (i.e., intervention only offered to those with positive screenings). Abbreviation: BDI-II=Beck Depression Inventory-II, GAD=Brief Generalized Anxiety Disorder-7, CES-D=Center for Epidemiological Studies Depression Scale, EPDS=Edinburgh Postnatal Depression Scale, PHQ-9=Patient Health Questionnaire-9, M=Mean, PSS=Perceived Stress Scale, QIDS=Quick Inventory of Depressive Symptomology-Self Report, SAMHSA=Substance Abuse Mental Health Services Administration, SD=Standard Deviation, SE=Standard Error, STAI=State-Trait Anxiety Inventory, T=Time.

More than half of the studies used home visiting programs either alone or combined with additional components such as group meetings (n=12).^{15–18,20,21,23,24,27,29,31,33} Four studies offered interventions at multiple community locations, often based on participants’ preferences, or used multiple modalities (e.g., combining clinic-delivered and telephone).^22,25,26,28 Two studies offered interventions in either non-clinical settings^13,14 or primary care settings^30,34, respectively. Interventions were offered during the prenatal period (n=6),^{13,14,22,25,30,34} postpartum period (n=7),^{15–19,21,32} or from prenatal to postpartum (n=7).^{20,23,24,26,27,29,31} Two studies had an unclear intervention time frame.^28,33

Fewer than half of the interventions were designed specifically to improve mental and psychological health outcomes (n=9).^{13,15,19,22,26,28,32–34} Most of the mental and psychological health outcomes were examined as a result of an intervention developed to improve or promote general maternal and child health outcomes (e.g., parenting efficacy, mother-infant bonding). Preventive intervention was operationally defined as interventions offered to all participants without eligibility criteria around mental and psychological health outcomes. Most studies were preventive and offered to all eligible women, mothers or families regardless of current mental or psychological health issues, whereas in about one-third of the studies (n=7), the intervention was only offered to those with positive screenings on mental and psychological health outcome measures.^{15,19,21,22,32–34}

Almost all studies examined depression outcomes (n=20)^{13–26,28–33} followed by stress outcomes including parenting stress, perceived stress, posttraumatic stress, and general distress (n=7).^{18,24,27,29,30,33,34} Anxiety^13,16,17 and substance or alcohol use^17,18,29 outcomes were examined in three studies respectively. Fewer studies also examined general mental health (n=1),¹⁸ affect (n=1),³⁰ or mood (n=2).^28,33 Among the studies that studied depression outcomes, the Edinburgh Postnatal Depression Scale was the most frequently used measure (n=11),^{13–17,21,22,24,25,31,32} followed by the Center for Epidemiological Studies Depression Scale (n=6),^{18–20,23,29,30} and Beck Depression Inventory-II (n=4).^15,22,26,33 Among studies that examined stress outcomes, the Parenting Stress Index was most frequently used instrument (n=4).^18,24,27,29

3.3. Effectiveness of community-based interventions

We also examined effectiveness of interventions on mental and psychological health outcomes, based on statistical significance comparisons reported in each study, using p < 0.05 as the cutoff for our determination (Table 3). More than half of studies (n=14) showed mixed effects,^{14,17,18,21–23,25–28,30–33} defined as effectiveness on some of the outcome variables, specific time points only, or only for certain participant subgroups. In about a quarter of studies (n=5), the intervention was not effective in any of the outcome variables measured.^{16,19,20,29,34} Three of the five studies with non-significant intervention effects did not specifically focus on mental and psychological health outcomes.^16,20,29 Although the other two studies had interventions focused on mental and psychological health outcomes, one had a very small sample (n=39)¹⁹ and the other had low intervention uptake, where about 57% of participants received all intervention sessions.³⁴ Among the studies where at least some of interventions were effective (n=17), 10 studies were focused on individuals with low income^{13–15,22,23,25–28,33} and 11 studies included racially and ethnically diverse participants.^{13,14,17,18,22–28}

3.4. Community-based interventions from community partners in New York City

Published research studies often do not include community perspectives and may not represent the voices of the people who are leading and carrying out this work on a day-to-day basis. For example, in New York City, Black-led grassroots and community-based organizations are making substantial contributions towards improving the mental health and well-being of birthing individuals and their families. They adopt a holistic approach to care, offering various services that support and empower birthing individuals of color throughout the perinatal period. Therefore, we sought insights from community partners at two organizations based in New York City, the Caribbean Women’s Health Association and Restore Forward (Black Women’s Blueprint) to augment this review. These human rights organizations provide community-based, culturally-specific, family-centered, and trauma-informed support for women and childbearing families. They offer timely and improved access to services that extend beyond singular interventions ranging from home visiting services, mobile healing and nutrition units, and trauma counseling to mental health support. Services are administered by various professionals including doulas, grief and bereavement counselors, licensed mental health clinicians, and peer-to-peer advocates with culturally responsive skills. In particular, the access and partnership with doulas as independent, grassroots community-based advocates have been key to their overall success.

Given that perinatal mental health is complex, both organizations aim to provide comprehensive support addressing clinical, emotional, and tangible needs of birthing individuals. Their work considers multiple, intersecting marginalized identities of birthing individuals such as race, ethnicity, gender, and socioeconomic and immigration status. Interventions include assisting birthing individuals with food and diaper insecurity and housing instability when needed as well as supporting optimal physical recovery in tandem to promote mental health and well-being. Another key to their success is investing time and resources to build trusting relationships with birthing individuals and community. This is fundamental, as birthing individuals may not disclose symptoms or concerns due to the significant stigma of mental illness, which remains prevalent in many minoritized communities. In both organizations, interventions are delivered by or in partnership with individuals who share similar cultural backgrounds to aid in building trusting relationships.

4. DISCUSSION

We systematically examined and summarized 22 studies that included community-based interventions and programs to support the mental and psychological health of birthing individuals in the US. We found that most community-based interventions and programs represented in the published research in the US were delivered at home or in other community locations and by those whom we believed to be community agents of change including community health workers, midwives, nurses, and doulas. This finding aligns with existing literature indicating that home visitation programs and community-based interventions delivered by community health nurses are common and effective for individuals with serious mental illness.⁵⁸ However, most interventions did not consistently prevent or improve adverse mental and psychological health outcomes across all time points and populations. Several reasons may contribute to these inconsistent findings. Most interventions and programs were designed to improve general maternal and child health outcomes and parenting skills, rather than specifically focusing on mental and psychological health. While comprehensive education on general maternal infant health and parenting could reduce potential stressors and have secondary positive effects on mental and psychological health outcomes, more focused therapeutic interventions (e.g., counseling) may have been needed to significantly improve mental and psychological health.

It is also important to note that most of the interventions in the reviewed studies consisted of primary (e.g. health education, anticipatory guidance) or secondary (e.g. screening) prevention for individuals without mental illness or poor psychological health. As such, interventions were potentially less likely to have significant effects in terms of reducing symptoms. Similarly, many studies had small sample sizes, which contributed to reduced statistical power to detect small effects or significant changes, even when mental health was the primary outcome. For example, some of the included studies showed observable improvements in mental and psychological health outcomes that were not statistically significant. In addition, few studies reported involving community members and birthing individuals on their study team or on community-advisory boards to inform intervention development. Understanding the needs and preferences of the intended perinatal populations is essential to developing effective interventions and improving mental health outcomes.⁵⁹ Our community partners agreed with this statement, and emphasized that it is critical to include families, communities, and affected individuals in intervention design and development for sustainable change and transformation to support perinatal mental and psychological health.

Another important consideration in interpreting the current review findings is that, although most of the interventions included racially and ethnically diverse participants, only a few were developed specifically for racially and ethnically minoritized populations. Other studies also suggest that community-based interventions and programs reported in the peer-reviewed literature often do not explicitly prioritize those at the highest risk for pregnancy-related morbidity and mortality in the US.^60,61 The historical legacy of obstetric violence and discrimination in the US, perpetuated by structural racism and provider bias, contributes to the twice greater risk of postpartum depression and pregnancy-related mortality among birthing individuals of color compared to their White counterparts.⁶² As such, previous studies and our review findings indicate that simply including diverse participants in research may not be sufficient and that interventions designed to address the unique circumstances and underlying stressors (e.g., racism and discrimination) of minoritized groups are urgently needed.

We also observed a limited range of mental and psychological health outcomes that were prioritized among the included studies. The vast majority focused on addressing depression outcomes, followed by stress, particularly parenting stress. Only three studies examined anxiety, another common perinatal mental and psychological health condition,⁶³ and no studies examined suicidal ideation or behaviors. This is a critical gap, considering that about 8% of pregnancy-related deaths are due to suicide and 1% are likely due to suicide.⁶ Anxiety also plays a critical role in suicidality. Among pregnant and postpartum women with perinatal mental health conditions, those with anxiety have higher rates of suicidal ideation.⁶⁴ Future research is needed to examine the effect of community-based interventions on more comprehensive perinatal mental and psychological health issues including reducing or preventing suicidal ideation and behaviors.

4.1. Recommendations for building sustainable community-based interventions

To further strengthen community-based interventions and research that prioritize the needs of racially and ethnically minoritized birthing individuals, our community partners consistently highlighted the importance of actively engaging and involving community members, leaders, and organizations representing minoritized populations. Community involvement should ideally occur in the intervention design, implementation, and evaluation phases to incorporate culturally relevant practices, languages, and traditions in the intervention content and delivery. It is also important to disseminate information about perinatal mental health in ways that resonate with specific cultural groups (e.g., information delivery by trusted figures or community leaders). Structurally, funding agencies should increase and expand the number of opportunities for long-term community-academic partnerships that can evaluate and replicate existing programs that are successful in individual communities, particularly for interventions intended for minoritized populations. Importantly, there is an urgent need for future research to formally and rigorously evaluate the effectiveness of existing programs, which currently rely on anecdotal evidence from success stories or testimonials of participants and community health workers or professionals. The National Institutes of Health’s IMPROVE initiative is one such example that requires academic partnerships with community agencies to address maternal morbidity and mortality, which could also facilitate the formal evaluation of existing programs through research.

Community partners also underscored the limitations and missed opportunities related to the measurement of perinatal mental health when using currently available tools and instruments. Many instruments have been psychometrically developed and tested among primarily Non-Hispanic White birthing individuals and often feature words and expressions that are not used in racially and ethnically diverse communities.⁶⁵ Most of the studies included in this review employed mental health measures that have been widely used in general perinatal populations and in racially and ethnically diverse communities. However, none of them have been developed specifically for these populations, threatening the full validity of these instruments. For example, Edinburgh Postnatal Depression Scale and the Patient Health Questionnaire-9 are two widely used depression measures in perinatal populations. The median predictive accuracy of these instruments is low among American Indian (67%) and Alaskan Native (26%) birthing individuals, due to limited discriminative validity.⁶⁵ Similarly, the differentiation between anxiety and depressive symptoms in the Edinburgh Postnatal Depression Scale may not align with the conceptualization or experiences of depression and anxiety among Black and Hispanic birthing individuals.^66,67 These limitations indicate that mental health may not be conceptually equivalent across different languages and cultures, and that improved strategies are required. Examples of such strategies include tailoring translations and versions of instruments to more accurately assess perinatal mental health for racially, ethnically, culturally, and linguistically diverse communities.

Lastly, our community partners emphasized that, while community-based interventions play an important role in addressing perinatal mental health issues, they may not be the enduring solution to the maternal mortality and morbidity crisis. Priority should be given to ensuring equitable access to high-quality prenatal and postpartum healthcare services for minoritized populations. Birth justice movements have been calling for expanding access to culturally competent healthcare providers and improving health insurance coverage.⁶⁸ The limitations of billing and reimbursement for community-based healthcare services, disinvestment in public health programs and community health agencies, as well as the reliance on women as unpaid caregivers place many successful interventions and programs at risk.^69–72 Many community health workers are overworked and grossly uncompensated.⁷³ While it is important for minoritized communities to build their capacity to advocate for their own health needs, mobilize resources, and support community-led initiatives, community partners suggested that greater political and organizational changes in the larger US obstetric health care system are required to improve perinatal mental and psychological health and reduce disparities.

4.2. Summary and limitations

In summary, we conducted a comprehensive review of the literature to summarize the evidence base for community-based perinatal mental health interventions in the US. Our findings are strengthened by the perspectives and expertise of leaders from community organizations conducting this work in the field. As a result, there are clear opportunities to improve perinatal mental health through clinical care, research, and community investment and collaboration. Like any study, this systematic review also has limitations. First, we operationally defined community-based interventions in relation to where the intervention was delivered and who delivered it. Given that ‘community-based’ can be broadly defined, a different definition of community-based intervention (e.g., only including non-clinical settings, online communities) may lead to different findings. Second, we excluded qualitative studies to provide a cohesive synthesis and examine intervention effects. However, relying on statistical significance as the primary measure to evaluate the success of interventions has its limitations; such evaluations could be strengthened and enriched by including qualitative interviews and other measures that more fully capture clinical significance. Thus, a future systematic review focused on qualitative evaluations may bring additional insights into why certain community-based interventions may or may not have been effective. Third, we excluded studies that were exclusively focused on adolescent populations. This was necessary considering that this specific age group may be at a higher risk of perinatal mental and psychological health issues and should be further examined in a separate review. Fourth, we defined the perinatal period as up to 1 year postpartum and excluded studies that measured mental and psychological health outcomes beyond the first year after birth. Our community partners noted that birthing individuals may require mental and psychological health support beyond the initial perinatal period. A future review focused on long-term perinatal mental and psychological health outcomes may add further insights. Lastly, non-peer-reviewed studies and grey literature were not included, which may have introduced publication bias.

5. CONCLUSION

In this systematic review, we have highlighted the complexity and challenges of addressing perinatal mental and psychological health through community-based interventions in the US. We found that community-based interventions were often delivered in home settings by community health workers, nurses, midwives, and doulas with mostly mixed effects on perinatal mental and psychological health outcomes. This may be due to interventions designed for the promotion of general maternal and child health rather than specifically addressing mental and psychological health concerns. This indicates a critical need for more focused and therapeutic approaches to effectively address mental and psychological health issues among birthing individuals. Our review findings further emphasize the importance of incorporating the needs and experiences of racially and ethnically diverse populations into intervention design. Inclusion of representatives from communities provides important insight into the deeper issues of discrimination and structural inequalities that contribute to perinatal mental and psychological health disparities. It is evident that while community-based interventions offer valuable support, there is a pressing need for tailored mental health measures, greater community involvement, and systemic changes in the broader US obstetric healthcare landscape to improve perinatal mental and psychological health outcomes.