The Impact of Covid-19 On Maternal Delivery Experiences and Breastfeeding Practices in China: A Cross-Sectional Study

The extensive on of the pandemic on maternal delivery and infant feeding experiences, what are the for future practical and support? Using the directed acyclic graphs to identify the minimum adjustment set of confounders, this study assessed the relationship between lockdown measures, maternal postnatal experiences, breastfeeding supports, and predictors for 6-month mostly breastfeeding practice during the COVID-19 pandemic. Notably, we found younger mothers were more likely to breastfeeding their infants in Beijing, China, presumably due to the new fashion of breastfeeding advertised by the social media.


Abstract Background
The COVID-2019 pandemic has placed extensive pressure on health systems and posed a severe public health challenge worldwide. Lockdown measures implemented in many countries have delayed virus spread. However, a considerable number of people have faced unprecedented pressure, especially the pregnant and breast-feeding women, because face-to-face professional support has been reduced during the lockdown in many countries.

Objectives
To compare the delivery and infant feeding experiences of women who delivered before (BL) versus during (DL) the Covid-19 pandemic in Beijing, China.

Methods
Women aged ≥ 18 years with an infant ≤ 18 months of age completed an anonymous survey. Information/links were shared online and via local clinics in Beijing. Logistic regression was performed to assess predictors of breastfeeding during the rst 6-months.
Results 2233 women provided data; BL 1241 (55.6%), DL 992 (44.4%). The 6-month mostly breastfeeding (MBF, mainly breastfeeding with few non-formula uids added) rate was signi cantly higher in the DL mothers (71.6%), compared to the BL mothers (60.6%). MBF was predicted by delivery during the lockdown period (OR1.42, 95% con dence interval 1.06, 1.90), younger maternal age, discussing infant feeding in a feeding support group, and getting support from friends or relatives. Moreover, relative to those little affected by COVID, those who were much affected were less likely to MBF.

Conclusion
The COVID-19 pandemic and lockdown measures in uenced mothers' planned birth location and breastfeeding intention. However, breastfeeding practice was maintained during the pandemic. The reported breastfeeding rates were high and our results highlight the importance of feeding support as well as potential bene cial effects of increased mother-infant contact during the lockdown period which is relevant even under normal circumstances.

Synopsis Introduction
The coronavirus disease 2019 (COVID-2019) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), began in December 2019 and has placed extensive pressure on health systems and posed a severe public health challenge worldwide [1,2]. COVID-19 affects people of all ages, with no signi cant gender difference [3][4][5]. Lockdown measures implemented in many countries have delayed virus spread. Nevertheless, confronted with the COVID-19 pandemic and pervasive disruption to daily life, a considerable number of people have faced unprecedented pressure. This is particularly the case for pregnant and breast-feeding women, because face-to-face professional support has been reduced during the lockdown in many countries.
Breast milk is the optimum food for infants, and provides the energy and essential nutrients required during the rst few months of life as well as non-nutritive bioactive components, many of which provide protection against infection [2]. Existing evidence indicates that SARS-CoV-2 is not transmitted via breastmilk [6,7]. Even so, breastfeeding, particularly for mothers who might potentially be infected with COVID-19, has been controversial since the pandemic began. The World Health Organization (WHO) has recommended that mothers with COVID-19 (or suspected  can breastfeed as long as they take appropriate precautions [4,8], because the bene ts substantially outweigh the potential risks of viral transmission from mother to infant [9,10]. Regardless, anecdotal media accounts have highlighted the separation of mothers and babies in hospitals, alongside the circulation of inaccurate stories about the safety of breastfeeding. In China, some clinicians have recommended that mothers with con rmed or suspected COVID-19 pneumonia should stop breastfeeding, although the Centres for Disease Control and Prevention (CDC) in China have suggested that those mothers can continue breastfeeding with appropriate precautions, because no viral nucleic acid has been detected in breast milk [7,11].
In Beijing, China, schools were closed from 26th January to 31st July 2020 [12], alongside most public places apart from essential stores. Travel restrictions were imposed and meeting with those from other households limited, other than for caring or work purposes. Additionally, as a means of ensuring the safety of mothers and babies, routine postpartum home visits were replaced with telephone consultations, although mothers were able to request face-to-face appointments if required. In this event, a triage process was followed to identify suspected cases of COVID-19. Although these strategies effectively protected postpartum women from infection, the impact of such strict policies on women's wellbeing breastfeeding practice remains unknown. Given that breastfeeding is optimally supported through high-quality professional and peer-to-peer support and positive maternal wellbeing [13], it is important to understand the impact of the pandemic on women's ability to breastfeed.
This study therefore aimed to investigate: 1) The delivery experience, feeding intentions and actual feeding methods in mothers who delivered before or during the lockdown; 2) The impact of lockdown upon breastfeeding practice and support; 3) Factors that predict infant feeding methods in the rst 6 months. The research is important to identify problems experienced by this vulnerable group, and will also contribute to the formulation of public health policies in the event of future public health emergencies.

Recruitment
The research was undertaken between 1st August 2020 and 31st October 2020, at several local clinics attached to Beijing Children's Hospital (BCH) in Beijing, China. Advertisements, including a brief introduction to the research and the inclusion criteria, were posted on the bulletin board in each local clinic's reception area. Eligible mothers were invited to complete a one-time, anonymous questionnaire, which consisted of 48 questions and took approximately 15-20 minutes to complete. Mothers could complete the questionnaire during the waiting time in the clinic. A trained nurse was available to assist the mothers in completing the questionnaire. The inclusion criteria were: 1) mother ≥ 18 years; 2) infant ≤ 18 months of age at the time of survey completion; 3) living in Beijing and breastfeeding their infant (exclusively or partially) for some or all of the lockdown period. Ethical approval was obtained from the Beijing Children's Hospital Research Ethics committee (2020-Z-102). The rst page of the survey provided information about the study.
The questionnaire was provided both online and in paper. The online version was supported by the WenjuanNet (https://wenjuan.net/s/NVveyew/), which could create online questionnaire with a large number of templates provided. A QR code was generated when the online survey started. Eligible mothers who were interested in participation could begin the survey on their smart phone by scanning the QR code, or by clicking the invitation link. Participants could refuse to answer any question if they did not want to.

Content Of The Survey
The questionnaire was adapted from one intended for use in the UK, namely the COVID-19 New Mum Study [14,15] with translation into Chinese. The translated Chinese version was produced by a native Chinese speaker on the team and re ned by a BCH paediatrician with clinical experience of infant feeding. The survey comprises of four sections: 1. Demographic characteristics: including mothers' socio-economic status, infant age, gender, weight and gestational age. 2. Infant feeding and birth experiences: including infant feeding goals, delivery place and method, postbirth skin-to-skin contact, sources of support for infant feeding, and the impact of COVID-19 on infant feeding. Mothers were also asked about the method of milk feeding used during the rst 6 months ("what is the source of milk you used for feeding for the rst six months after birth?"). The response options were exclusive breastfeeding, mixed feeding or formula-feeding. However, given that a strict de nition of exclusive breastfeeding was not provided, and the question did not ask about complementary foods, the exclusive breastfeeding category was renamed as 'mostly breastfeeding' (MBF) for the analyses, re ecting the majority source of the infant's milk intake during this period. 3. Maternal mood, daily activities, living circumstances and supports received on infant feeding and personal heath during the lockdown.
4. The impact of COVID-19 on maternal mental health and life patterns.
In the nal version of the questionnaire, questions from different sections were mixed to avoid anxiety or boredom of participants that might arise from repeatedly answering similar questions.

Statistical Analyses
Online data were exported from https://wenjuan.net/s/NVveyew/. Statistical analysis was conducted in SPSS version 22.0 (IBM., Armonk, NY, USA). Descriptive data are shown as mean (standard deviation, SD), or n (%). Data are presented separately for women who delivered before or during the lockdown, with comparisons between groups using independent t-test, Mann-Whitney or chi-squared test as appropriate. Directed acyclic graphs (DAG) (EFigure 1) were sketched to identify the minimum adjustment set of confounders and to identify ancestors of the outcome that were dropped from the regression model. On this basis, logistic regression was performed to assess predictors of breastfeeding during the rst 6months. Mothers were included in the regression analysis if their infants had reached 6-months of age.
Adjusted odds ratio (OR) and 95% con dence intervals (CI) are presented, and P < 0.05 was considered as statistically signi cant.

Results
Demographic characteristics of the participants From 1st August to 31st October 2020, 2233 mothers completed the questionnaire. All participants were married. At the time of survey completion, mean maternal age was 31.4 ± 4.1 years, and mean infant age 8.2 ± 4.0 months (range 1-18 months). The background characteristics are shown in Table 1. There were 992 (44.4%) and 1241 (55.6%) mothers who gave birth during the lockdown (DL) and before the lockdown (BL), respectively. Maternal education and infant gestation did not differ between the BL and DL groups. Infant age was higher as expected in BL (10.8 ± 3.1 months) than in DL (5.0 ± 2.3 months).
The BL group contained more male infants (52.5%) than the DL group (48.2%). Mothers in the BL group were also older than those in the DL group (31.7 ± 4.1 years vs. 31.2 ± 4.0 years) and more were in the lowest category for education and household income compared to the DL group. Notes: * P < 0.05. SD = standard deviation a T-apartment is a type of apartment building with around ten residents living on each oor. b B-apartment is a type of apartment with less than three residents living on each oor.

Birth experiences and daily life
For DL and BL mothers, 64.3% and 60.4% were vaginally delivered and the post-birth skin-to-skin contact was 97.1% and 96.3% respectively; with no signi cant differences between groups. Over 90% of the women delivered at a national hospital, although 13.2% of DL mums reported the delivery place had been altered. Notably, over half of women in the DL group considered COVID-19 to have had a moderate-tohigh impact on their household income (53.9%) and on the convenience of purchasing daily necessities (50.7%), compared to 50% and 44.1% in the BL group (Table 2). Note: * P < 0.05 Infant feeding during the lockdown Infant feeding intentions did not differ between DL and BL groups ( Table 3). Among DL women, 32.6% reported a moderate-to-high impact of COVID-19 on their infant feeding practice, compared to 24.2% in the BL group (, Table 3). Besides, 25.9% of mothers in the BL group with a feeding goal of EBF before delivery had changed to formula or mixed feeding after birth, compared to 17.8% in DL mothers (Fig. 1).
In the DL group, signi cantly more mothers reported 6-month MBF compared to the BL mothers (71.6% vs. 60.6%). As expected, given the younger age of the infants of DL mothers, a signi cantly higher proportion were MBF (52.8%) or fed a combination of breast milk and infant formula (11.5%) at the time of completing the questionnaire, while a greater proportion of older infants born BL also consumed solid foods (Table 3). professional, friends and relatives, online support sources or reported no support during the pandemic (Table 4). No difference was found between BL and DL mothers regarding the frequency of participating in a breastfeeding support group. Signi cantly more DL women reported having contacted health professionals more than four times a week (8.9%), compared to those from the BL group (7.2%). Of the mothers who had contacted health professionals, 97.7% thought they had obtained enough help with breastfeeding in the DL group, which was signi cantly higher than BL mothers (95.3%). Moreover, more mothers in the DL group reported receiving su cient help from family, compared to BL mothers (97.2% vs. 94.8%), suggesting that mothers who had given birth during the pandemic may have received more support than those who had given birth before.

Predictors of infant feeding during the rst 6-months
Logistic regression was performed based on the DAG (EFigure 1), including 1231 BL and 576 DL mothers who were feeding an infant ≥ 6 months at the time of completion of the questionnaire. MBF during the rst 6 months was positively predicted by delivery DL versus BL (OR1.42, 95%CI 1.06, 1.90). Positive predictors of MBF (Table 5) also included younger maternal age, infant feeding discussion in a support group more than four times a week, and receiving support from friends or relatives. Moreover, relative to those little affected by COVID, those who were much affected were less likely to MBF. Notes: MBF = Mostly breastfeeding; CI = con dence interval; DL = mothers who delivered during the lockdown; BL = mothers who delivered before the lockdown. Factors controlled for included maternal socio-economic status (education, income, social class), living conditions, cohabitation with family members ("Who did you live with during the lockdown?"), gestational age, health advice-seeking behaviour ("Frequency of discussing with health professionals/chatting in breastfeeding support group"), support ("Did you receive enough support from health professionals/family members?"), and the impact of the COVID-19 lockdown on maternal daily life.
a Class IV&V mainly refer to manual labor jobs which don't require high education experience; Jobs in class III & II require certain training, certi cations, licenses and degree to qualify; Jobs in class I involve professional careers which require advanced degree, high-end skills or expertise.
b T-apartment is a type of apartment building with around ten residents live in each oor. c B-apartment is a type of apartment with less than three residents live in each oor. Mode of delivery Notes: MBF = Mostly breastfeeding; CI = con dence interval; DL = mothers who delivered during the lockdown; BL = mothers who delivered before the lockdown. Factors controlled for included maternal socio-economic status (education, income, social class), living conditions, cohabitation with family members ("Who did you live with during the lockdown?"), gestational age, health advice-seeking behaviour ("Frequency of discussing with health professionals/chatting in breastfeeding support group"), support ("Did you receive enough support from health professionals/family members?"), and the impact of the COVID-19 lockdown on maternal daily life.
a Class IV&V mainly refer to manual labor jobs which don't require high education experience; Jobs in class III & II require certain training, certi cations, licenses and degree to qualify; Jobs in class I involve professional careers which require advanced degree, high-end skills or expertise.
b T-apartment is a type of apartment building with around ten residents live in each oor. c B-apartment is a type of apartment with less than three residents live in each oor. Notes: MBF = Mostly breastfeeding; CI = con dence interval; DL = mothers who delivered during the lockdown; BL = mothers who delivered before the lockdown. Factors controlled for included maternal socio-economic status (education, income, social class), living conditions, cohabitation with family members ("Who did you live with during the lockdown?"), gestational age, health advice-seeking behaviour ("Frequency of discussing with health professionals/chatting in breastfeeding support group"), support ("Did you receive enough support from health professionals/family members?"), and the impact of the COVID-19 lockdown on maternal daily life.
a Class IV&V mainly refer to manual labor jobs which don't require high education experience; Jobs in class III & II require certain training, certi cations, licenses and degree to qualify; Jobs in class I involve professional careers which require advanced degree, high-end skills or expertise.
b T-apartment is a type of apartment building with around ten residents live in each oor. c B-apartment is a type of apartment with less than three residents live in each oor.

Summary of the results
Our ndings show that, despite the di culties imposed by the pandemic, hospital facilities in Beijing continued to implement measures based on WHO guidelines [8,16,17]. Current evidence and clinical experience con rms the safety of breastfeeding for mothers with con rmed or suspected COVID-19 [18][19][20] and the CDC of China have modi ed the relevant guidelines from initially not recommending breastfeeding for women with suspected or con rmed COVID-19 to currently recommending that breastfeeding mothers in isolation for COVID-19 can continue breastfeeding if they follow protective measure to reduce the risk of droplet transmission [11] Infant feeding practice and support 60.6% and 71.6% mothers in the BL and DL group chose EBF for the question "what is the source of milk you used for feeding for the rst six months after birth?", which is signi cantly higher than the EBF rate reported in previous studies in China [21]. This may re ect differences in the de nition of EBF used in different studies. We did not provide a speci c de nition of EBF and, whilst nurses were available to help mothers if they had any queries, some mothers may not have asked for clari cation. Furthermore, the question in the survey referred to milk feeding but did not speci cally ask about the use of complementary foods before 6 months. For these reasons, we renamed the EBF variable as 'mostly breastfeeding (MBF)' -to re ect the infant's main source of milk during the rst 6 months. The high rate for MBF during the rst 6 months is also likely to re ect the inclusion criteria for the study which speci ed that participants must have breastfed their infant during the lockdown. Rates of EBF vary across provinces and cities in China. According to the data published in 2013, the EBF rate for 14539 children from 30 provinces in China was 20.8% [22], whilst a large cohort study conducted in Zhejiang province involving 42,550 children reported a 3-to 5 month EBF rate of 51.3% for girls and 46.8% for boys [21]. However, in that cohort, EBF ≥ 6 months was de ned as self-reported EBF at each of the 3 clinical visits (1, 3, and 6 months), thus the EBF rate may have been over-estimated. In another cohort study in Zhejiang in 2005, 74% of 1520 mothers reported "any breastfeeding" de ned as the infant has received breastmilk (direct from the breast or expressed) with or without other drinks, formula, or infant food at 6 months.
Zhejiang is an eastern coastal province in China, where health care facilities and baby-friendly hospitals are well-developed, and, compared to Zhejiang, the development of baby-friendly hospitals is even better in Beijing. As reported by Beijing Municipal Health Commission [23], in 2019, there were 114 baby-friendly hospitals in Beijing, where most of our participants were recruited. A Chinese review showed that after the baby friendly hospital initiative became established in 1992, the 4-month EBF rate (EBF was not de ned in the methodology) increased from 28-40% [24]; a retrospective study reported the same trend with an increase in the 4-month EBF rate in Beijing from 16.77% in 1991-1994 to 58.77% in 1995-1998 [25].
While more recent data are need to con rm this nding, it supports the importance of baby friendly hospitals in supporting EBF.
Interestingly, the ndings show that a signi cantly lower proportion of DL mothers changed their initial feeding goal from EBF to formula/mixed feeding compared to BL mothers (18% vs. 26%), suggesting that the lockdown measures may have indirectly promoted breastfeeding, perhaps by allowing mothers to spend more time with their babies, or feeling less pressure due to fewer visitors during the lockdown [26]. A similar nding was reported in the COVID-19 New Mum Study in the UK [27], where 13% of women reported changes to feeding method due to the lockdown; a higher frequency and longer duration of breastfeeding was observed in 30% and 17% women who were breastfeeding during the pandemic, re ecting more time spent at home as well as a greater contribution to childcare from themselves and partners during the lockdown period. However, as suggested by the researchers, the effects of the pandemic on infant feeding may differ depending on access to support and special circumstances experienced by individual mothers. Indeed, our results suggest that DL mothers received signi cantly more infant feeding support than BL mothers. As suggested by the literature, both peer and professional support are important for the success of breastfeeding [28].
Research on breastfeeding support has consistently identi ed the husband or partner as an important source of support for women, with in uence on four aspects in particular: breastfeeding decisions, assistance at rst feeding, assistance during breastfeeding, and risk factors for bottle feeding [29]. Our study found that women from both groups considered their partner as the greatest source of support for infant feeding, consistent with the UK New Mum study [27]. Compared to BL mothers, DL mothers received even more support from their partners, possibly re ecting the "work from home" measures in place during the lockdown, which increased the time husbands could spend with their wives. As suggested by Vazquez et al [27], husbands who contribute more to childcare during this exceptional period may represent a valuable source of support for women, especially for those with limited access to friends or relatives who would provide this support under normal circumstances.
According to Chinese tradition, mothers normally adhere to a month long "con nement period" after birth.
Although an increasing number of mothers use a con nement centre where they can receive care and professional support, most centres were closed during the lockdown period; this may have led to differences in the con nement experience for mothers in the DL and BL groups. However, a recent study found that the 1-and 6-month EBF rate did not differ between mothers who chose a con nement centre and mothers who stayed at home (37% vs. 42%, p = 0.5) [30]. Whilst more studies are needed to con rm this nding, we suggest that mothers in the DL group may have received more support from their partner, so the total support received by the DL and BL mothers might have been the same.

Predictors of infant feeding during the lockdown
Mothers who delivered during the lockdown were more likely to mostly breastfeed their infant during the rst six months. This may re ect the increased time available for both mothers and their partners to spend with the infant and the breastfeeding support provided during the lockdown. Higher rates of MBF were also observed among mothers who delivered at a younger age. This result is consistent with a previous study in Zhejiang, China, which indicated that EBF was positively related to younger maternal age [31]. However, this contrasts with studies in many high-income countries which indicated that older maternal age was associated with a higher rate of breastfeeding initiation and the duration of EBF [32][33][34]. This may re ect social and cultural factors, and emphasizes the importance of considering these factors when developing health policy to promote breastfeeding. In recent years, EBF has become a "new fashion" in several modern cities in China, including Beijing, Shanghai, and some large cities in Zhejiang province, possibly re ecting the increasing number of Chinese female celebrities sharing their EBF experiences and healthy mother-infant relationships through vlogs or posts. Young women from these cities are keen to follow this new fashion and are very proud of themselves if they successfully EBF to 6 months. Additionally, our results suggest that discussing infant feeding in a feeding support group more than four times a week, getting support from friends or relatives, and considering that the COVID-19 pandemic had little or no impact on their planned birth location positively predicted MBF in the rst 6months, re ecting the vital role of infant feeding support and the normal operation of health care facilities in promoting breastfeeding for mothers.

Strengths and Limitations
To our knowledge, this is the rst cross-sectional study investigating infant feeding practices and potential predictors of infant feeding during the COVID-19 pandemic in China. Using an adapted version of questionnaire that has been used in the COVID-19 New Mum Study [14,27] in the UK makes the results comparable between countries. However, the study has some limitations. Firstly, although Beijing is a metropolitan city with 20 million residents from different regions of China [35], study participants are not representative of all new mothers in China; since the signi cant higher gross domestic product per capita in Beijing results in a higher level of family income and education of the population [36], which may partly explain the higher MBF rate in this study [37]. However, considering the duration of lockdown and the type of lockdown measures varied between provinces and cities in China, it was di cult to compare the infant feeding outcomes between cities even in a national survey. Secondly, there may be recall bias since at the time of survey completion lockdown measures had just ended in Beijing. Last but not least, we did not clearly de ne EBF in our questionnaire so; consequently, we relabeled this variable as 'mostly breastfeeding'.

Conclusions
Our ndings highlight that despite di culties imposed by the pandemic, feeding support was generally well preserved for mothers who delivered a baby during the lockdown, re ected in high rates of mostly breastfeeding in the rst 6 months. This highlights the importance of breastfeeding support during public health emergencies. Moreover, our results suggest that breastfeeding may have been facilitated by mothers and fathers spending more time at home, allowing both to contribute to childcare during the lockdown period, which may also be relevant under normal circumstances. The study has been performed in accordance with the Declaration of Helsinki. We con rm that all methods in this study were performed in accordance with the STROBE checklist. Ethical Approval was obtained from the Beijing Children's Hospital Research Ethics Committee (reference number: 2020-Z-102).
Ethics approval and consent to participate Ethical approval was obtained from the Beijing Children's Hospital Research Ethics Committee (reference number: 2020-Z-102). The rst page of the survey provided information about the study. Written informed consent was obtained from a parent or guardian for participants under 16 years old.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare no con ict of interest with respect to this study. Professor Mary Fewtrell receives an unrestricted donation for research on infant nutrition from Philips. The remaining authors declare no other con icts.

Funding
No funding received.

Authors' contributions
All authors contributed to the design and planning of the study. ZW collected the data, JY and MG drafted the manuscript. MG contributed to the data analysis. MF and JW re ned the manuscript. All authors read and approved the nal manuscript. Differences in infant feeding plan in mothers delivered before and during the lockdown Figure Legend: EBF=exclusive breastfeeding; MBF= mostly breastfeeding; Others=other feeding method during the rst six months except the exclusive breastfeeding (such as formula feeding and mixed feeding). Signi cant difference found between groups (Chi-square test).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. eFigure1.tif