The role of government assistance, housing, and employment on postpartum maternal health across income and race: a mixed methods study

Abstract

Background

Stressful large-scale events, such as the COVID-19 pandemic and natural disasters, impact birthing individuals’ postpartum experiences and their mental health. Resultant changes in government assistance, housing, and employment may further exacerbate these impacts, with differences experienced by varying income levels and races. This study aimed to examine maternal depression and anxiety in postpartum individuals by income and race during a stressful large-scale event, and the mediating role of government assistance, housing, and employment.

Methods

An explanatory sequential mixed methods study was conducted (QUANT + QUAL). For aim 1 (quantitative), birthing individuals who delivered during peak pandemic (June 2020 - September 2021) completed questionnaires related to their perinatal experiences and mental health. Macrosystem factors (government assistance, housing, and employment changes) were assessed using the Psychosocial Recommended Measures. The Edinburgh Postnatal Depression Scale (EPDS) and the Generalized Anxiety Disorder-7 (GAD7) assessed depression and anxiety, respectively. Serial linear regression models assessed the relationship between race and income with mental health and macrosystem factors. For aim 2 (qualitative), 40 individuals from the quantitative study balanced by income (low vs. high income) and race (Black vs. White) completed one-on-one semi-structured interviews which were analyzed using thematic analysis.

Results

Amongst 1582 birthing individuals, Black individuals had a significantly higher EPDS score compared to White counterparts. Not receiving government assistance, unstable housing, and experiencing various employment changes were all related to worse mental health during stressful large-scale events. In semi-structured interviews, low-income individuals discussed that government assistance helped alleviate a financial and mental burden. Low- and high-income individuals reported varying job changes that impacted their mental health (low-income: job loss, high-income: increased hours).

Conclusions

This research spotlights the negative impact of large-scale events most affected both Black and low-income individuals’ postpartum mental health, and the role of government assistance, stable housing, and secure employment in helping to alleviate these disparities between income levels.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-024-20745-w.

Background

The first 1,000 days of life, pregnancy to 24 months, are a critical time for maternal and child physical and mental health [1, 2]. The major shift in maternal time and energy demands associated with caregiving may increase risk for depression and anxiety, which have short- and long-term consequences on maternal and child health [3–6]. For birthing individuals, depression during the perinatal period is associated with increased maternal morbidity and mortality, including pregnancy complications [7]. For the infant, maternal postpartum anxiety and depression may lead to difficulties breastfeeding in the short term [5], and negatively impact infant growth and development in the long term [3, 5]. These impacts necessitate supporting birthing individuals for immediate and lasting benefit.

Large-scale stressful events, such as natural disasters, may further impact individuals’ mental health and ability to care for their newborns [8, 9]. Indeed, the main driving factor in disaster-related changes to child development is not the event, but maternal mental health [9]. During the COVID-19 pandemic, individuals had a higher risk for developing anxiety and depression compared to pre-pandemic individuals, especially perinatal individuals [10–12]. This increase is attributed to unique perinatal stressors such as disruptions in healthcare, visitor restrictions, and perceived risk or fear of contracting COVID-19 during pregnancy [13, 14]. Other stressors were specific to mothers of young children, including additional childcare responsibilities while working at home, employment disruptions, and income loss [15, 16].

COVID-19 hardships may have been amplified by a coinciding shortage of infant formula and local natural disasters (i.e., Hurricanes Laura [2020] and Ida [2021]) [17, 18]; these hardships may have especially impacted individuals with low-income [19] and people of color [20], including Black individuals, who are already at risk for mental health problems and adverse pregnancy outcomes [21, 22]. The U.S. government enacted policies and provided financial support to families in an effort to negate these hardships, including increased Supplemental Nutrition Assistance Program (SNAP) and Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) benefits, as well as allowing flexibilities such as certification for eligibility to occur remotely [23]. However, despite government assistance, those simultaneously experiencing financial hardship may still have an increased risk of anxiety and depression as found one study of 28,842 working adults [24]. Possibly additional macrosystem factors such as housing insecurity and employment changes further negatively impacted mental health [20, 25]. The mental health of perinatal individuals may have been additionally impacted by the widespread shortage of infant formula. This shortage occurred primarily due to a safety recall by Abbott Nutrition, a supplier responsible for 40% of the U.S. infant formula supply and whose products are included in WIC food packages. Black and low-income individuals may have been particularly impacted by the formula shortage as they are less likely to breastfeed [26–29]. The global pandemic, national formula shortage, and local natural disasters all coincided to create an increasingly difficult time for individuals of low-income and Black individuals.

COVID-19 pandemic stressors affected everyone [10, 11], but the literature is mainly focused on the mental health of higher income and non-Hispanic White perinatal individuals [3, 12], rather than Black individuals, especially individuals pregnant during this time. Investigating the potential role that macrosystem factors (government assistance, housing, and employment) may play in mediating differences in mental health by race and income may elucidate options to mitigate the impact of large-scale stressful events on other groups. We recognize race is not a biological variable but rather a surrogate for individuals’ lived experiences related to systemic/structural racism and discrimination [30, 31]. This study aimed to examine maternal depression and anxiety in postpartum individuals by income and race during a stressful large-scale event, and the mediating role of government assistance, housing, and employment. We hypothesized low-income and Black individuals who were pregnant during the COVID-19 pandemic would have an increased risk for perinatal anxiety and depression compared to high-income and White counterparts, and that changes in housing (i.e., loss of housing) and employment (i.e., hours reduction or job loss) would partly explain these differences.

Methods

Overall study

This study used a sequential explanatory mixed methods design (QUANT → QUAL) to collect cross-sectional quantitative data via an online survey, followed by one-on-one semi-structured interviews. A mixed methods strategy was employed to combine the strengths of quantitative and qualitative research to generate a more complete understanding of research questions and their context [32]. Thus, a sequential (QUANT → QUAL) explanatory mixed methods study was conducted to examine prevalence of depression and anxiety, whereas semi-structured interviews provided context for these findings. Two populations contributed to the survey and interviews to create one postpartum cohort; these populations included (1) recipients of the WIC in a southern U.S. state, Louisiana, and (2) patients of large woman’s speciality hospital in a metropolitan city of the same southern U.S. state of all income statuses. This report follows the mixed methods reporting standards (see Supplementary Table 1) [33, 34]. Pennington Biomedical Research Center’s institutional review board approved this study (IRB#: 2021-042). The intent of this study was to examine the role of the COVID-19 pandemic on maternal mental health. However, coinciding major stressful events also took place during the data collection, including two category-4 hurricanes (Hurricane Laura: 2020, Hurricane Ida: 2021), and a national formula shortage (2022).

QUANT: online survey

From the two populations, we invited women who were ≥ 18 years old, recently postpartum (delivery date between June 10, 2020 – September 30, 2021), English speaking, and provided permission to link survey responses to medical records to participate. Individuals who were unwilling or unable to provide consent were excluded. Individuals were recruited through direct text message, phone contacts, and contacts at postpartum appointments (hospital population only).

Potential participants were sent a link to a secure website where they were asked to provide consent, then provided completed surveys related to demographic information, their pregnancy, perinatal mental health, and pandemic-related life experiences. At the end of the survey, they indicated interest in the interview portion of the study [35]. Quantitative data collection occurred between December 2021-December 2022, thus respondents were 6-months – 1.5-years postpartum when completing the survey.

Demographics and macrosystem factors

Individuals reported their age, race, ethnicity, educational level, household income, insurance status, and delivery hospital. Household income had 10 different options, the first five options were in $10,000 USD increments (beginning with $0 to $49,999 USD), then $50,000–74,999 USD, $75,000–99,999 USD, $100–199,999 USD, $200,000 USD and finally, an option to “rather not say”. Household income was then categorized into low (annual household income <$50,000 USD), middle ($50,000–99,999 USD), and high-income (≥$100,000 USD). Questions were derived from the National Institutes of Health Recommendations for Common Data Elements for COVID-19 studies Tier-1 Psychosocial Recommended Measures [36]. We specifically assessed the socioeconomic status, housing, and emergent financial strain domain, including government assistance, loss/change of housing, changes in employment situation, and distress from changes.

Government assistance was assessed by three questions including (1) “enrollment in WIC”, (2) “receiving one or more stimulus checks from the government”, and (3) “any amount of unemployment benefits from the government”. Response options included “yes”, “no”, “not sure”, or “decline to answer”. To assess loss/change of housing, participants were asked “Has the COVID-19 pandemic led to any of the following” with three sub-questions (1) “relocation or moving from where you lived before the pandemic (e.g., downsizing, moving in with family, etc.)”, (2) “facing possible eviction”, and (3) “becoming homeless”. Response options were either “yes” or “no” [37]. Changes in employment were assessed by querying “Which of the following changes in employment have occurred due to the COVID-19 outbreak?”, and participants were asked to check all that applied [38] from eight options: move to remote work/telework, loss of hours, decreased pay, loss of job, decreased job security, disruption due to childcare challenges, increased hours, and other changes. Response options were “yes”, “no”, or “not applicable”. Employment-related stress was further investigated via the question “How bothersome or distressful were changes to your family’s employment?”, with five response options ranging from “not at all” to “very much” [39].

Mental health

The primary outcome of this analysis was depression, which was measured with the Edinburgh Postnatal Depression Scale (EPDS); this 10-item questionnaire asks about participants’ feelings over the past 7-days, including ability to laugh, being anxious or worried, and coping [40]. The questionnaire was reduced to 9-items by removing the question on self-harm for the purposes of this study. All questions had four response options that varied by question. Response options, ranging from “as much as I always could” to “not at all,” or “not at all” to “yes, most of the time” depending on the question, are reverse coded so that higher responses are indicative of depressive symptoms. Responses are summed, with a maximum score of 27, and probable depression is indicated by a score ≥ 12 [41]. Maternal anxiety was measured by the Generalized Anxiety Disorder-7 (GAD-7) scale, which is a 7-item questionnaire that inquires about problems related to anxiety (e.g., trouble relaxing, worrying too much about different things) across the past two weeks. There were four response options, ranging from “not at all” to “nearly every day”, with a higher score indicating a higher level of anxiety. The maximum score was 21, and probable anxiety was defined as any score ≥ 10 [42]. Both measures have been validated in pregnant and postpartum individuals [40, 43, 44].

QUAL: semi-structured interviews

Individuals who completed the entire QUANT survey and indicated their race as “Black” or “White” and interested in being contacted for an interview were eligible for this aim of the study. Low-income individuals were recruited from the WIC sample (estimated annual household income <$50,000 USD), while individuals who self-reported a household income $100,000 USD in the hospital sample were recruited for the high-income sample. The low and high-income samples were further stratified by race (Black and White), and then invited in random number order to participate. We sought 10 participants from each income and race combination (40 total) based on best practices for qualitative studies and previous investigations [45].

One-on-one interviews were conducted via a secure online platform by researchers trained in qualitative methods (CLK, MSD, BJ, ES, JB). Qualitative interviews occurred between March – June 2023, so participants were 1.75–3-years postpartum and approximately two years into the pandemic. Interview procedures included obtaining informed electronic consent and following a pre-piloted interview guide. Specifically, individuals were asked how the pandemic impacted their (1) daily lives (employment, housing, family, and personal life), (2) perinatal experiences, and (3) mental health (Supplementary Table 2). All interviews were video recorded, transcribed verbatim by Microsoft Stream, and checked for accuracy.

Data analysis

QUANT: statistical analysis

Only individuals who indicated they were Black or White and had complete data were included in analysis. Participants with “n/a” for some variables were included in the data set but not used for yes/no comparison. For our first research question, mean EPDS and GAD-7 scores and probable depression (yes/no) and anxiety (yes/no) were compared between Black and White individuals, and by income using a linear model for EPDS and GAD-7 scores, and chi-square analyses for categorical analysis, respectively. Linear and logistic models were constructed for both outcomes, including fixed effects for race, income, and their interaction with each other, with adjustments for age, insurance (private, public, or other), education level, and sample (hospital vs. WIC). Covariates in adjusted models were investigated for multicollinearity and were within acceptable ranges. Next, potential mediators (government assistance, housing status, employment changes due to the COVID-19 pandemic, and pandemic stress) were examined by predictor (race, income) and outcome (mental health), using chi-square, Pearson correlations, and independent t-tests. Linear regression models between identified exposure, mediator, and outcome were explored with adjustment for the same covariates. Finally, a pre-planned mediation analysis was conducted to test the direct and indirect effects of variables that were associated with both exposure (race or income) or outcome (mental health). This approach was conducted with the PROCESS vs3.5 macro with 10,000 bootstrap intervals with unstandardized estimates, and adjustment for the same covariates [46]. Analyses were conducted using R statistical software, except for PROCESS which was completed in SAS (version 9.4, Cary, N.C.) and statistical significance was set at p < 0.05.

QUAL: thematic analysis

Saturation was reached at 40 interviews (10 per race and income categories), and no additional interviews were conducted. Transcribed interviews were coded using inductive thematic analysis (i.e., iterative coding and coding tree development [topic, root codes, and code names]) [47]. Three authors (CLK, KOG, and MSD) trained in qualitative analysis reviewed the transcripts to create an initial codebook, and a final codebook was created after a second transcript review. Transcripts were coded independently, and disagreements were resolved by discussion. Selected codes were read by each reviewer independently to determine themes, allowing themes to be developed from participant responses across codes [48]. The finalized themes were informed by study aims and quantitative analysis results [49]. Atlas TI software was used for data management (Berlin, Germany). Description of the sample compared to the full quantitative sample was completed with maternal mental health outcomes (depression, anxiety), and macrosystem factors variables were compared by race and income using a chi-square or Fisher exact test.

Integration

Quantitative and qualitative results were integrated to compare inferences. Results from the same concepts were presented in a joint display to visualize the overall results of the mixed methods study [50]. The joint display includes respective concept, quantitative findings, qualitative findings, and their meta-inference.

Results

QUANT: Survey

In total, 3706 individuals initiated the survey, and 1582 individuals with complete data were included. Excluded individuals did not self-report as White or Black, did not report race or did not complete this item (n = 1424); reported “rather not say” for any variable (e.g., income, COVID-19 variables) (n = 490); did not provide insurance information (n = 61); were outside the age range (i.e., too young or improbable age, n = 10); or reported “not sure” for government assistance (n = 52).

Those who had complete data and were included in the analysis (n = 1582) were mainly White (69.2%), with slightly more low-income individuals (42%) than high (31%) or middle-income individuals (27%). Most individuals received one or more government stimulus payments (90%), and a third enrolled in WIC (38%) or received unemployment benefits (23%). The high frequency of government stimulus payments in this cohort may be attributed to families with a household income up to $200,000 USD qualifying for government stimulus payments. Few relocated (14%), faced possible eviction (8%), or became homeless (4%) during the pandemic. Around a third of individuals reported remote work (32%) or loss of hours (33%), and some reported decreased pay, loss of job, decreased job security, or increased hours (< 20% for all). Half of individuals reported disruptions due to childcare challenges (46%) and reported employment changes were “somewhat” to “very much” bothersome to daily life (44% total). A quarter of individuals were classified as having depression (23%) or anxiety (26%, Table 1). Additional details on demographic characteristics, and comparisons between included and not included participants can be found in Supplementary Table 3, respectively.

Table 1.

Comparison of mental health outcomes by race and income of individuals pregnant during the pandemic (n = 1582)

		EPDS Score			Depression
	N (%)	Mean (SD)	P- Valuea	P- Valueb	Yes, N (%)	No, N (%)	OR (95% CI)^	aOR (95% CI)^
Total sample		8.0 (4.9)			367 (23)	1215 (77)
		GAD-7 Score			Anxiety
Total sample		6.4 (6.0)			413 (26)	1169 (74)
Race
		EPDS Score			Depression
White	490 (31)	7.7 (5.4)	0.10	0.003	129 (26)	361 (74)	Ref	Ref
Black	1092 (69)	8.1 (4.7)			238 (22)	854 (78)	0.78 (0.61; 1.00)	0.97 (0.75; 1.27)
		GAD-7 Score			Anxiety
White		6.4 (6.2)	0.68	0.30	144 (29)	346 (71)	Ref	Ref
Black		6.3 (5.8)			269 (25)	823 (75)	0.79 (0.62; 1.00)	0.96 (0.75; 1.25)
Household Income
		EPDS Score			Depression
High (>$100,000 USD)	493 (31)	7.4 (4.3)	0.01	0.47	82 (17)	411 (83)	Ref	Ref
Middle ($50,000-$99,999 USD)	426 (27)	8.2 (4.6)			96 (23)	330 (77)	1.46 (1.05; 2.03)	1.17 (0.82; 1.66)
Low (<$49,999 USD)	663 (42)	8.3 (5.6)			189 (29)	474 (71)	2.00 (1.50; 2.68)	1.26 (0.85; 1.88)
		GAD-7 Score					Anxiety
High (>$100,000 USD)		5.5 (5.3)	0.001	0.32	95 (19)	398 (81)	Ref	Ref
Middle ($50,000-$99,999 USD)		6.6 (5.8)			113 (27)	313 (73)	1.51 (1.11; 2.07)	1.19 (0.81; 1.75)
Low (<$49,999 USD)		6.9 (6.4)			205 (31)	458 (69)	1.88 (1.42; 2.48)	1.24 (0.89; 1.74)

Data presented as mean (SD) or N (%) as specified. ^aunadjusted; ^badjusted models; Effect estimates and corresponding P values derived from linear model including either race or income as predictor and adjusted for education, insurance, age, and study recruitment site.; ^Odds ratios (OR) alongside 95% confidence interval (CI) are derived from binomial logistic generalized linear model with either race or income as predictor; and adjusted odds ratios include adjusted for education, insurance, age, and study recruitment site. EPDS = Edinburgh Postnatal Depression Scale; Depression is indicated by a EPDS score ≥ 12; GAD-7 = Generalized Anxiety Disorder-7; Anxiety = GAD-7 score ≥ 10

Race, income, and mental health

Race

Black individuals had a significantly higher total score on the EPDS compared to White individuals in adjusted models (p = 0.003, Table 1); this difference was around 1-point (0.85 [95% CI: 0.29; 1.40]). However, there was no significant difference in depression classification from the EPDS, GAD-7 total score, or anxiety classification based on the GAD-7 in adjusted models (p‘s > 0.05).

Income

In unadjusted models, middle and low-income individuals reported higher scores and worse mental health outcome prevalence compared to high-income individuals (p’s < 0.05), but these were no longer significant after adjustment for covariates (p’s > 0.05). In adjusted models, the interaction between race and income was not significant for EPDS score (p = 0.11) or GAD-7 score (p = 0.19) indicating there were no differences in these associations based on the level of the other predictor.

Macrosystem factors and mental health

In adjusted models, any housing change (relocation, possible eviction, becoming homeless), or multiple employment changes were associated with a higher EPDS score (see Supplementary Table 4). This translated into significantly higher odds of depression for any housing change and all employment changes, except moving to remote work. Individuals who experienced any of the housing or employment changes had a higher GAD-7 score, and higher risk of anxiety (see Supplementary Table 5). Notably, having childcare disruptions was related to 81% higher risk of having anxiety (aOR: 1.81, 95% CI: 1.43; 2.30, p < 0.001, Supplemental Table 5).

Employment-related stress was positively correlated with EPDS score (r = 0.33) and GAD-7 score (r = 032, p’s < 0.05). Government assistance was not related to differences in depression classification, GAD-7 scores, or anxiety classification (p’s > 0.05, Supplementary Tables 4 and 5).

Race, income, and macrosystem factors

Race

There was a significant difference in receiving multiple government assistance, experiencing unstable housing, and multiple employment changes between races (p’s < 0.05, Supplementary Table 6). White individuals were more likely to report WIC enrollment, possible eviction, becoming homeless, and job loss compared to Black individuals (p’s < 0.05). Black individuals were more likely to report government unemployment benefits, loss of hours, decreased pay, and decreased job security compared to White individuals (p’s < 0.05). White individuals reported slightly higher employment-related stress (2.69 ± 1.89) on the five-point scale (range 1–5) compared to Black individuals (2.44 ± 1.37, p = 0.001), which represents that changes to their family’s employment were “a little bit” and “somewhat” bothersome or distressful.

Income

All macrosystem factors differed by income. Compared to high-income individuals at least twice as many low-income individuals reported using government assistance options, experiencing housing changes, loss of hours, decreased pay, job loss, and decreased job security (p’s < 0.05, Supplementary Table 6). High-income individuals reported slightly more remote work and increased hours, and middle-income individuals reported lower childcare disruptions relative to low-income individuals (p’s < 0.05). Middle and low-income individuals reported higher employment-related stress than high-income individuals (p’s < 0.05).

Mediation model: race, macrosystem factors, and mental health

Employment-related stress significantly differed across all macro system factors except receiving government COVID-19 stimulus payments (Supplementary Table 7). Given this overlap, it was pursued as the primary mediator since it was associated with both predictor (race) and outcome (EPDS score). After adjustment for covariates, there was no difference between races in employment-related stress score (p = 0.43). In mediation analysis, the association between race and employment-related stress continued to be non-significant (p = 0.30), though the paths from both races and to EPDS score were significant including their indirect effects (p’s < 0.05, Supplementary Fig. 1). These results indicate race and employment-related stress had independent effects on the EPDS score in this postpartum cohort.

QUAL: semi-structured interviews

In total, 40 individuals balanced by race and income participated in interviews. Prevalence of mental health outcomes (depression: 17.5%; anxiety: 25%), and macrosystem factors mirrored the full sample (Supplementary Table 8). As expected by study design, and in agreement with the larger cohort, racial differences were not present (p’s > 0.05), but low-income individuals had higher prevalence of WIC enrollment, unemployment benefits, relocation, and eviction, and lower prevalence of telework compared to high-income individuals (p’s < 0.05). Themes were later categorized by covariates explored in the quantitative survey to align findings, including government assistance, housing, and employment. One to two themes align with each macrosystem factor, and illustrative quotes are shown in Table 2. Differences in experiences were prevalent by income and described in these themes.

Table 2.

Thematic findings related to government assistance, employment, and housing amongst birthing individuals pregnant during the COVID-19 pandemic^

Theme	Quotes Representative of theme
Government assistance
Low-income birthing individuals discussed that government assistance helped alleviate a financial and mental burden.	• “So I think I’ve got COVID stamps like every month and that helped me with making groceries and being able to buy, you know, things that may have been a little bit more expensive, you know. So it helped me like buy things that’s actually really healthy for me instead of actually getting cheaper items that maybe a little bit healthy but not that maybe the healthiest option, you know?” (Birthing Individual #1, Low-income, Black) • “I applied like literally once I lost a job in March that’s when I applied. And I guess they felt like June and July came, they kind of felt like, okay, you should have a job like you should be looking for a job now…And then by the time, that time came where, you know, it was kind of given that way to where you could go back to work, I was due to have a child anytime soon, so I couldn’t even work if I wanted to” (Birthing Individual #2, Low-income, Black) • “So I lost my food stamps and I was struggling to keep food in the house. And I was like explaining to people, I’m like, I have, I’m pregnant you know, I’m about to have a baby. I, I I’m not working a whole lot. I’m working just enough to get by, you know, and I need my food stamps… so then I finally, right before I had the baby, um, or right after I had the baby, I got my food stamps back and it was a, a big release, a really big release.” (Birthing Individual #3, Low-income, White) • “I was doing good for a couple months whenever they shut down the casino and we were all getting that extra six hundred dollars a week in unemployment. I was getting almost like eight hundred dollars a week in unemployment and I was like, that was actually more than what I was making working. So I was happy about it.” (Birthing Individual #4, Low-income, White)
The formula shortage revealed the importance of WIC supplied formula, and negatively impacted Low-income birthing individual’s mental health.	• “[child] used a certain kind of formula and it was out like everywhere. I had to ask around to friends and, and find like little places, like little shops that you wouldn’t usually expect to go to and call places to make sure they had it. It was, it was a big pain for a while… because I had WIC, so you had to find which WIC stores had it on top of which stores you know had it. So that made it even more difficult.” (Birthing Individual #5, Low-income, White) • “When you had to order the formula [online], which is what we were doing because I would never ever find it in the stores, I could not use my WIC for it…because every month you would just lose [WIC spending], you know? Because, they it wasn’t building up so it was, you know it was tough because formula got expensive. Thank God I have family you know that was willing to help because if not I don’t know how we would have done it.” (Birthing Individual #6, Low-income, Black,) • “…this store didn’t have it because you got to think about, I had a baby. We need a formula too. So then that was a worry about not getting what she needed. But then I ended up I think ordering stuff from Amazon for her thankfully. And they had what I was able to order ahead and what she needed and get like a supply on a monthly basis kind of type thing.” (Birthing Individual #7, High-income, Black)
Housing
Low-income birthing individuals reported a higher degree of traumatic events which influenced mental health, whereas High-income birthing individuals seemed buffered from extreme life situations.	• “I feel like I’ve been through it, a, a greater experience with the loss of my daughter and I, and it’s like I don’t contribute any of those things, pandemic related.” (Birthing Individual #1, Low-income, Black) • “I turned 40 in the pandemic, pregnant with gestational diabetes. So I wasn’t able to like have a birthday party. I wasn’t able to eat cake, I wasn’t able to celebrate with friends, like I wasn’t able to do anything.” (Birthing individual #9, High-income, White) • “the hurricane happened and then we lost our house, but then we moved, but we had to start all over and then we didn’t have a vehicle. And it’s just a lot going on all at once.” (Birthing individual #7, Low-income, White) • “Honestly, it’s like we’re playing house sometimes because [staying home] was just like so nice and pleasant. It was spring. It was just nice.” (Birthing individual #8, High-income, White)
Employment
Birthing individual of all income levels experienced job disruptions, but of varying types that impacted mental health postpartum	• “Ah yes, employment definitely changed tremendously because I was out of work for maybe like 5 months because of my, uh high risk situation. Um. I work a real strenuous physical job. I was working in a warehouse … My doctor did, write a letter to my, write a letter letting my job know my restrictions…but it just being too much of bending and lifting or pushing and pulling… so. I end up having to take a leave from work…I did still receive payment due to my medical leave… eventually me and caseworker just came up with a plan put in my resignation, you know? I didn’t have enough money in my account to have more time off, you know, the time that I actually needed. So I just decided to leave.” (Birthing individual #11, Low-income, Black) • “So it was kind of adjustment to come back from my pregnancy to go back working with inmates…I had to just stay masked up. I had to take my clothes off when I got home from a 12 hour shift, put them in a bag, wash them, take a shower, basically like undress before I even get to see my child because my mom didn’t want me to bring work and sickness into the house.” (Birthing individual #12, Low-income, Black) • “Well, the biggest thing was in the early pandemic when there was a lot of daycare closures. So, then me and my husband trying to juggle working from home, who’s going to take care of the baby…it was kind of the juggle of work and childcare was the biggest change to home life.” (Birthing individual #13, High-income, White)

^Participant numbers indicate their self-reported income and race; WIC = women, infant, and children supplemental nutrition assistance program;

Government assistance

Low-income individuals discussed that government assistance helped alleviate a financial and mental burden

Low-income individuals enjoyed WIC benefits being extended to contactless food purchasing (i.e., pick up or delivery), which aligned with their concerns of the virus and reduced stress. Government stimulus checks and food stamps alleviated the burden of unemployment amongst concerns of bringing the virus home and finding a job while pregnant or recently postpartum. High-income individuals did not discuss stimulus checks, though 90% reported receiving this benefit.

The formula shortage revealed the importance of WIC-supplied formula, and negatively impacted low-income individuals’ mental health

More low-income individuals reported they had to rely exclusively on formula or supplement their breastmilk with formula to feed their infant compared to high-income individuals. Though WIC provides formula to postpartum individuals, many low-income individuals reported additional time and travel to find formula during the formula shortage, which impacted their mental health. WIC’s limited coverage of specialty formula (e.g., milk allergies) created further challenges for low-income individuals. In comparison, most high-income individuals reported breastfeeding their infant, but they had the time and resources available to facilitate obtaining needed formula.

Housing

Low-income individuals reported a higher degree of traumatic events which influenced mental health, whereas high-income individuals seemed buffered from extreme life situations

Low-income individuals reported many other major events, such as deaths within the family (e.g., grandparents or children) and housing changes (i.e., displacement or damage to house), compared to high-income individuals who had to delay events (e.g., birthday party, wedding, surgery, buying a house). Low-income individuals felt that exposure to these traumatic events made them more resilient to adapt to the major life disruptions during the pandemic. As one individual described: “I already had a hard life, you know, so I was able to adapt to difficult situations” (Individual #3, low-income, White). High-income individuals also reported they were adapted to taking precautions during the pandemic, and many acknowledged that they had no major life events co-occur. Many individuals, regardless of income, experienced hurricanes during postpartum and short-term distress. Only low-income individuals reported significant negative long-term impact, such as changes to housing or infrastructure, attributed to hurricanes and COVID-19 restrictions.

Employment

Individuals of all income levels experienced job disruptions, but of varying types that impacted mental health postpartum

High-income individuals reported more flexible jobs, i.e., telework and adaptation for COVID, which allowed additional leave and flexibility to resume work. However, these individuals’ daily life and workload were negatively impacted by childcare disruptions. Some high-income individuals quit their jobs postpartum to stay home with their children, or their partner’s job allowed the partner to stay home and help with childcare. These events contrasted with low-income individuals who reported losing jobs during pregnancy (mid-peak pandemic). Reasons for job loss included a few layoffs but many reported it was related to being pregnant (e.g., employer virus concerns, in-person labor demands). Low-income individuals then continued the postpartum period unemployed, as maternity leave (paid or unpaid) was available to only a few. Low-income individuals who worked postpartum were able to sustain income, and continued precautions to not bring the virus home from work, which impacted their mental health.

Integration

A joint display of quantitative and qualitative analysis on major constructs is presented within Table 3. Quantitative results revealed that both Black individuals and middle or low-income individuals had worse mental health outcomes; after adjusting for confounders, this relationship was only significant in Black individuals. Employment-related stress may not explain the relationship between race and mental health, even though it was related to many macrosystem factors. Qualitative results further revealed other considerations (e.g., virus concerns), and contextual factors (i.e., formula shortage and hurricanes) that may have influenced individuals’ mental health, and further exacerbated differences in macrosystem factors.

Table 3.

Joint Display of quantitative and qualitative findings, and Meta-inference

Key finding	Quantitative Findings	Qualitative Findings	Meta inference
Race, Income, and mental health	• Middle and low-income individuals had worse mental health outcomes in unadjusted models. • After adjusting for covariates (insurance, site, education, and age), only statistically significant results were that Black individuals had worse depressive symptoms.	Experiences that impacted mental health were notably different by income.	• Black, and middle to low-income birthing individuals had higher rates of depressive and anxiety symptoms relative to their counterparts. • Other factors, e.g., education, do not account for racial differences in mental health.
Government assistance and maternal mental health	• Lower income individuals were more likely to receive government assistance, but high-income individuals did still receive government stimulus checks • WIC and unemployment redemption differed by race. • In adjusted models, receiving government assistant was not related to differences in anxiety and depression.	• Government assistance relieved employment-related stress for low-income birthing individuals. • High-income birthing individuals did not discuss the impact of government benefits on their mental health though some received government stimulus checks. • WIC benefits also caused stress for formula feeding individuals.	• Government assistance was helpful for mental health, especially low-income birthing individuals, but possibly not during the formula shortage. • Redemption and enrollment in benefits differed by race.
Housing events and maternal mental health	• Housing changes were double or more as likely in low-income birthing individuals. • Housing changes were related to worse anxiety and depression.	• Major housing changes occurred due to income constraints, but other personal events and natural disasters that occurred during this time frame. • Though high-income birthing individuals experienced the same natural disasters, low-income birthing individuals were the only ones to report changes in housing.	• Major housing changes for low-income birthing individuals may be attributed to changes in both employment and other natural disasters. • Regardless of origin, these events significantly impacted individuals’ mental health.
Employment and maternal mental health	• Black and white birthing individuals experienced different employment-related changes. • Employment related changes also differed by income. • Each employment change was related to higher employment-related stress, anxiety, and worse depression.	• High and low-income birthing individuals described different employment-related stressors. • Another factor that impacted their employment and support was the virus restrictions they imposed for their newborn.	• Multiple events contributed to employment-related stress by race and income. • A mediation model revealed race and employment-related stress had independent effects on depressive score postpartum. • Qualitative results emphasized that income and other restrictions contributed to individual’s employment-related stress experienced.

WIC = Special Supplemental Nutrition Program for Women, Infants, and Children

Discussion

The purpose of this study was to examine the relationship of race and income on mental health, and the mediating role of other macrosystem factors, among birthing individuals who were pregnant during stressful periods; these periods include COVID-19 pandemic, Hurricane Laura and Ida and an infant formula shortage. Black and low-income individuals had worse mental health outcomes compared to their counterparts, but after considering confounders, only Black individuals had worse depressive scores. All types of macrosystem factors were related to worse mental health outcomes, especially changes related to employment. Employment-related stress and race were separate factors in maternal mental health, based on our mediation analyses. The qualitative results revealed co-occurring national events not captured by the quantitative analysis that impacted individuals’ postpartum lives and mental health. Integration of these results underscore the negative impact of natural disasters on both Black and low-income individuals’ postpartum mental health, and the role of flexible government assistance, stable housing, and secure employment that may help alleviate these disparities. These findings spotlight the importance of continued unemployment benefits for low-income households, reducing possible constraints with WIC food packages, and promoting flexible and financially supportive employment options for all individuals pregnant during natural disasters.

We found partial support for our hypothesis that Black and low-income individuals would have higher perinatal depression and anxiety during this stressful time relative to White individuals and individual of other incomes. These results did not translate to the qualitative results, likely due to income-balanced design. Even so, these racial differences in mental health are supported by past literature that Black individuals were disproportionately impacted by the COVID-19 pandemic, resulting in poorer mental health compared to White counterparts [20]. The current study adds that disparities in mental health were prevalent in postpartum individuals by race, expanding our understanding beyond primarily White and high-income investigations [3, 12], and previous papers on individual coping practices in Black perinatal populations [51]. Given changes to mental health may be the main driver of a disasters’ impact on maternal health outcomes [9], these results highlight the need for supporting birthing individuals during natural disasters. It is unclear if traditional approaches would apply (e.g., coping practices and in-person meetings), given the high burden of depression, anxiety, and potentially post-traumatic stress disorder, and potential infrastructure recovery needed. One review identified social support was related to fewer mental health problems during the pandemic and hurricanes [52], and others have spotlighted the benefits of a free emergency hotline to support birthing individuals during these times [18].

We found partial support for our hypotheses that changes in housing and employment situation would explain differences in mental health by race and income. For housing, few (< 14%) reported major housing changes, potentially due to participants being allowed to continue to remain in their living situation due to other government support (e.g., stimulus checks, unemployment, or the eviction moratorium). Still, our qualitative sample experiencing housing insecurity described major effects on mental health in line with nationally representative surveys [25]. This survey and our quantitative results also found that employment may account for the relationship between housing insecurity and mental health [25]. Employment-related stress was related to job insecurity and additional responsibilities, which impact finances in opposing directions. Job loss may have reduced income but this was offset with COVID-19 unemployment benefits in low-income individuals who were then able to take that time with their infant [53]. The converse was presented for high-income birthing individuals, who were able to keep employment, but this came with additional responsibilities that were impacted by childcare disruptions (e.g., childcare closures) [16, 54]. In this perinatal sample, both directions of employment changes negatively impacted maternal mental health. Aligning these results with our first hypothesis, we found independent relationships between race and employment-related stress with mental health, emphasizing these are separate but important factors amongst the COVID-19 pandemic.

Different types of government assistance (i.e., WIC, stimulus checks, and unemployment) exhibited different effects on mental health. These findings add to a limited literature on stimulus and mental health [55, 56]. Other studies in this area found little effect of government assistance on the relationship between having low income or losing stable employment and mental health [24, 55, 56]. In these studies, government assistance was defined as an income-support policy amongst a larger cohort of the general population [55, 56], rather than specific to pregnant and postpartum individuals. Indeed, another notable government assistance received amongst postpartum individuals was the WIC food package. While this package was described by birthing individuals as helpful, during the formula shortage it was also difficult to navigate. Incomplete formula redemption (i.e., not using the benefit) by WIC participants was 342% higher during the formula shortage relative to pre-pandemic, potentially due to having trouble finding formula that met the package requirements, and many WIC birthing individuals pursued other solutions [17, 57]. Breastfeeding may have reduced formula-related stress but there was minimal discussion amongst low-income individuals in qualitative interviews on this topic. High-income individuals mainly reported breastfeeding and increased food access which had protected them from formula-shortage hardships. Together, these findings demonstrate the benefit of supplemental income, and needed flexibility for existing governmental supplemental nutrition programs to support low-income individuals’ mental health.

Strengths of the current study include the mixed methods analysis, harmonized COVID-19 survey collection, exploration of a peak period in natural disasters (formula shortage, hurricane, COVID-19 pandemic), and sampling strategy to allow for exploration across race and income. Limitations of this study are specific to study design and generalizability. Specific to study design, the quantitative survey and approach could have been improved in four major areas: (1) demographic variables, (2) other natural disasters’ impact, (3) feeding modes, and (4) postpartum timeframe for sampling for quantitative and qualitative components. First, the additional demographic variables (e.g., parity) would have been beneficial for context but may be represented in other macrosystem factors (e.g., childcare challenges) and existing covariates (e.g., education). This study unfortunately did not capture history of depression or anxiety, which we acknowledge can also negatively contribute to mental health during the perinatal period. Second, the coinciding natural disasters and major events were unexpected; the qualitative analysis organically obtained this information and highlights the mixed methods approach strengths. Third, feeding modes and duration in this survey may have also been prematurely assessed as some individuals completing the survey were still 6 months postpartum and continuing these modes of feeding. The later timing of the qualitative interviews allowed birthing individuals to reflect on their past feeding experiences. Fourth, these results may not represent their immediate postpartum mental health outcomes or experiences, which may have produced different results, but also allowed individuals to reflect on major stressful components in the postpartum experience. Even a year or so after giving birth, this postpartum sample reported similar mental health scores (EPDS score: 8.0; GAD-7 score: 6.4), anxiety (26.0%) and depression (23.1%) rates as early pandemic mothers in Canada (EPDS Score: 8.1, GAD-7 score: 5.8) [58], and in other pooled analyses (anxiety: 30.0.%; depression 25.6%) [59]. As for the second major limitation category, the birthing individuals included were confined to a southern U.S. state and findings may not be applicable to all races in the U.S., non-English speaking individuals, other U.S. states, or other countries.

There are four clear future directions for research and practice from this study, related to social policies and support to reduce discrepancies between macrosystem factors. First, further investigation into paid maternity leave policies and their impact on mental health of individuals across race and income may expand upon our findings of supplemental income and improving maternal health. Second, these results support a nimble WIC package with more brand offerings and specialty formula. Flexible options for WIC enrollment and flexibility for food redemption should be continued to minimize individuals’ stress. Third, investigating existing government assistance may expand upon the current study. Translation of “Food As Medicine” initiatives, whereby participants are provided groceries during the week as part of a program to improve health behavior and pregnancy outcomes [60], provide an opportunity to identify the active ingredient (i.e., food access, composition, financial offset) in such government assistance [61]. Finally, this study captured the negative impacts of hurricanes that impacted the Gulf Coast in 2020–2022. Prioritizing recovery and appropriate government assistance during these times may reduce the impact of maternal distress from the event on child development.

Conclusion

The current study documents the mental health impact of an infectious disease pandemic, national formula shortage, and local natural disaster recovery on postpartum birthing individuals of varying race and income. Black and low-income birthing individuals experienced a different COVID-19 pandemic and subsequent mental health outcomes relative to White and high-income counterparts. Postpartum individuals’ employment-related stress was related to poor mental health, and related to receipt of government assistance, housing changes, and employment disruptions. Ultimately, nimble government assistance packages, stable housing, and varied employment and childcare options to reduce employment-related stress are needed to support pregnant individuals during natural disasters. Multi-level supports and social policies are required to reduce adverse macrosystem factors, and racial disparities, to ultimately improve maternal mental health for long-term benefit.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

COVID-19

SaRS COVID-19

EPDS

Edinburgh Postnatal Depression Scale

GAD-7

Generalized Anxiety Disorder-7

QUANT

Quantitative

QUAL

Qualitative

SNAP

Supplemental Nutrition Assistance Program

WIC

Special Supplemental Nutrition Program for Women, Infants, and Children

Author contributions

CLK contributed to this manuscript through conceptualization, data curation, formal analysis, investigation, methodology, project administration, writing – original draft, writing – review and editing. KOG contributed to this manuscript through formal analysis, writing – original draft, writing – review and editing. MD contributed to this manuscript through formal analysis, investigation, writing – original draft, writing – review and editing. BJ, ES and JB contributed to this manuscript through implementing the investigation, and writing – review and editing. MK contributed to this manuscript through conceptualization, funding acquisition, investigation, methodology, project administration, and writing – review and editing. KF contributed to this manuscript through - data curation, formal analysis, writing – original draft, writing – review and editing. EWH contributed to this manuscript through conceptualization, funding acquisition, methodology, writing – review and editing. EFS contributed to this manuscript through conceptualization, funding acquisition, methodology, project administration, supervision, writing – review and editing. LMR contributed to this manuscript through conceptualization, funding acquisition, methodology, project administration, supervision, writing – review and editing.

Funding

This work was supported by the National Institutes of Health [grant numbers: K99HD107158, P20GM144269, R01 NR017644, R01 DK124806, and U54 GM104940]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The funder did not have a role in the design and conduct of the study; data collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data availability

The dataset used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Pennington Biomedical Research Center provided institutional review board and ethics approval. The institutional review board approved the method of collecting consent. Participants provided their consent to the quantitative portion by submitting their response. Prior to the beginning of the survey, participants were given instructions that explained consent, and notified them if they do not consent to this procedure that they should not proceed with the survey. Verbal consent was collected from participants prior to the semi-structured interviews.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.