Abstract
We compared 572 pregnant women (319 first-time mothers) surveyed in spring 2020, during the first wave of COVID-19 lockdowns in the United States, with 99 pregnant women (all first-time mothers) surveyed before the pandemic (2014–2020). Compared with the prepandemic sample, women assessed during the pandemic showed elevated depression, anxiety, and stress and weaker prenatal bonding to their infants. These findings remained significant when restricting the pandemic sample to first-time mothers only and held after controlling for race/ethnicity, education, and pregnancy stage. Average levels of depression and anxiety within the pandemic group exceeded clinically significant thresholds, and women who estimated that the pandemic had more negatively affected their social relationships reported higher distress. However, pandemic-related changes to social contact outside the household were inconsistently associated with mental health and with some positive outcomes (fewer depressive symptoms, stronger prenatal bonding). Given that prenatal stress may compromise maternal and child well-being, the pandemic may have long-term implications for population health.
In spring 2020, the COVID-19 pandemic transformed the global social landscape as the need to maintain social distance quickly reshaped work, travel, and leisure behaviors. In particular, the experience of pregnancy, birth, and new parenthood changed for many families (Poon et al., 2020; Tran et al., 2020). In the early days of the pandemic, women preparing for childbirth—often a time of joyful anticipation—reported fear and uncertainty. There were restrictions on the number of people permitted to attend births, and some women delivered infants alone, without partner or family support (Farewell et al., 2020; Morano & Calleja-Agius, 2020). Pregnant women also experienced disruptions to their prenatal medical care and to their ability to seek social connection. The transition to parenthood often prompts an outpouring of extended family and community support, but many expectant parents relinquished baby showers, bris and christening ceremonies, and other rituals to help to mark their new baby’s arrival.
Although the pandemic engendered societal upheaval, loss of community and connection, and increased health-related anxiety across a broad swath of the population, these experiences may have been particularly acute for expectant parents, given heightened needs for both medical and social support during the transition to parenthood (Saxbe et al., 2018). New parenthood is also recognized as a time of vulnerability for mood and anxiety disorders, which can have pervasive effects on both parent and child well-being (Biaggi et al., 2016; Stein et al., 2014). Exploring impacts of the pandemic on perinatal mental health is thus a continued public-health priority, even after the initial threat of COVID-19 recedes. In the current article, we compare psychological distress levels in 572 pregnant women (319 first-time mothers), surveyed in spring 2020 at the height of social distancing behaviors in the United States, with a prepandemic sample of 99 pregnant women.
Perinatal mental-health disorders are the most common pregnancy complication, and prevalence ranges from 10% to 40% of women during the perinatal period (Biaggi et al., 2016). These disorders have implications for the well-being of not only mothers but also their children, including higher rates of birth complications and admission to the neonatal intensive care unit (Stein et al., 2014). Moreover, risks to child development extend past the perinatal period. Two reviews summarize a large body of work showing increased risks for psychological and developmental problems in children born to mothers with perinatal mood and anxiety disorders (Hoffman et al., 2017; Stein et al., 2014).
Implications of COVID-19 for Perinatal Mental Health
Studies of pandemic-related stress have the potential to extend previous research on prenatal exposure to mass disasters (e.g., natural disasters, terrorist attacks), which have been linked with adverse long-term effects on maternal and infant health, including maternal mood disorders and infant temperament and development (Harville et al., 2010). The COVID-19 pandemic is a particularly unusual population-scale event; although it entails acute threat to life and loss of life, as would a natural disaster, it also occasioned a prolonged period of behavioral change. Natural disasters often catalyze an increase in social affiliation and support as communities come together to grieve and rebuild. But the pandemic necessitated sudden and often dramatic reductions in social contact. Compounding this, pregnant women are categorized as high risk and may have been advised to quarantine strictly (Fakari & Simbar, 2020). Social support has been identified as a key buffer of risk for perinatal mental-health disorders and a significant, consistent predictor of both mental and physical health in women during the perinatal period (Biaggi et al., 2016; Emmanuel et al., 2012), yet during the COVID-19 pandemic, “safer-at-home” orders limited access to social support and connection. Moreover, the pandemic also introduced additional multifaceted stressors: caregiving and bereavement demands, economic losses, and disruptions to work and medical care (Pfefferbaum & North, 2020). Thus, the acute decrease in social support and additional stressors of the COVID-19 pandemic may present unique risks for perinatal mental health compared with other large-scale disasters.
A few preliminary studies have reported that women experiencing pregnancy during the pandemic described low mood (Milne et al., 2020) and heightened anxiety (Corbett et al., 2020) in the context of the unique stressors of COVID-19. Specifically, Lebel and colleagues (2020) found substantially elevated depression and anxiety symptoms in a sample of almost 2,000 pregnant women in Canada, of whom 37% reported clinically relevant symptoms of depression and 57% reported clinically relevant symptoms of anxiety. Another Canadian survey study of 520 pregnant women reported similar rates; 40.7% of women reported depression above the clinical threshold, and 72% of women reported moderate to high anxiety, significantly higher rates than the women’s estimated prepandemic mental health, although with the caveat that the prepandemic estimates may have been affected by retrospective reporting bias (Davenport et al., 2020). A Stanford study found even higher rates of depressive symptoms; 51% of pregnant women in the San Francisco Bay Area during the pandemic scored above the clinical cutoff compared with 25% of a demographically matched prepandemic sample (King et al., 2020). Moreover, a Turkish study found that 35.4% of women endorsed clinically significant depressive symptoms in the early months of the COVID-19 pandemic, significantly higher than in prepandemic control subjects (Durankuş & Aksu, 2020). Another Turkish study conducted in June and July 2020 found that 64.5% and 56.3% of women reported clinically significant symptoms of anxiety and depression, respectively (Sut & Kucukkaya, 2021). Only one study to date has focused exclusively on mental health in first-time mothers. This study found high levels of distress among 49 U.S. women expecting their first child who were sampled between June and July 2020 (McMillan et al., 2021). In sum, reports from the first wave of the pandemic indicate elevated prenatal distress compared with existing prepandemic samples.
Mothers’ ability to bond with their unborn infants during the COVID-19 pandemic has not yet been investigated. Prenatal stress and distress have been associated with poorer prenatal bonding, which in turn may negatively affect the postpartum parent–child relationship (Glover & Capron, 2017; Walsh et al., 2014). In addition, more work is needed to understand the mechanisms of pandemic-related distress, such as changes in social contact and connectedness associated with lockdowns. Studies of COVID-19 impacts can also benefit from specificity in terms of timing and sample. For example, the May 25, 2020, murder of George Floyd sparked a racial-justice movement that changed the cultural landscape just as COVID-19 was reshaping social behavior. In the current study, we focus only on data collected within a 6-week period, between April 6, 2020, and May 24, 2020, to limit the potentially confounding effects of these and other geopolitical events. In addition, many published studies of perinatal women during the COVID-19 pandemic have not considered parity, but first-time expectant mothers may have had a different experience of the pandemic than women who were already mothers given that the pandemic occasioned child-care disruptions that may have affected women with other children at home.
The Current Study
We report on a survey of expectant parents conducted in spring 2020, during the first wave of COVID-19 lockdowns in the United States. We hypothesized that, compared with a prepandemic sample of pregnant women, the pandemic sample would report heightened levels of psychological distress (depressive symptoms, anxiety symptoms, and perceived stress) and decreased levels of prenatal bonding. We also tested whether pandemic-related changes to social contact and social relationship quality would explain these differences. Our comparison sample focused on first-time mothers; given the possibility that parity might affect the experience of pregnancy during the pandemic, we tested all hypotheses twice, first with the full pandemic sample and then with first-time mothers only.
Method
Prepandemic comparison data were drawn from a sample of 100 couples recruited in Los Angeles, California, between March 2014 and early February 2020 and followed from pregnancy across the first year postpartum. All procedures were approved by a university institutional review board. Eligible couples were expecting their first child, cohabiting, and able to complete study measures in English. Couples completed two lab visits and a postbirth hospital visit; data included in this article are drawn from the questionnaire battery administered to pregnant women at the prenatal visit, which occurred in mid to late pregnancy. Couples were recruited primarily through social media (e.g., Facebook pregnancy/parenting groups) and word of mouth. Of the 100 first-time mothers recruited through this project, 99 had complete data on all measures and were included in this article.
In April 2020, the questionnaire battery used in this study was adapted into an online survey for expectant parents, the Coronavirus, Health, Isolation, and Resiliency in Pregnancy (CHIRP) study, launched on April 6, 2020. According to smartphone GPS data, the week of April 7, 2020, represented the peak of “sheltering in place” behavior in the United States, during which Americans spent 93% of their time at home (Schaul et al., 2020). For the present analyses, we focused on responses collected within the first 6 weeks after the survey was launched. (We excluded 24 pregnant mothers who responded to our survey after May 24, 2020, to avoid potentially confounding effects of the May 25, 2020, murder of George Floyd, which occasioned widespread racial-justice protests.) We dropped an additional 45 participants who completed less than 28% of the survey. The resulting sample size for analyses was 572, of which 319 were first-time mothers.
Participants provided consent and completed a 20- to 30-min questionnaire through the Qualtrics platform (http://www.qualtrics.com). All study procedures were approved by the university institutional review board. Expectant parents were recruited via social media, including posts on Twitter, Facebook, and Instagram; online pregnancy/parenting groups and message boards; and paid advertising on Facebook and Instagram. Participants were compensated through entry into a gift-card lottery. Per institutional review board guidance, after completing the survey, participants were offered a downloadable list of mental-health resources, including substance-abuse, suicide-prevention, and domestic-violence hotlines.
Participants
Participants in the prepandemic sample were recruited across the Los Angeles area and were diverse and representative of the region, expecting their first child, and between 20 and 35 weeks pregnant at study entry. The pandemic sample was mostly based in the United States (97.7%) and could participate at any point during pregnancy. Most participants were married (84% of the prepandemic sample and 88% of the pandemic sample); rates of marriage or cohabitation did not differ across the two groups. As shown in Table 1, the prepandemic sample was more racial/ethnically diverse and slightly more educated than the pandemic sample. Furthermore, the prepandemic sample was, on average, 3 to 4 weeks further along in their pregnancies and 1 year older than the pandemic sample. We adjusted for these differences by controlling for race/ethnicity, education, days pregnant, and maternal age in all analyses. There were no significant moderating effects of race/ethnicity on levels of depression, anxiety, stress, and bonding when comparing the two samples.
Participant Characteristics
Note: Values are percentages with ns in parentheses unless otherwise indicated.
Measures
COVID-19 impacts
We asked about exposure to COVID-19, COVID-19-related bereavement, and additional behavioral and other impacts (for more details, see Morris et al. 2021). We operationalized changes in social connection in two ways. First, participants were asked, “Overall, what impact has COVID-19 had on your social relationships?” and responded on a 5-point scale (1 = very negative, 3 = no impact, 5 = very positive). Next, participants were also asked, “As compared to before COVID-19, how much total contact (including in-person, phone, or online) do you have with the following people?” Participants were asked to respond to this question on a 5-point scale (1 = much less, 5 = much more) in relation to their (a) neighbors/community members, (b) coworkers, (c) friends, (d) families, and (e) partners. Responses across the first three groups were averaged to create a total score reflecting social contact outside the household.
Depression
The Beck Depression Inventory (BDI-II; Beck et al., 1996), a 21-item self-report questionnaire that has been used in many studies of depression, assesses mental and somatic complaints related to depression, including loss of pleasure and changes to sleep and appetite. Respondents rate items on a 4-point scale (0 = not at all, 3 = severely). Scores are summed so that higher scores indicate worse depressive symptoms.
Anxiety
The state subscale of the State-Trait Anxiety Inventory (Spielberger, 2010), a well-validated and widely used 20-item questionnaire measuring acute anxiety, uses a 4-point scale (0 = not at all, 3 = very much so) to rate items such as “I feel nervous” and “I am tense.” Responses summed for a total score can range from 20 to 80, and higher scores reflect higher state anxiety.
Stress
The 10-item Perceived Stress Scale (Cohen et al., 1994) is the most widely used measure of self-reported stress. Responses for each question (0 = never, 4 = almost always) are summed, and higher scores reflect higher perceived stress.
Prenatal bonding was measured using the Antenatal Attachment Scale (AAS; Condon, 1993). The AAS, a 19-item self-report questionnaire, assesses expectant parents’ thoughts and feelings about their developing baby over the past 2 weeks. All responses are on a 5-point scale and are averaged into a total score; higher scores reflect stronger prenatal bonding (example items: “I have thought about or been preoccupied with the baby” [1 = not at all, 5 = almost all the time]; “When I have spoken about or thought about the developing baby I got emotional feelings which were” [1 = very weak, 5 = very strong]).
Analyses
Analysis of covariance was used to assess differences in mental health and prenatal bonding in the prepandemic and pandemic samples of pregnant women. Partial correlations were used to assess the relationship between COVID-19-related changes in social connection and prenatal mental health and bonding within the pandemic sample. All analyses controlled for number of days pregnant, identification as a racial/ethnic minority, maternal age, and educational attainment. Because not all participants completed every questionnaire item, sample sizes differ slightly for each analysis, as indicated in the tables and text. All analyses were run twice, first with the full pandemic sample and then with first-time mothers only.
Results
For descriptive statistics for both the prepandemic and pandemic samples, see Table 1. At the time of the CHIRP survey, direct impacts of the pandemic were relatively low: Approximately 95% of respondents had not experienced either a suspected or confirmed case of COVID-19 or a COVID-19-related death of someone close to them. However, 97.7% of women reported that their community had issued stay-at-home orders restricting large gatherings and closing nonessential businesses. The majority of women also reported that the pandemic had a negative impact on their social relationships; about 60% reported negative impacts on their social relationships and reduced social contact with individuals outside their household, and almost 90% indicated an overall decrease in contact.
Mental health
As shown in Table 2, women’s self-reported depressive symptoms, state anxiety, and perceived stress were all significantly greater in the pandemic group compared with the prepandemic group, both across the full sample and in first-time mothers only. As shown in Table 1, levels of depressive symptoms and perceived stress were approximately 1 SD higher in the pandemic sample compared with the prepandemic baseline; for anxiety, the difference was about 1.8 SD. Within the pandemic sample, the mean BDI-II score of depressive symptoms (15.46) exceeded the threshold (14) that has been used to indicate clinically significant depression (Smarr & Keefer, 2011). Likewise, the mean anxiety score in the pandemic group exceeded the clinical threshold that has been used with perinatal populations (Tendais et al., 2014). In comparison, prepandemic means fell far below clinically significant thresholds. Estimated marginal means that reflect these differences, after adjusting for covariates, are shown in Table 3.
Analysis of Covariance and for Pandemic and Prepandemic Samples on Measures of Depressive Symptoms, Anxiety Symptoms, Perceived Stress, and Prenatal Bonding
Note: BDI-II = Beck Depression Inventory (Beck et al., 1996); STAI-S = State-Trait Anxiety Inventory (Spielberger, 2010); PSS = Perceived Stress Scale (Cohen et al., 1994); AAS = Antenatal Attachment Scale (Condon, 1993).
p < .05. **p < .01. ***p < .001.
Estimated Marginal Means for Pandemic and Prepandemic Samples on Measures of Depressive Symptoms, Anxiety Symptoms, Perceived Stress, and Prenatal Bonding
Note: Estimated marginal means are based on the analysis of covariance analyses shown in Table 2. BDI-II = Beck Depression Inventory (Beck et al., 1996); STAI-S = State-Trait Anxiety Inventory (Spielberger, 2010); PSS = Perceived Stress Scale (Cohen et al., 1994); AAS = Antenatal Attachment Scale (Condon, 1993).
As shown in Table 4, mothers reporting more negative pandemic-related impacts on their social relationships also reported higher levels of depressive symptoms, anxiety, and stress (within both the full sample and first-time mothers only). Reductions in social contact because of the pandemic were correlated with stress in the full sample of mothers. However, this association became nonsignificant when the analysis was restricted to the sample of first-time mothers only. Counterintuitively, first-time mothers reporting reduced social contact outside the household reported fewer depressive symptoms. Social contact was not associated with anxiety in either the full sample or the first-time-mother sample.
Partial Correlations Between Social Impact and Contact and Mental Health and Bonding
Note: Correlations are adjusted for maternal age, days pregnant, racial/ethnic minority status, and educational attainment. BDI-II = Beck Depression Inventory (Beck et al., 1996); STAI-S = State-Trait Anxiety Inventory (Spielberger, 2010); PSS = Perceived Stress Scale (Cohen et al., 1994); AAS = Antenatal Attachment Scale (Condon, 1993).
p < .05. **p < .01. ***p < .001.
Prenatal bonding
Also shown in Table 2 is that women who were pregnant during the pandemic reported weaker prenatal bonding with their unborn infant compared with women who completed the same bonding measure before the pandemic. This difference amounted to approximately two thirds of a standard deviation, as shown in Table 1. Estimated marginal means reflecting differences in prenatal bonding, after adjusting for covariates, are shown in Table 3. Mothers’ estimated impact of the pandemic on their social relationships was not associated with their reports of prenatal bonding, as shown in Table 4. However, pandemic-related decreases in total social contact outside the household were associated with stronger prenatal bonding. These findings were consistent across the full sample and within first-time mothers only.
Discussion
We found that women experiencing pregnancy during the COVID-19 pandemic reported significantly higher levels of psychological distress and weaker prenatal bonding with their babies compared with a prepandemic sample of pregnant women. Strikingly, within our pandemic sample, women’s depression and anxiety exceeded clinical thresholds. Women who reported that the pandemic had a more negative impact on their social relationships also reported more depression, anxiety, and stress, which points to pandemic-related decreases in social connectedness as a potential mechanism explaining differences between pandemic and prepandemic cohorts. However, pandemic-related changes in social contact outside the household were not consistently associated with mental health. In fact, pandemic-related reductions in social contact were actually associated with fewer depressive symptoms in first-time mothers and with stronger prenatal bonding with infants in all mothers. Thus, mothers’ perceptions of negative social impacts may be more strongly associated with their distress, whereas actual changes in social contact outside the household yielded more mixed results. The ability to reduce social contact may reflect occupational privilege, which is typically associated with better mental-health and pregnancy outcomes. Our results suggest that research on social connection and mental health during the pandemic should carefully consider global perceptions of negative impact compared with more granular reports of social contact with others.
Given that prenatal mental-health problems are associated with a wide array of negative developmental outcomes for children (Stein et al., 2014), this evidence for heightened distress in pregnant women during the pandemic is of concern. Prenatal stress and distress may have long-term effects on both maternal and infant health, as other research on population-scale disruptions has suggested. Therefore, these findings may indicate risk to both mothers and infants in the postpartum period and beyond. Our results were generally consistent across first-time mothers compared with all mothers, which suggests that the pandemic had negative impacts on pregnant women across the board.
In addition to significantly higher levels of mental-health symptoms in our pandemic cohort, we also found that women reported weaker prenatal bonding with their unborn babies. This finding held after adjusting for duration of pregnancy and highlights the potential for future challenges for this population given that prenatal mother-child bonding has been associated with future bonding and parenting behaviors (Maas et al., 2016). We found that decreases in the amount of social contact with community members, coworkers, and friends was associated with stronger feelings of prenatal bonding. It is possible that decreased engagement with others may result in more time spent thinking about and connecting with the unborn baby. Although we found a relationship between social contact and prenatal bonding, we did not find a similar relationship between prenatal bonding and perceptions of how the pandemic affected women’s social relationships.
Conclusions
In this article, we sought to highlight perinatal mental health and prenatal bonding as particular areas for concern and attention in the wake of the COVID-19 pandemic. We sought to further extend the literature by exploring prenatal bonding and examining the impact of the COVID-19 pandemic on social relationships and social contact as a potential mechanism for these effects. Note that our pandemic sample was recruited during the first wave of COVID-19 lockdowns, when social-distancing measures were most widely adopted in the United States, which allowed us to assess the impacts of social isolation more closely than in other samples with a larger range of participation dates. Furthermore, this is one of the first studies to assess rates of mental health and prenatal bonding in pregnant women transitioning to motherhood during the COVID-19 pandemic.
Given the known detrimental impacts of poor maternal mental health on both mother and infant outcomes (Harville et al., 2010), our findings of heightened risk for perinatal mental-health problems, poorer prenatal bonding, and decreases in social connection during the pandemic are of great concern. Efforts are needed to support pregnant and postpartum women even as the pandemic resolves. Research should continue to track children born during and shortly after the COVID-19 pandemic to assess potentially enduring effects of pandemic-related prenatal stress.
Limitations
These data were collected cross-sectionally in spring 2020, shortly after many COVID-19 restrictions were implemented. Responses to this online survey came from a well-educated convenience sample recruited across social media platforms. The comparison group was drawn from an existing study and comprised only first-time mothers. Although this is a strength for our analyses of first-time mothers only, it is a limitation of our larger analyses with the full pandemic sample including non–first-time parents. In addition, our pandemic group was less racially/ethnically diverse than our prepandemic sample and was not limited to geographic region, compared with our prepandemic group, which was recruited from the Los Angeles area. Furthermore, the prepandemic group participated in a more time-intensive, in-person study, compared with the pandemic group, which participated in a 45-min online survey. Thus, although we controlled for race/ethnicity, age, days pregnant, and educational attainment, our samples are not equivalent in background or motivation to participate in research. Given well-documented associations between higher socioeconomic status and perinatal mental health in the United States, a more ethnically/racially and socioeconomically diverse sample might reveal even more striking levels of mental-health vulnerability.
Future directions
There is a clear need to follow up and explore the long-term implications of the perinatal experience during the COVID-19 pandemic. Previous work highlights the pandemic’s disproportionate impact on low-income communities and communities of color and suggests that these impacts magnify preexisting disparities in maternal and infant health (Dongarwar et al., 2020). Therefore, more work is needed to explore the impact of the pandemic in more racially/ethnically diverse samples, and the need for mental-health services is likely even more acute than reflected in the current data. Future work should follow families longitudinally to better understand the effects of pandemic-related stress across time, explore effects on fathers and children in addition to mothers, and assess potential positive changes linked with the pandemic (i.e., more time at home with baby, greater ease of breastfeeding). In sum, research on the mental-health and social dynamics of the COVID-19 pandemic can shed light on stress, health, and resilience in pregnancy and beyond.
Footnotes
Transparency
Action Editor: Jennifer L. Tackett
Editor: Jennifer L. Tackett
Author Contributions
A. R. Morris and D. E. Saxbe developed the study concept and study design, oversaw data collection, and performed data analyses. Both authors contributed to the writing and revisions of the manuscript and approved the final manuscript for submission.
