Prenatal stress, health, and health behaviours during the COVID-19 pandemic: An international survey

Background Pregnant women’s stress, mental and physical health, and health behaviours can have important implications for maternal and child health outcomes. Aim To examine pregnant women’s levels of stress, mental and physical health, and health behaviours during the COVID-19 pandemic. Methods A cross-sectional survey was conducted online, with recruitment and data collection occurring between 16/6/20 and 17/7/20. Participants were pregnant women recruited via online pregnancy/parenting communities. Participants self-reported their levels of general stress, pregnancy-specific stress and COVID-19 related stress, mental and physical health, general health behaviours, and COVID-19 related health behaviours. Findings 573 pregnant women participated in the survey. Participants were most commonly resident in the United States (42.6%, n = 243), Ireland (41.2%, n = 235) or the United Kingdom (10%, n = 57). The majority (80.0%, n = 457) were married and educated to degree level or above (79.3, n = 453). Pregnant women reported high levels of pregnancy-specific and COVID-19-related stress, and low levels of mental and physical health, during the pandemic. Encouragingly, pregnant women in this study generally reported high levels of adherence to public health advice and pregnancy health behaviours. Stress and general mental health outcomes were best predicted by well-being factors (including stress and social support). Health impairing behaviours (e.g. poor diet) were predicted by both well-being and demographic factors. Discussion Interventions targeting pregnancy- and pandemic-specific stress at the population level will be essential to support mental health and minimise adverse outcomes for women and children during the pandemic.


Introduction
Many women experience disruptions to mental health during pregnancy. It is estimated, for example, that between 7% and 30% of women exhibit symptoms of depression during their pregnancies [1]; between 18.2% and 24.6% report anxiety [2]; and up to 84% experience stress [3]. Depression, stress and anxiety are related yet distinct psychological constructs [4,5] which have been associated with a range of poor outcomes including increased risk of premature birth and reduced length of gestation [6] and disruptions to hormonal regulation in children [7]. Prenatal mental health issues may also be associated with poorer physical health and health behaviours, with further direct and indirect effects on maternal and infant outcomes [8]. For instance, low physical activity, poor diet, and substance use during pregnancy are associated with increased risk of adverse prenatal [9], health [9][10][11][12], and child developmental [13] outcomes.
Prenatal stress and mental health issues can arise from multiple factors, including from experiencing stressful life events like the COVID-19 pandemic [14,15]. COVID-19, which was declared a pandemic by the World Health Organisation in March 2020 [16], is a significant global public health crisis. The pandemic and pandemic-related restrictions, such as lock-downs and quarantining, are already demonstrating significant negative effects for psychological well-being [17]. In frontline workers and general populations, COVID-19 has been associated with increased risks for depression, anxiety [17][18][19] vicarious trauma, and sleep disruptions [17,19]. High rates of substance use and suicidal ideation have also been reported among a general population during the pandemic [17].
Pregnant women are potentially at increased risk of adverse outcomes during the pandemic due to compounding pregnancyspecific concerns. For example, women may experience distress due to fear of coronavirus infection during pregnancy and/or in fetuses and children [20]; lack of clarity about pregnancy-specific risks of COVID-19 and which precautions are most effective at reducing infection risks to infants [21]; and changes to maternity services and birth plans due to public health restrictions [20,21]. In addition, COVID-19 restrictions have placed significant limitations on women's levels and types of social support, which has the potential to further strain their coping resources [20,22]. Given the proposed stress-buffering effects of social support, this may also increase women's risk of adverse outcomes both during pregnancy and beyond [23].
To date, some negative outcomes of the COVID-19 pandemic have been demonstrated in pregnant populations. For instance, one Canadian study reported increased incidences of anxiety, depression, post-traumatic stress, and dissociative symptoms among women pregnant during the COVID-19 pandemic, compared to a pre-pandemic cohort [24]. Another study reported increased levels of depressive symptoms in pregnant women in China after COVID-19 was declared an epidemic there in late January, again compared to a pre-pandemic baseline in the same population [25]. Reduced perceived social support has also been found among women pregnant in Ireland during the pandemic [26], and women also self-reported lower physical activity, compared to those pregnant before the pandemic [27]. These early findings suggest that women pregnant during the pandemic may be at risk for impaired prenatal psychological well-being and health behaviours. Examinations of stress have been limited however, with studies tending to focus on depression and/or anxiety symptoms. Similarly there has been little to no examination of prenatal health behaviours. Given the importance of maternal mental health, stress and health behaviours for maternal and child outcomes, examining these factors during the COVID-19 pandemic is essential to inform approaches to supporting pregnant women during this, and potential future, pandemics. The aim of this study is to examine pregnant women's levels of stress, mental and physical health, and health behaviours during the COVID-19 pandemic.

Design
A cross-sectional online survey of pregnant women was conducted. This study forms part of a larger project examining prenatal stress, social support and health behaviours [28].

Ethical considerations
Ethical approval was granted by the [redacted] Research Ethics Committee and the Clinical Research Ethics Committee at [redacted]. Funding organisations were not involved in conceptualising the study, collecting data, analysing/reporting results, or any decisions related to publication.

Participants
Participants in this study were pregnant women over 18 years of age. No exclusions were made based on nationality or current place of residence. Participants were recruited between June 16th, 2020 and July 17th, 2020. All participants were recruited online via public postings on forums associated with pregnancy websites (e.g What To Expect, Mumsnet), and on social media pregnancy/ parenting communities (e.g. Facebook, Twitter, Instagram, Reddit); a full list of targeted communities is available in Supplement 1. Participants were provided with a brief study description and link to the study survey, containing the full study information leaflet and consent form. Participants provided consent prior to completing the survey.

Measures
The survey included open and closed questions on demographic factors, stress, mental and physical health, health behaviours, social support, and perceived satisfaction and quality of antenatal care.
Details and findings in relation to social support, quality of antenatal care, stress-reduction strategies, and experiences of distress during the pandemic are presented in our complementary paper [29]. As such they are not reported as outcomes in the current paper but are presented here for completeness.

Demographic factors
Data on the following demographic factors were collected: age, marital status, education, household income, ethnicity, nationality, country of residence, weeks of pregnancy, and number of previous pregnancies.

General stress
General stress was measured using the 4-item Perceived Stress Scale Short Form (PSS-4) [30]. This validated scale asks participants to report frequency of experiencing stress on a 5-point Likert scale ranging from 0 = "never" to 4 = "very often". Items include "How often have you felt that you were unable to control the important things in your life?" In the current study the PSS-4 had a reliability coefficient of α = .75.

Pregnancy-specific stress
Pregnancy-specific stress was measured using the validated 17item Revised Prenatal Distress Scale (NuPDQ) [31]. Items ask participants to report whether they are feeling "bothered, upset, or worried at this point in their pregnancies" about pregnancyspecific issues such as concerns about their pregnancy or baby, and changes to self-image and social relationships. Items are rated on a 3-point Likert scale ranging from 0 = "not at all" to 2 = "very much." In this study the NuPDQ had a reliability coefficient of α = .76.

COVID-19 related stress
At the time of this study there was no published validated scale available to measure COVID-19 related stress during pregnancy. As such, and guided by available evidence from a recent study by Zhong et al. [32], 6 items were developed to examine women's concerns about the COVID-19 pandemic in this study. Participants were asked to indicate how worried they were about the coronavirus overall using a single item, and 5 items assessing concerns about coronavirus infection, being able to sleep because of coronavirus concerns, restrictions in place due to the coronavirus, and access to antenatal care due to the coronavirus. Items were measured using a 3-point Likert scale ranging from 0 = "not at all" to 2 = "very much." The internal validity for this measure was α = .77.

Mental and physical health
Mental and physical health were measured using the mental and physical domains of the validated 12-item Short Form Survey (SF-12) [33]. Participants indicated their general perceived health on a 5-point Likert scale (ranging from 1 = "excellent" to 5 = "poor"). Participants were also asked to indicate health-related limitations to activities using a 3-point Likert scale (ranging from 1 = "Limited a lot" to 3 = "Not at all limited"), and by answering "yes" or "no" to questions about specific limitations. Participants indicated how much pain was interfering with their ability to work using a fivepoint Likert scale (ranging from 1 = "not at all" to 5 = "extremely").
Participants were also asked about the frequency of emotional experiences, such as "feeling down-hearted or blue" and whether their health interfered with social functioning, on a 6-point Likert scale (ranging from 1 = "all of the time" to 6 = "none of the time").

Prenatal health behaviours
Frequency of engaging in prenatal health behaviours was measured using the validated 20-item Prenatal Health Behaviour (PHBS) scale [34]. Health behaviours were grouped into a healthpromoting behaviours subscale (α = .72) and a health-impairing behaviours subscale (α = .55) as per previous guidance [35]. Healthpromoting behaviours included exercise, stretching, getting enough sleep, taking vitamins, and maintaining a good diet. Health-impairing behaviours included skipping meals, consuming snack foods instead of balanced meals, consuming drinks with caffeine, and smoking cigarettes. Items are measured on a Likert scale ranging from 0 = "never" to 4 = "very often". The internal consistency of the health-promoting subscale was α = .72. The internal consistency of the health-impairing subscale was α = .55, which, though lower than desired, is in keeping with previous examinations [34].

COVID-19 related health behaviours
Six items were developed to assess women's adherence to COVID-19 related health behaviours. Women were asked to report whether they were using a mask outside the home and in hospital, avoiding crowded places, leaving the house only when essential, shaking hands or making close contact with people, and disinfecting frequently touched objects, by answering either "yes" or "no."

Social support
Social support was measured using the 12-item Multidimensional Perceived Social Support Scale (MSPSS) [36]. This validated scale asks women rated their perceived degree of support from significant others, family and friends, by rating statements (for example: "There is a special person with whom I can share my joys and sorrows") on a 7-point Likert scale ranging from 1 = "very strongly disagree" to 7 = "very strongly agree." The overall scale reliability was α = .96; the reliability of the significant other, family, and friends subscales were α = .98, α = .94, and α = .96 respectively.

Open-ended questions
Women were also invited to provide additional qualitative feedback on their prenatal stress experiences including what they found stressful, what stress-reduction strategies they engaged in, and what services and supports they use/want/need during their pregnancies.

Analysis
Descriptive statistics were used to summarise information on participant demographics, stress, mental and physical health, social support, health behaviours, and COVID-19 related behaviours. Associations between demographic factors, stress, mental and physical health, social support, and all health behaviours were examined using Spearman's rank order correlations. Hierarchical multiple regressions were conducted to examine the role of sociodemographic and well-being factors (stress, mental and physical health, and social support) in predicting general stress, pregnancy-specific stress, COVID-19 related stress, mental and physical health, and health behaviours. Sociodemographic factors were entered at Step 1 of the hierarchical regressions; followed by well-being factors at Step 2 to determine the impact of well-being factors when controlling for sociodemographic factors. Dummy variables were created from multilevel categorical variables for inclusion in the regression models. Chi square tests of independence were conducted to examine potential demographic differences in COVID-19 related health behaviours. Due to the use of multiple inferential tests, Bonferroni corrections were applied with a more stringent significance value, calculated by dividing an alpha of .05 by the number of regression tests (n = 7), which results in a significance value rounded up to p < .01. This more stringent significance value was used in all analyses.
Participants reported high frequency of engagement in health-promoting behaviours and low engagement in health-impairing behaviours. For example, most participants reported abstaining entirely from alcohol (87.2%) and cigarettes (96.2%); as such these variables were excluded from further analyses. Participants also reported high levels of adherence to COVID-19 public health advice; for instance, 95.6% of participants reported that they had stopped shaking hands or making close contact with others, while 76.3% reported wearing masks when leaving home (see Table 3).
Findings of associations between stress, mental and physical health, and health behaviours using Spearman's correlations are presented in Supplement 2.
Predictors of stress, mental and physical health, and health behaviours Hierarchical multiple regressions were conducted to examine the role of sociodemographic and well-being factors in predicting stress, mental and physical health, and health behaviours. Preliminary analyses confirmed normality, linearity, multicollinearity and homoscedasticity. A summary of regression outcomes is provided in Table 4 below; full details of each regression are available in Supplements 3-5.

COVID-19 related stress
Sociodemographic factors were entered at Step 1 and explained 4.9% of the variance in perceived stress. After entry of the wellbeing factors (general stress, pregnancy-specific stress, social support, and mental and physical health) at Step 2, the total variance explained by the model was 31.1%, F (23, 474) = 9.31, p < .001. The well-being factors explained an additional 26.2% of the variance in stress after controlling for sociodemographic factors; R squared change = .26, F change (5, 474) = 36.07, p < .001. In the final model, only general stress (beta = -.41, p < .001) and pregnancy-specific stress (beta = .17, p = .002) predicted COVID-19 related stress. See Supplement 3.

Pregnancy-specific stress
Sociodemographic factors entered at Step 1 explained 7.5% of the variance in pregnancy-specific stress. After entry of the COVID-19 related stress, general stress, social support, and mental and physical health at Step 2, the total variance explained by the model was 39.4%, F (23, 474) = 13.42, p < .001. The well-being factors explained an additional 32% of the variance in stress, after controlling for sociodemographic factors; R squared change = .32, F change (5, 474) = 50.07, p < .001. In the final model, the number of children women had (beta = -.21, p < .001), COVID-19 related stress (beta = .36, p < .001), and mental health (beta = À.23, p < .001) predicted pregnancy-specific stress. See Supplement 3.

Perceived physical health
Sociodemographic factors, entered at Step 1, explained 7.6% of the variance in physical health measured by the SF-12. After entry of the stress, social support, and mental health measures at Step 2,

Demographic differences in COVID-19 health behaviours
Chi square tests conducted to examine potential demographic differences in COVID-19 related health behaviours indicated that wearing a mask when leaving home was significantly associated with relationship status (χ 2 (3) = 19.49, p < .001): married women were most likely to wear a mask. Wearing a mask was also associated with country of residence (χ 2 (3) = 87.30, p < .001), with women in America more likely to report wearing a mask than their counterparts in Ireland or the United Kingdom. Wearing a mask when in hospital/maternity services was also associated with the country of residence (χ 2 (3) = 41.08, p < .001): women in America were most likely to do so. Going to crowded places was associated with education (χ 2 (5) = 22.92, p < .001), women with a postgraduate degree being least likely to go to crowded places; and with country of residence (χ 2 (3) = 24.34, p < .001), women in Ireland being most likely to go to crowded places. Disinfecting and cleaning, leaving home only when essential, and avoiding handshakes/close contact were not significantly associated with any demographic factors. See Supplement 6 for complete details of findings regarding demographic differences in COVID-19 related health behaviours.

Discussion
This study examined pregnant women's levels of stress, mental and physical health, and frequency of health behaviours during the COVID-19 pandemic. Building upon existing evidence that pregnant women are experiencing more anxiety, depression, and post-traumatic stress during the pandemic [24,25], this study found high levels of pregnancy-specific stress and COVID-19 related stress, but low levels of mental and physical health and social support. The majority of women in this study engaged in recommended prenatal health behaviours and COVID-19 related behaviours. Psychological outcomes and health-promoting behaviours were mostly predicted by well-being variables including stress and social support; sociodemographic factors had limited impact on outcomes.
Our findings indicated that during the pandemic, pregnant women in our study were experiencing normative levels of general stress [37], but reported high levels of pregnancy-specific stress [31] and COVID-19 related stress. That prenatal and COVID-19 stress were heightened in this sample, but general stress was not, suggests that the experience of being pregnant during the pandemic may function as a contextual stressor. This is supported by our findings that nearly all participants reported feeling somewhat or very concerned about COVID-19/related restrictions, many were concerned about their babies becoming unwell due to the coronavirus, and many reported disruptions to sleep associated with thinking about the coronavirus. That is in line with previous evidence that the coronavirus has prompted pregnancy-specific concerns [20,21]. More generally, our findings of increased COVID-19 related stress are in keeping with previous examinations that indicated high levels of anxiety during the pandemic [24] as stress and anxiety are highly related, though distinct, constructs [5]. These findings provide insight on the specific content of women's stress during the pandemic, and specifically suggest that stress reduction interventions which target pregnancy and coronavirusrelated concerns may be particularly useful.
In addition, although significant differences in stress were not previously observed between women pregnant before or during the pandemic in Ireland [26], the current paper did not find any sociodemographic predictors of any stress outcomes in this international sample. This suggests that while experiences of stress during the pandemic may be contextually specific, they may not be place-or demographic-specific, a finding which suggests that population-level approaches to addressing and supporting prenatal mental health may be effective in the pandemic context. However, given the overall homogeneity of this sample in terms of ethnicity and education, and in light of existing evidence of the relationship between sociodemographic factors and mental health concerns [1], this likely bears further examination.
In this sample, the strongest predictors of stress outcomes were other types of stress and general mental health. Stressors identified by participants included concerns about COVID-19 infection during pregnancy and/or in infants, limitations on travel and social contact imposed by public health restrictions, and limitations on access to prenatal care as a result of public health measures. Additional qualitative feedback from this sample (presented in a complementary paper) suggests that changes to antenatal care, conflict with social support networks, and feelings of uncertainty and isolation were also important sources of stress for this population [29]. This suggests that the development and implementation of prenatal mental health supports and interventions should include a specific focus on pregnancy-and COVID-19-related concerns. Further, though our findings did not suggest an influence of demographic factors, predisposing psychological factors should be considered in the development and implementation of interventions.
In line with previous findings about the relationship between prenatal stress and health outcomes [8], general stress was shown to have a significant impact on both mental and physical health in this sample. Notably, pregnancy-specific stress was also a significant predictor of mental health, suggesting that pregnancy-related concerns may put women at increased risk for mental health impairment and could be useful targets for mental health intervention in pregnant populations. This may be of particular significance as increased mental health symptoms have been reported in pregnant populations since the beginning of the pandemic [24,25]. In addition, our findings suggest a bi-directional relationship between mental and physical health. Previous findings have demonstrated associations between prenatal stress and engagement in health behaviours [8]. However, our findings indicate that despite high levels of pregnancy related and COVID-19 related stress, women in this study reported high adherence to health-promoting behaviours (e.g. eating healthily, exercising, taking vitamins). Similarly, women reported low engagement in health-impairing behaviours (e.g. alcohol consumption and smoking). This finding differs from two recent studies demonstrating lower reported physical activity during the pandemic [26,27]. It is important to note that high levels of selfreported appropriate health behaviours may reflect socially desirable responding in the current study. Unlike stress outcomes, sociodemographic factors including education, number of children, and country of residence, were predictors of healthpromoting and health-impairing prenatal behaviours in the current study. The finding that the number of children women have impacts health behaviours is likely related to childcare and caring responsibilities; as there is evidence that changes to schooling and childcare provision during the pandemic have placed significant strain on families with children [39]. Finally, while women generally reported very high levels of adherence to COVID-19 related health behaviours (such as social distancing, wearing masks, and disinfecting surfaces), some sociodemographic differences in adherence were observed. Observed associations between some pandemic-related behaviours (avoiding public places and wearing masks) and country of residence, suggest that national guidance, regulations, and sociocultural contexts may significantly impact on pregnant women's health-related decision-making during the pandemic. Unlike the findings related to stress outcomes, these findings suggest a need for potentially more targeted interventions within countries and groups to improve and/or maintain appropriate prenatal behaviours during a pandemic.

Strengths/limitations
The cross-sectional nature of this study is limited as we are unable to track how the pandemic may impact participants' wellbeing-and behaviours over time/as perceived COVID-19 risk and restrictions change. In addition, while assessment of stress during the pandemic was conducted using validated measures of general stress and pregnancy-specific stress, there was not, at the time of data collection, a published validated scale to measure COVID-19 related stress in pregnant women. While the measure of COVID-19 related stress designed for this study had good internal validity, a validated measure of prenatal stress and anxiety has since been published [40], and this would prove useful for future research in this area. We also did not assess whether women had received a positive COVID-19 test result, which might have influenced their well-being and behaviours.
Furthermore, participants in this study were largely married or in a relationship, educated to degree-level or higher, socioeconomically advantaged, and resident in Ireland, the United Kingdom or America. Thus these results may tend to understate the economic and social drivers of stress and adherence to health advice during the pandemic. However, the international sample enhances generalisability of our findings and interpretations. In addition, our comprehensive approach to examining multiple types of stress and a range of health behaviours provides important insights into as yet under-examined aspects of women's experiences of pregnancy during the pandemic.

Conclusion
This study found high levels of pregnancy-specific and COVID-19 related stress, and low levels of mental and physical health during the COVID-19 pandemic. In this sample, these variables were mostly predicted by well-being factors rather than sociodemographic factors. Pregnant women in this study reported good engagement with healthy behaviours and high adherence to COVID-19 public health recommendations during the pandemic.
Our findings highlight the need for prenatal mental health supports that target pregnancy-and coronavirus-specific concerns, and that can be implemented at population levels to support prenatal mental health. While experiences of stress did not appear to be place-or demographic-specific in this study, the association between certain health behaviours (including mask-wearing and avoiding crowded places), and sociodemographic factors, suggest that some elements of prenatal health behaviour may be influenced by national guidance and cultural context, further illustrating the need for structured interventions for women pregnant during the pandemic. These supports are clearly greatly needed and may help to mitigate adverse effects for women and children.

Author agreement
The authors hereby certify that the work presented in this submission is the authors' original work, has not been published (and is not under consideration for publication) elsewhere, and has been reviewed and approved by all authors. The authors also agree to abide by the copyright terms and conditions of Elsevier and the Australian College of Midwives.

Author contributions
KMS conceptualised and designed the study, collected and analysed the data, wrote and revised the final manuscript.
JP contributed to study design, collected data and assisted with data analysis, and critically reviewed and revised the manuscript.
EKO designed the study, contributed to analyses and critically reviewed and revised the manuscript.
SL conceptualised and designed the study and critically reviewed and revised the manuscript.
SM conceptualised and designed the study, contributed to analyses and critically reviewed and revised the manuscript.