Predictors of exclusive breastfeeding practice among migrant and non-migrant mothers in urban China: results from a cross-sectional survey

Objective To explore and compare the predictors for exclusive breast feeding (EBF) among migrant and non-migrant mothers in China. Design A large-scale cross-sectional study. Setting 12 counties/districts were covered in China. Participants A total number of 10 408 mothers were recruited, of whom 3571 mothers of infants aged 0–5 months in urban China were used for analysis. Outcome The practice of EBF was calculated based on the foods and drinks consumed in the last 24 hours, as recommended by WHO. Results Around 30% of Chinese mothers with infants aged 0–5 months practised EBF in urban areas, with no significant difference between migrant and non-migrant mothers (p=0.433). Among the migrant mothers, factors associated with EBF included residence in big cities (adjusted OR, AOR 1.68 (95% CI 1.20 to 2.34)), premature birth (AOR 0.27 (95% CI 0.09 to 0.81)), knowledge about EBF (AOR 2.00 (95% CI 1.51 to 2.65)), low intention of breast feeding in the first month postpartum (AOR 0.59 (95% CI 0.36 to 0.97)) and mothers working in agriculture-related fields or as casual workers (AOR 1.77 (95% CI 1.18 to 2.64)). Among non-migrant mothers, in addition to similar predictors including residence in big cities (AOR 1.40 (95% CI 1.13 to 1.73)), knowledge about EBF (AOR 1.25 (95% CI 1.02 to 1.53)) and low intention of breast feeding in the first month post partum (AOR 0.46 (95% CI 0.31 to 0.70)], early initiation of breast feeding (EIBF) (AOR 1.78 (95% CI 1.35 to 2.33)) and caesarean delivery (AOR 0.74 (95% CI 0.60 to 0.89)) were also factors associated with EBF. Conclusions There was no significant difference in the prevalence of EBF between migrant and non-migrant mothers in urban China. Premature birth and maternal occupation in agriculture-related fields or casual work were distinctive factors associated with EBF for migrants, while EIBF and caesarean delivery were unique predictors for non-migrants. Trial registration number ChiCTR-ROC-17014148; Pre-results.

Rapid economic development is inseparable from an increasing number of domestic migrants [4]. China has witnessed an unprecedented internal migration in world history.
On the one hand, better health infrastructures in cities can provide migrants with better health care services [5]. On the other hand, the household registration system and limited insurance coverage put migrants in a vulnerable situation with unfair access to urban benefits [6].
A large number of previous studies focused on immigrant mothers, who moved to another country where was different from their countries of origin [7]. Little attention has been paid to breastfeeding practice and its associated factors among migrant mothers who moved domestically. Previous literature in China disclosed that disparities in breastfeeding practice existed between urban and rural areas [3,8], employed and unemployed mothers [9], Han and other minority groups [10], high-educated and loweducated mothers [11]. Yet few studies have examined the predictors of breastfeeding  Calculation of sample size was based on the expected prevalence of EBF for 0-5 months old infants.
We adopted a multi-stage stratified cluster sampling approach when selecting the survey sample. In the first stage, all districts/counties were categorized into 4 strata, of which 12 districts/counties were selected (4 from large cities, 4 from medium and small cities, 2 from normal rural areas and 2 from poor rural areas). Selection of districts/counties considered geographic distributions, the executive capacities, and collaboration of the provincial level CDC as well as their population sizes. In the second stage, four clusters were randomly selected via probability proportional to size sampling in each selected district/county. One cluster usually had one community health center. In the last stage, mothers who brought their 0-11 months old infants to immunization clinics were randomly invited to participate in the survey until the designed sample size was met. The inclusion criteria of this study were mothers of 0-11 months old infants who signed informed consent, had no psychiatric disorders and were able to answer questions clearly. In order to decrease reporting bias of feeding practice, we only invited mothers who were the primary caregivers of their infants. In total, 10408 mothers of 0-11 months old infants were interviewed. Breastfeeding-related factors included early initiation of breastfeeding (EIBF) (yes and no), knowledge on EBF (yes and no), low intention of breastfeeding within 1 month postpartum (yes and no) and breastfeeding difficulties postpartum (yes and no). Social influence denoted whether her partner, friend, mother or mother-in-law agreed that breastmilk is better than breastmilk substitutes. Health-seeking behaviors were comprised of attending antenatal visits, breastfeeding education session or mother groups.

Patient and public involvement
We then performed logistic regression to identify predictors of exclusive breastfeeding practice. The selection of covariate variables was based on the p-values of Chi-square tests (P<0.05). Odds ratios and 95% confidence intervals were displayed.
All of the data analysis was performed by STATA version 15.0 (Stata Corporation, USA). We set the level of statistical significance at α=0.05.

Sample description and factors associated with EBF
Considering that migrants usually concentrate in urban areas, we only focus on 3571 mothers of 0-5 months old infants who lived in big, medium and small cities.
Migrant mothers were defined as those who had resided in a specific county for one month or more, and been in a place different from their registered county. We excluded 153 observations due to missing values of potential predictors, which left 3418 (>95% of the original sample) observations for statistical analysis, of which 2208 were nonmigrants and 1210 were migrants. Sensitivity analysis using the full sample revealed very similar results to those shown in this study. More than 60% of mothers were between 26-35 years old in both groups, while the proportion of mothers above 36 years old was significantly higher in non-migrant mothers. The likelihood of being Han ethnicity and being formally employed was statistically higher in non-migrant mothers than migrant mothers as well (Table 1)  10 education, who were formally employed and lived in big cities were more likely to practice EBF in both groups. Mothers from the Han ethnic group were more likely to practice EBF than mothers from minority groups among the migrant group (Table 1).
Among factors about obstetrical history, the percentage of having only one child among migrant mothers was statistically higher than non-migrant mothers (P<0.001).
The proportions of premature birth and low birth weight were quite similar between migrant and non-migrant mothers, while the rate of cesarean section was statistically higher in non-migrant mothers than migrant mothers. Cesarean section significantly Even though more than 80% of mothers received positive influence from their partners, friends, mothers or mothers-in-law on the fact that breastmilk is better than breastmilk substitutes, these influences differed among migrant and non-migrant mothers. The influence of partner and friend was significantly associated with EBF for non-migrant mothers, while the influence of partner and their own mothers were significantly associated with EBF for migrant mothers.
Non-migrant mothers were more likely to adopt health-seeking behaviors like attending antenatal visits, breastfeeding education sessions and mother groups than migrant mothers. However, only the proportion of mothers who attended antenatal visits was statistically different between these two groups. Attending breastfeeding education sessions only significantly affected the EBF practice of non-migrant mothers.

Predictors of breastfeeding practice among migrant and non-migrant mothers
We then performed logistic regression to determine the predictors of exclusive breastfeeding respectively. We only kept covariates with p-values smaller than 0.05 in Table 1.

Discussion
In this study, we found that breastfeeding practice was suboptimal in both migrant and non-migrant mothers. The prevalence of EBF among non-migrant women was not significantly different from that of migrant mothers. This is the first and largest study to explore and compare breastfeeding practice and their potential predictors among migrant and non-migrant mothers in China. Multivariate regression results suggested that similar predictors of EBF for mothers in both groups included residence in big cities, knowledge on EBF and low intention of breastfeeding in the first month postpartum. Mothers in big cities were approximately 1.5 times more likely to practice EBF. Among migrant mothers, having premature infants and breastfeeding difficulties postpartum were associated with EBF practice. However, for nonmigrant mothers, EIBF and cesarean delivery were unique predictors of EBF practice.
Living in big cities was the only common variable in demographic characteristics that were significantly associated with EBF practice in both groups. Previous literature also suggested a higher rate of EBF in big cities than in middle and small cities in China using nationally representative data derived from the Chinese National Nutrition and Health Survey conducted in 2013 [3]. This is possibly due to better knowledge and supporting environment for mothers in big cities than those living in small and medium cities regardless of their migration status.
For example, the proportion of mothers who received information on encouraging breastfeeding was higher in big cities of 86.7% than that in small and medium cities of 76.1%, and more than 90% of them received encouragement from the delivery hospital in big cities.
This revealed that both migrant and non-migrant mothers may benefit from better breastfeeding-related education and support in big cities in China.
Another common predictor of EBF practice for both migrant and non-migrant mothers was knowledge of EBF. Some of the previous studies used knowledge scores on breastfeeding benefits as a predictor of breastfeeding practice, but they failed to provide consistent evidence 14 on its association with EBF [8,14]. However, we only focused on the key message that infants should be exclusively breastfed within 6 months. This implies that future education intervention projects should pay more attention to disseminating core messages in promoting optimal breastfeeding practice.
Additionally, we also found that the low intention of breastfeeding in the first month postpartum significantly decreased the odds of practicing EBF for both groups. In alignment with previous studies, nipple pain or fissure within the first month postpartum was the commonest reason for the low intention of breastfeeding among mothers regardless of their migration status [15,16]. Given that nipple pain was often attributed to incorrect positioning and attachment [16], health care providers need to find effective measures to help to establish proper positioning of mothers, good attachment of the babies to the breast as well as effective sucking.
In addition to common predictors of EBF practice explained above, there were unique risk factors for migrant and non-migrant mothers respectively.
Among migrant mothers, having premature infants significantly decreased the prevalence of EBF. The EBF rate for migrant mothers with preterm infants was 9.8%, which was largely lower than non-migrant mothers of 31.0%. The low EBF rate can be attributable to motherinfant separation in neonatal intensive care units (NICUs) in China [17], which resulted in many difficulties and challenges for mothers with premature infants. The disparity in the EBF rate of premature infants may come from different choices of hospitalization services between migrant and mon-migrant mothers. A recent study in China found that migrants especially those from rural areas were less likely to use tertiary hospitals, where well-trained health workers and good service were concentrated [18].
Different from previous studies, we failed to find that breastfeeding difficulties were barriers to optimal breastfeeding practice [19]. In contrast, migrant mothers who encountered This is possibly because around 95% of the breastfeeding difficulties encountered by migrant mothers were solved successfully, and over 80% of them felt that these difficulties didn't impede future breastfeeding. A systematic review of previous literature suggested that the combination of peer support and professional support can effectively ensure the continuation of breastfeeding [20]. Experience in developed countries suggested that health professionals such as midwives, health visitors as well as lactation consultants had started to get involved in peer support programs collaborating with voluntary breastfeeding organizations [21]. The government may take the initiative to fund breastfeeding support projects and provide incentives to health professionals to actively provide support to lactating mothers.
Among non-migrant mothers, EIBF and cesarean section were distinctive predictors of EBF. The rate of EIBF was 12.8% in non-migrant mothers, which was a little bit higher than a study in Sichuan province of 9% [22], but much lower than the national prevalence in 2013 of 28.7% [23] and a study in central and western China of 59.4% [24]. A large regional disparity of EIBF in China may be due to the delayed process of implementing Early Essential Newborn Care (EENC) in China. Even in places where EENC had been implemented, skin-to-skin contact was often interrupted due to inadequate facility resources and early initiation of breastfeeding was difficult to be ensured [25].

Disclaimers
The views expressed in the article are those of the authors and do not necessarily reflect the views of the institution and funder.

Competing interests
None of the authors had a conflict of interest related to any part of this study or manuscript.

Availability of data and materials
This dataset is available from the corresponding author on reasonable request.

Not required
Disclosure of prior presentation of study data as an abstract or poster

Acknowledgements
The authors would like to thank research teams from 12 sample sites for their hard work in the data collection. The authors also want to thank all the mothers who participated in this study.

Strengths and limitations of this study
 This is the first and largest study to explore and compare breastfeeding practices and its associate factors among migrant and non-migrant mothers in urban China.
 Mothers in large cities were more likely to practice EBF when compared with mothers in medium and small cities, regardless of their migrant status.
 The mothers covered in this study may not be representative of mothers in urban areas since we only invited mothers who brought their infants to immunization clinics to participate, which may overestimate the prevalence of exclusive breastfeeding.
 Since it is not a nationally representative study, the results cannot be generalized to all the urban areas of China.  [2,3]. Rapid economic development is inseparable from an increasing number of domestic migrants [4]. China has experienced an unprecedented internal migration in world history. On the one hand, better health infrastructures in cities can provide migrants with better health care services [5]. On the other hand, the household registration system and limited insurance coverage put migrants in a vulnerable situation with less potential access to urban benefits [6]. In addition to socio-economic and cultural differences between migrant and non-migrant women in China, previous studies also found that migrant women had poor knowledge and limited utilization of maternal health care services [7][8][9].
A large number of previous studies have focused on immigrant mothers, who moved to a country which was not their country of origin [10]. Immigrant mothers were found to be more likely to initiate and continue breastfeeding than non-immigrant mothers. However, EBF remained a challenge for both groups [10]. Little attention has been paid to breastfeeding practices and its associated factors for domestic migrant mothers. Previous literature in China disclosed the disparities in breastfeeding practices between urban and rural areas [3,11], employed and unemployed mothers [12], Han and other minority groups [13], and highly-educated and less-educated mothers [14]. Yet, few studies have examined the predictors of breastfeeding practices among migrant mothers. To fill this knowledge gap, we analyzed a large-scale cross-sectional survey

Study design and participants
Data for this study was derived from a large-scale cross-sectional survey in China Calculation of sample size was based on the expected prevalence of EBF for 0-5 monthold-infants.
We adopted a multi-stage stratified cluster sampling approach while selecting the survey sample. In the first stage, all districts/counties were categorized into 4 strata, of which 12 districts/counties were selected (4 from large cities, 4 from medium and small cities, 2 from normal rural areas and 2 from poor rural areas). The selection of districts/counties considered their geographic distributions, executive capacities, collaboration of the provincial level CDC, and their population sizes. In the second stage, four clusters were randomly selected using the probability proportional to size sampling in each selected district/county. One cluster usually had one community health center. In the last stage, mothers who brought their 0-11 month-old-infants to immunization clinics, were randomly invited to participate in the survey until the designed sample size was met. The inclusion criteria of this study were mothers of 0-11 month-old-infants who signed the informed consent form, had no psychiatric disorders and were able to answer questions clearly. To decrease the reporting bias of feeding practice, we only invited mothers who were the primary caregivers of their infants. In total, 10,408 mothers of 0-11 month-old-infants were interviewed.

Patient and public involvement
This survey was undertaken by mothers of 0-11 month-old-infants in China to understand the prevalence of EBF and its associated factors. Neither these mothers nor the public were involved in the study design and implementation. The study results will be disseminated to the public through media briefings and scientific publications.

Data collection
This survey was conducted by the NINH, China CDC in collaboration with 12 teams at the provincial level CDC from September 2017 to January 2018. Data were collected through face-to-face interviews using a structured questionnaire programmed into smartphones or tablets. The questionnaire was developed by NINH, China CDC, which comprised eight sections: demographic characteristics, breastfeeding practices, maternal and child health, supportive environment for breastfeeding, health service, workplace, social environment and culture, and household financial situation related factors. The study design was approved by the Medical Ethics Committee at the NINH, China CDC. We obtained written consent forms from all mothers.

Statistical Analysis
We used the indicators recommended by the World Health Organization (WHO) to assess the feeding practice of 0-5 month-old-infants, based on the foods and drinks consumed in the last 24-hours. We defined the practice of being fed exclusively with breast milk in the last 24-hours as exclusive breastfeeding.  Social influence denoted whether her partner, friend, mother or mother-in-law agreed that breastmilk was better than breastmilk substitutes. Health-seeking behaviors comprised the timing of considering how to feed the infants and attending antenatal visits, breastfeeding education session or mother groups.
We then performed logistic regressions to identify predictors of EBF. The selection of covariate variables was based on the p-values of Chi-square tests (P<0.05).

Sample description and factors associated with EBF
Considering that migrants usually concentrate in urban areas, we only focused on 3,571 mothers of 0-5 month-old-infants who lived in big, medium and small cities.
Migrant mothers were defined as those who had resided in a specific county for a month or more and had been in a place different from their registered county. We excluded 172 observations due to the missing values of potential predictors, which left 3,399 (>95% of the original sample) observations for statistical analysis, of which 2,199 were non-migrants and 1,200 were migrants. A sensitivity analysis using the full sample revealed very similar migrant distribution to the one shown in this study. Mothers or their partners with a college or higher education were more likely to practice EBF in both groups. The prevalence of EBF was higher in mothers who had a job, were formally employed, and lived in big cities among both migrant and non-migrant mothers. Mothers from the Han ethnic group were more likely to practice EBF than mothers from minority groups among the migrant group (Table 1).
Among factors about obstetrical history, the percentage of having only one child among migrant mothers was statistically higher than for non-migrant mothers (P<0.001). The proportions of premature birth and low birthweight, were quite similar between migrant and non-migrant mothers, while the rate of cesarean section was statistically higher in non-migrant mothers than migrant mothers. Cesarean section was significantly associated with lower prevalence of EBF for both migrant and nonmigrant mothers. However, for migrant women, premature birth and low birthweight of infants, were also negatively related to EBF. A significantly larger proportion of migrant mothers delivered their infants in hospitals at the municipal level or above (P<0.001). Mothers who delivered infants in hospitals at the district/county level were less likely to practice breastfeeding in both groups.
EIBF, having a low intention of breastfeeding within 1 month postpartum and breastfeeding difficulties postpartum, occurred more frequently among migrant mothers than non-migrant mothers, while only the prevalence of having breastfeeding difficulties postpartum was statistically different between the two groups. Unlike the non-migrant group where EIBF was associated with EBF significantly (P<0.001), migrant mothers were more likely to practice EBF if they had breastfeeding difficulties of them had a low intention of breastfeeding within 1 month postpartum in both groups.
Mothers in both groups were more likely to practice EBF if they knew its definition and were less likely to do it if their intentions of breastfeeding were low within 1 month postpartum.
Even though more than 80% of mothers received positive influence from their partners, friends, mothers or mothers-in-law about breastmilk being better than breastmilk substitutes, these influences differed among migrant and non-migrant mothers. The influence from partner and friends was significantly associated with EBF for non-migrant mothers, while the influence from their partners and their own mothers were significantly associated with EBF for migrant mothers.
Non-migrant mothers were more likely to adopt health-seeking behaviors like considering how to feed their infants before pregnancy and attending antenatal visits, breastfeeding education sessions, and mother groups, than migrant mothers. Only the timing of considering how to feed their infants was statistically different between the two groups. Migrant mothers who considered how to feed their infants before pregnancy were more likely to practice EBF. Attending breastfeeding education sessions significantly affected the EBF practice for non-migrant mothers only.

Predictors of breastfeeding practice among migrant and non-migrant mothers
We then performed logistic regression to determine the predictors of EBF, respectively. We only kept covariates with p-values smaller than 0.05 in Table 1  Odds ratios were adjusted for all the covariates in the model. * P < 0.05, ** P < 0.01, *** P < 0.001.  Odds ratios were adjusted for all the covariates in the model. * P < 0.05, ** P < 0.01, *** P < 0.001. Living in big cities was the only common variable in demographic characteristics that were significantly associated with EBF practice in both the groups. Previous literature also suggested a higher rate of EBF in big cities than in middle and small cities in China using nationally representative data derived from the Chinese National Nutrition and Health Survey conducted in 2013 [3]. This is possibly due to better knowledge and supportive environment for mothers in big cities than those living in small and medium cities regardless of their migration status.
For example, the proportion of mothers who received information on encouraging breastfeeding was higher in big cities (55.3%) than in small and medium cities (44.7%), and more than 92% of them received encouragement from the delivery hospital in big cities. This revealed that both migrant and non-migrant mothers may benefit from better breastfeedingrelated education and support, in big cities in China.
Another common predictor of EBF for both migrant and non-migrant mothers was knowledge on EBF. Some of the previous studies used knowledge scores on breastfeeding benefits as a predictor of breastfeeding practice, but they failed to provide consistent evidence on its association with EBF [11,18]. However, we only focused on the key message that infants should be exclusively breastfed for the first 6 months. This implies that future education intervention projects should pay more attention to disseminating core messages in promoting optimal breastfeeding practices.
Additionally, we also found that having a low intention of breastfeeding in the first month postpartum significantly decreased the odds of practicing EBF for both groups. Cracked nipples or nipple pain, and insufficient breastmilk supply ranked as the top two reasons, which together accounted for around half of all the answers causing low breastfeeding intention. In alignment with previous studies, nipple pain or fissure within the first month postpartum was the commonest reason for the low intention of breastfeeding among mothers regardless of their migration status [19,20]. Given that nipple pain was often attributed to incorrect positioning and attachment [20], health care providers need to find effective measures to help establish proper positioning for mothers, good attachment of the babies to the breast as well as effective sucking. In addition, insufficient milk supply was frequently reported by previous studies as the leading cause of breastfeeding cessation or its exclusivity [21,22]. However there is limited evidence on how to design interventions to address maternal perceptions or concerns of insufficient milk supply [23,24]. Further studies are needed to identify effective interventions regarding the maternal perception of milk supply.
In addition to the common predictors of EBF, as explained above, there were unique risk factors for migrant and non-migrant mothers, respectively. Among migrant mothers, having premature infants significantly decreased the prevalence of EBF. The EBF rate for migrant mothers with preterm infants was 9.8%, which was vastly lower than that of non-migrant mothers (31.0%). The low EBF rate could be attributable to mother-infant separation in neonatal intensive care units (NICUs) in China [25], which resulted in many difficulties and challenges for mothers with premature infants. The disparity in the EBF rate of premature  [27].
Implementation of KC was hindered by the limits on parental visitation, inconsistent guidelines and safety fears of the medical staff [27]. Medical practice in NICUs may need to encourage more involvement of parents to care for their infants.
Additionally, migrant mothers working in agriculture-related fields or as casual workers, were more likely to practice EBF. This is possibly because when compared with mothers working in other fields, they have more flexible time, and the shortest commuting time of 29-minutes for those who have already gone backed to work. When compared with unemployed mothers, these migrant mothers were more likely to adopt health seeking behaviors like starting to consider how to feed their infants before pregnancy and attending breastfeeding education sessions in hospitals. These findings were in alignment with previous studies, finding that mothers who worked full-time were less likely to initiate or continue breastfeeding [28]. Our results suggested that working was not necessarily a barrier for breastfeeding, allowing lactating mothers to work with a flexible schedule that may provide support for breastfeeding.
Among non-migrant mothers, EIBF and cesarean section were distinctive predictors of EBF. The rate of EIBF was 12.8% in non-migrant mothers, which was a little higher than the 9% found in a study in the Sichuan province [29], but much lower than the national China [31]. A large regional disparity of EIBF in China may be due to the delayed process of implementing Early Essential Newborn Care (EENC) in China. Even in places where EENC had been implemented, skin-to-skin contact was often interrupted due to inadequate facilities and early initiation of breastfeeding was difficult to be ensured [32]. Hospitals in China need to speed up their process of adopting EENC recommended practices and national policies are needed to be established to help hospitals and health workers overcome obstacles in implementing these practices.
For obstetrical history-related factors, we also found that the prevalence of cesarean delivery was high and negatively associated with EBF among non-migrant mothers only. The prevalence of cesarean delivery was 39.6% among non-migrant mothers, which was higher than the national average of 36.7% in 2018[33]. Even though China had made efforts in decreasing cesarean delivery in the past decade [34], the national prevalence is still far higher than the suggested prevalence of 10-15% by the WHO [35]. The government and health facilities need to make effective policies and measures in addressing non-clinical reasons for preferring a cesarean delivery, such as maternal request and perceived convenience [36].
This study derived data from a large-scale cross-sectional survey covering 8 sample sites in urban areas, which is the first and largest study to explore and compare breastfeeding practices and their potential predictors among migrant and non-migrant mothers in China.
Findings in this study filled the knowledge gap in the similarities and differences in the predictors of EBF among migrant and non-migrant mothers, which had important implications for future studies as well as health interventions to promote optimal breastfeeding practices in China.
However, our study still faced the following limitations. First, differences in the predictors of EBF among migrant and non-migrant mothers revealed in this study, may come Future studies need to take a closer look at the differences in health facility choices among migrant and non-migrant mothers and identify potential obstacles faced by the migrants.
Second, mothers covered in this study may not be representative of the mothers in general.
That is because we only interviewed mothers who brought their infants into immunization clinics at the time of the interview, which naturally excluded children who were left behind in their early childhood and may overestimate the prevalence of EBF for both groups [37,38]. A low rate of EBF would be really problematic since it will not only result in nutrition-related harm to the early childhood development of children, but also increase the health risks of mothers in China[1]. Third, when selecting sample sites in the first stage, we considered the executive capacities, and collaboration of the provincial level CDC. Thus, sample sites selected in this study to some extent reflected a higher efficiency of health systems at the county/district level. Caution is needed when interpreting these results and the results disclosed in this study cannot be generalized to the whole urban areas of China.
To conclude, this study found that there was no significant difference in the prevalence of EBF between migrant and non-migrant mothers in our sampled urban areas of China. Even though the predictors of EBF shared some similarities between migrant and non-migrant mothers, future interventions may still need to adopt different strategies in promoting optimal breastfeeding practices in different groups.

Disclaimers
The views expressed in the article are those of the authors and do not necessarily reflect the views of the institution and funder.

Competing interests
None of the authors had a conflict of interest related to any part of this study or manuscript.

Availability of data and materials
This dataset is available from the corresponding author on reasonable request.

Patient consent for publication
Not required

Acknowledgements
The authors would like to thank research teams from 12 sample sites for their hard work in the data collection. The authors also want to thank all the mothers who participated in this study.

Strengths and limitations of this study
 This is the first and largest study to explore and compare breastfeeding practices and its associate factors among migrant and non-migrant mothers in urban China.
 Mothers in large cities were more likely to practice EBF when compared with mothers in medium and small cities, regardless of their migrant status.
 The mothers covered in this study may not be representative of mothers in urban areas since we only invited mothers who brought their infants to immunization clinics to participate, which may overestimate the prevalence of exclusive breastfeeding.
 Since it is not a nationally representative study, the results cannot be generalized to all the urban areas of China.  [4]. China has experienced an unprecedented internal migration in world history. On the one hand, better health infrastructures in cities can provide migrants with better health care services [5]. On the other hand, the household registration system and limited insurance coverage put migrants in a vulnerable situation with less potential access to urban benefits [6]. In addition to socio-economic and cultural differences between migrant and non-migrant women in China, previous studies also found that migrant women had poor knowledge and limited utilization of maternal health care services [7][8][9].
A large number of previous studies have focused on immigrant mothers, who moved to a country which was not their country of origin [10]. Immigrant mothers were found to be more likely to initiate and continue breastfeeding than non-immigrant mothers. However, EBF remained a challenge for both groups [10]. Little attention has been paid to breastfeeding practices and its associated factors for domestic migrant mothers. Previous literature in China disclosed the disparities in breastfeeding practices between urban and rural areas [3,11], employed and unemployed mothers [12], Han and other minority groups [13], and highly-educated and less-educated mothers [14]. A wide range of individual, socio-demographic, cultural, psychosocial and environmental factors had been identified as risk factors of exclusive breastfeeding in China [15,16], such as maternal and child characteristics [12,13,[17][18][19], support from family members and friends [18,20,21], maternity facility education, support and practice [11,17,19,22] as well as breastfeeding intention [18]. Yet, few studies have examined the difference in breastfeeding practice and its predictors among migrant and non-migrant mothers in China. To fill this knowledge gap, we analyzed large-scale cross-sectional survey data to examine the similarities and disparities in EBF and its associated factors, among migrant and non-migrant mothers in China.

Study design and participants
Data for this study was derived from a large-scale cross-sectional survey in China Calculation of sample size was based on the expected prevalence of EBF for 0-5 monthold-infants.
We adopted a multi-stage stratified cluster sampling approach while selecting the survey sample. In the first stage, all districts/counties were categorized into 4 strata, of which 12 districts/counties were selected (4 from large cities, 4 from medium and small cities, 2 from normal rural areas and 2 from poor rural areas). The selection of districts/counties considered their geographic distributions, executive capacities, collaboration of the provincial level CDC, and their population sizes. In the second stage, four clusters were randomly selected using the probability proportional to size sampling in each selected district/county. One cluster usually had one community health center. In the last stage, mothers who brought their 0-11 month-old-infants to immunization clinics, were randomly invited to participate in the survey until the designed sample size was met. The inclusion criteria of this study were mothers of 0-11 month-old-infants who signed the informed consent form, had no psychiatric disorders, and were able to answer questions clearly. To decrease the reporting bias of feeding practice, we only invited mothers who were the primary caregivers of their infants. In total, 10,408 mothers of 0-11 month-old-infants were interviewed.

Patient and public involvement
This survey was undertaken by mothers of 0-11 month-old-infants in China to understand the prevalence of EBF and its associated factors. Neither these mothers nor the public were involved in the study design and implementation. The study results will be disseminated to the public through media briefings and scientific publications.

Data collection
This survey was conducted by the NINH, China CDC in collaboration with 12 teams at the provincial level CDC from September 2017 to January 2018. Data were collected through face-to-face interviews using a structured questionnaire programmed into smartphones or tablets. The questionnaire was developed by NINH, China CDC, which comprised eight sections: demographic characteristics, breastfeeding practices, maternal and child health, supportive environment for breastfeeding, health service, workplace, social environment and culture, and household financial situation related factors. The study design was approved by the Medical Ethics Committee at the NINH, China CDC. We obtained written consent forms from all mothers. We used the past 24 hours infant and young child feeding indicator method recommended by the World Health Organization (WHO) to assess the feeding practice of 0-5 month-old-infants to generate internationally comparable rates of exclusive breastfeeding. We defined the practice of being fed exclusively with breast milk in the last 24-hours as exclusive breastfeeding. Even though previous studies indicated that providing prelacteal feeds is a long-held tradition in many parts of China, regional disparities were quite large [23]. Additionally, water is the top one first drink received by the newborns in China and sometimes mothers revert to exclusive breastfeeding after breast milk came in [24]. Using the method can reduce the recall bias of mothers with older infants as well as the possibility of underestimating exclusive breastfeeding [24].

Statistical Analysis
We first used descriptive analysis to report the selected characteristics, and then assessed their differences between migrant and non-migrant mothers, using Pearson's  We defined mothers who were always, very often or sometimes unwilling to breastfeed within 1 month postpartum as having a low intention of breastfeeding within 1 month postpartum. Social influence denoted whether her partner, friend, mother or mother-inlaw agreed that breastmilk was better than breastmilk substitutes. Health-seeking behaviors comprised the timing of considering how to feed the infants and attending antenatal visits, breastfeeding education sessions or mother groups.
We then performed logistic regressions to identify predictors of EBF. The selection of covariate variables was based on the p-values of Chi-square tests (P<0.05).
Odds ratios and 95% confidence intervals were displayed. The data analyses were performed by Stata version 15.0 (Stata Corporation, USA). We set the level of statistical significance at α=0.05.

Sample description and factors associated with EBF
Considering that migrants usually concentrate in urban areas, we only focused on 3,571 mothers of 0-5 month-old-infants who lived in big, medium, and small cities.
Migrant mothers were defined as those who had resided in a specific county for a month or more and had been in a place different from their registered county. We excluded   Even though more than 80% of mothers received positive influence from their partners, friends, mothers or mothers-in-law about breastmilk being better than breastmilk substitutes, these influences differed among migrant and non-migrant mothers. The influence from partner and friends was significantly associated with EBF for non-migrant mothers, while the influence from their partners and their own mothers were significantly associated with EBF for migrant mothers.
Non-migrant mothers were more likely to adopt health-seeking behaviors like considering how to feed their infants before pregnancy and attending antenatal visits, breastfeeding education sessions, and mother groups, than migrant mothers. Only the timing of considering how to feed their infants was statistically different between the two groups. Migrant mothers who considered how to feed their infants before pregnancy were more likely to practice EBF. Attending breastfeeding education sessions significantly affected the EBF practice for non-migrant mothers only.

Predictors of breastfeeding practice among migrant and non-migrant mothers
We then performed logistic regression to determine the predictors of EBF, respectively. We only kept covariates with p-values smaller than 0.05 in Table 1  Odds ratios were adjusted for all the covariates in the model. * P < 0.05, ** P < 0.01, *** P < 0.001.  Odds ratios were adjusted for all the covariates in the model. * P < 0.05, ** P < 0.01, *** P < 0.001. Living in big cities was the only common variable in demographic characteristics that were significantly associated with EBF practice in both groups. Previous literature also suggested a higher rate of EBF in big cities than in middle and small cities in China using nationally representative data derived from the Chinese National Nutrition and Health Survey conducted in 2013 [3]. This is possibly due to better knowledge and supportive environment for mothers in big cities than those living in small and medium cities regardless of their migration status.
For example, the proportion of mothers who received information on encouraging breastfeeding was higher in big cities (55.3%) than in small and medium cities (44.7%), and more than 92% of them received encouragement from the delivery hospital in big cities. This revealed that both migrant and non-migrant mothers may benefit from better breastfeedingrelated education and support, in big cities in China.
Another common predictor of EBF for both migrant and non-migrant mothers was knowledge on EBF. Some of the previous studies used knowledge scores on breastfeeding benefits as a predictor of breastfeeding practice, but they failed to provide consistent evidence  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y 20 on its association with EBF [11,18]. However, we only focused on the key message that infants should be exclusively breastfed for the first 6 months. This implies that future education intervention projects should pay more attention to disseminating core messages in promoting optimal breastfeeding practices.
Additionally, we also found that having a low intention of breastfeeding in the first month postpartum significantly decreased the odds of practicing EBF for both groups. Cracked nipples or nipple pain, and insufficient breastmilk supply ranked as the top two reasons, which together accounted for around half of all the answers causing low breastfeeding intention. In alignment with previous studies, nipple pain or fissure within the first month postpartum was the commonest reason for the low intention of breastfeeding among mothers regardless of their migration status [27,28]. Given that nipple pain was often attributed to incorrect positioning and attachment [28], health care providers need to find effective measures to help establish proper positioning for mothers, good attachment of the babies to the breast as well as effective sucking. In addition, insufficient milk supply was frequently reported by previous studies as the leading cause of breastfeeding cessation or its exclusivity [29,30]. However, there is limited evidence on how to design interventions to address maternal perceptions or concerns of insufficient milk supply [31,32]. Further studies are needed to identify effective interventions regarding the maternal perception of milk supply.
In addition to the common predictors of EBF, as explained above, there were unique risk factors for migrant and non-migrant mothers, respectively. Among migrant mothers, having premature infants significantly decreased the prevalence of EBF. The EBF rate for migrant mothers with preterm infants was 9.8%, which was vastly lower than that of non-migrant mothers (31.0%). The low EBF rate could be attributable to mother-infant separation in neonatal intensive care units (NICUs) in China [33], which resulted in many difficulties and challenges for mothers with premature infants. The disparity in the EBF rate of premature  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   21   infants may come from the different choices of hospitalization services between migrant and non-migrant mothers. Even though migrant mothers are more likely to deliver their infants in higher level hospitals in general, the proportion of migrant mothers with premature infants who delivered in hospitals at the municipal or higher levels was 65.9%, lower than that of nonmigrant mothers (73.6%). In China, high-level hospitals are believed to have well-trained health workers and high-quality medical care [34]. An internationally well-known practice such as Kangaroo Care (KC), has only been implemented in some NICUs of high-level hospitals in China as pilot studies, for promoting the development of premature infants [35].
Implementation of KC was hindered by the limits on parental visitation, inconsistent guidelines, and safety fears of the medical staff [35]. Medical practice in NICUs may need to encourage more involvement of parents to care for their infants.
Additionally, migrant mothers working in agriculture-related fields or as casual workers were more likely to practice EBF. This is possibly because when compared with mothers working in other fields, they have more flexible time, and the shortest commuting time of 29-minutes for those who have already gone back to work. When compared with unemployed mothers, these migrant mothers were more likely to adopt health-seeking behaviors like starting to consider how to feed their infants before pregnancy and attending breastfeeding education sessions in hospitals. These findings were in alignment with previous studies, finding that mothers who worked full-time were less likely to initiate or continue breastfeeding [36]. Our results suggested that working was not necessarily a barrier to breastfeeding. Allowing lactating mothers to work with a flexible schedule that may provide support for breastfeeding.
Among non-migrant mothers, EIBF and cesarean section were distinctive predictors of EBF. The rate of EIBF was 12.8% in non-migrant mothers, which was a little higher than the 9% found in a study in the Sichuan province [37], but much lower than the national  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   22 prevalence of 28.7% in 2013 [38] and the 59.4% found in a study in central and western China [39]. A large regional disparity of EIBF in China may be due to the delayed process of implementing Early Essential Newborn Care (EENC) in China. Even in places where EENC had been implemented, skin-to-skin contact was often interrupted due to inadequate facilities and early initiation of breastfeeding was difficult to be ensured [40]. Hospitals in China need to speed up their process of adopting EENC recommended practices and national policies are needed to be established to help hospitals and health workers overcome obstacles in implementing these practices.
For obstetrical history-related factors, we also found that the prevalence of cesarean delivery was high and negatively associated with EBF among non-migrant mothers only. The prevalence of cesarean delivery was 39.6% among non-migrant mothers, which was higher than the national average of 36.7% in 2018 [41]. Even though China had made efforts in decreasing cesarean delivery in the past decade[42], the national prevalence is still far higher than the suggested prevalence of 10-15% by the WHO [43]. The government and health facilities need to make effective policies and measures in addressing non-clinical reasons for preferring a cesarean delivery, such as maternal request and perceived convenience [44].
This study derived data from a large-scale cross-sectional survey covering 8 sample sites in urban areas, which is the first and largest study to explore and compare breastfeeding practices and their potential predictors among migrant and non-migrant mothers in China.
Findings in this study filled the knowledge gap in the similarities and differences in the predictors of EBF among migrant and non-migrant mothers, which had important implications for future studies as well as health interventions to promote optimal breastfeeding practices in China.
Future studies need to take a closer look at the differences in health facility choices among migrant and non-migrant mothers and identify potential obstacles faced by the migrants.
Second, mothers covered in this study may not be representative of the mothers in general.
That is because we only interviewed mothers who brought their infants into immunization clinics at the time of the interview, which naturally excluded children who were left behind in their early childhood and may overestimate the prevalence of EBF for both groups [45,46]. A low rate of EBF would be really problematic since it will not only result in nutrition-related harm to the early childhood development of children, but also increase the health risks of mothers in China[1]. Third, when selecting sample sites in the first stage, we considered the executive capacities, and collaboration of the provincial level CDC. Thus, sample sites selected in this study to some extent reflected a higher efficiency of health systems at the county/district level. Caution is needed when interpreting these results and the results disclosed in this study cannot be generalized to the whole urban areas of China.

Disclaimers
The views expressed in the article are those of the authors and do not necessarily reflect the views of the institution and funder.

Strengths and limitations of this study
 This is the first and largest study to explore and compare breastfeeding practices and its associate factors among migrant and non-migrant mothers in urban China.
 Mothers in large cities were more likely to practice EBF when compared with mothers in medium and small cities, regardless of their migrant status.
 The mothers covered in this study may not be representative of mothers in urban areas since we only invited mothers who brought their infants to immunization clinics to participate, which may overestimate the prevalence of exclusive breastfeeding.
 Since it is not a nationally representative study, the results cannot be generalized to all the urban areas of China.  [2,3]. Rapid economic development is inseparable from an increasing number of domestic migrants [4]. China has experienced an unprecedented internal migration in world history. On the one hand, better health infrastructures in cities can provide migrants with better health care services [5]. On the other hand, the household registration system and limited insurance coverage put migrants in a vulnerable situation with less potential access to urban benefits [6]. In addition to socio-economic and cultural differences between migrant and non-migrant women in China, previous studies also found that migrant women had poor knowledge and limited utilization of maternal health care services [7][8][9].
A large number of previous studies have focused on immigrant mothers, who moved to a country that was not their country of origin [10]. Immigrant mothers were found to be more likely to initiate and continue breastfeeding than non-immigrant mothers. However, EBF remained a challenge for both groups [10]. Little attention has been paid to breastfeeding practices and its associated factors for domestic migrant mothers. Previous literature in China disclosed the disparities in breastfeeding practices between urban and rural areas [3,11], employed and unemployed mothers [12], Han and other minority groups [13], and highly-educated and less-educated mothers [14]. A wide range of individual, socio-demographic, cultural, psychosocial and environmental factors had been identified as risk factors of exclusive breastfeeding in China [15,16], such as maternal and child characteristics [12,13,[17][18][19], support from family members and friends [18,20,21], maternity facility education, support and practice [11,17,19,22] as well as breastfeeding intention [18]. Yet, few studies have examined the difference in breastfeeding practice and its predictors among migrant and non-migrant mothers in China. To fill this knowledge gap, we analyzed large-scale cross-sectional survey data to examine the similarities and disparities in EBF and its associated factors, among migrant and non-migrant mothers in China.

Study design and participants
Data for this study was derived from a large-scale cross-sectional survey in China Calculation of sample size was based on the expected prevalence of EBF for 0-5 monthold-infants.
We adopted a multi-stage stratified cluster sampling approach while selecting the survey sample. In the first stage, all districts/counties were categorized into 4 strata, of which 12 districts/counties were selected (4 from large cities, 4 from medium and small cities, 2 from normal rural areas and 2 from poor rural areas). The selection of districts/counties considered their geographic distributions, executive capacities, collaboration of the provincial level CDC, and their population sizes. In the second stage, four clusters were randomly selected using the probability proportional to size sampling in each selected district/county. One cluster usually had one community health center. In the last stage, mothers who brought their 0-11 month-old-infants to immunization clinics, were randomly invited to participate in the survey until the designed sample size was met. The inclusion criteria of this study were mothers of 0-11 month-old-infants who signed the informed consent form, had no psychiatric disorders, and were able to answer questions clearly. To decrease the reporting bias of feeding practice, we only invited mothers who were the primary caregivers of their infants. In total, 10,408 mothers of 0-11 month-old-infants were interviewed.

Patient and public involvement
This survey was undertaken by mothers of 0-11 month-old-infants in China to understand the prevalence of EBF and its associated factors. Neither these mothers nor the public were involved in the study design and implementation. The study results will be disseminated to the public through media briefings and scientific publications.

Data collection
This survey was conducted by the NINH, China CDC in collaboration with 12 teams at the provincial level CDC from September 2017 to January 2018. Data were collected through face-to-face interviews using a structured questionnaire programmed into smartphones or tablets. The questionnaire was developed by NINH, China CDC, which comprised eight sections: demographic characteristics, breastfeeding practices, maternal and child health, supportive environment for breastfeeding, health service, workplace, social environment and culture, and household financial situation related factors. The study design was approved by the Medical Ethics Committee at the NINH, China CDC. We obtained written consent forms from all mothers. We used the past 24 hours infant and young child feeding indicator method recommended by the World Health Organization (WHO) to assess the feeding practice of 0-5 month-old-infants to generate internationally comparable rates of exclusive breastfeeding. We defined the practice of being fed exclusively with breast milk in the last 24-hours as exclusive breastfeeding. Even though previous studies indicated that providing prelacteal feeds is a long-held tradition in many parts of China, regional disparities were quite large [23]. Additionally, water is the top one first drink received by the newborns in China and sometimes mothers revert to exclusive breastfeeding after breast milk came in [24]. Using the method can reduce the recall bias of mothers with older infants as well as the possibility of underestimating exclusive breastfeeding [24].

Statistical Analysis
We first used descriptive analysis to report the selected characteristics, and then assessed their differences between migrant and non-migrant mothers, using Pearson's Chi-square test. Potential predictors were selected through reviewing relevant literature [3,[12][13][14][15]18,25]. Chi-square tests were then used to examine the potential predictors of EBF for migrant and non-migrant mothers, respectively.

Sample description and factors associated with EBF
Considering that migrants usually concentrate in urban areas, we only focused on 3,571 mothers of 0-5 month-old-infants who lived in big, medium, and small cities.
Mothers in both groups were more likely to practice EBF if they knew its definition and were less likely to do it if their intentions of breastfeeding were low within 1 month postpartum.
Even though more than 80% of mothers received positive influence from their partners, friends, mothers or mothers-in-law about breastmilk being better than breastmilk substitutes, these influences differed among migrant and non-migrant mothers. The influence from partner and friends was significantly associated with EBF for non-migrant mothers, while the influence from their partners and their own mothers were significantly associated with EBF for migrant mothers.
For example, the proportion of mothers who received information on encouraging breastfeeding was higher in big cities (55.3%) than in small and medium cities (44.7%), and more than 92% of them received encouragement from the delivery hospital in big cities. This revealed that both migrant and non-migrant mothers may benefit from better breastfeedingrelated education and support, in big cities in China.
Another common predictor of EBF for both migrant and non-migrant mothers was knowledge on EBF. Some of the previous studies used knowledge scores on breastfeeding benefits as a predictor of breastfeeding practice, but they failed to provide consistent evidence  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y 20 on its association with EBF [11,18]. However, we only focused on the key message that infants should be exclusively breastfed for the first 6 months. This implies that future education intervention projects should pay more attention to disseminating core messages in promoting optimal breastfeeding practices.
Additionally, we also found that having a low intention of breastfeeding in the first month postpartum significantly decreased the odds of practicing EBF for both groups. Cracked nipples or nipple pain, and insufficient breastmilk supply ranked as the top two reasons, which together accounted for around half of all the answers causing low breastfeeding intention. In alignment with previous studies, nipple pain or fissure within the first month postpartum was the commonest reason for the low intention of breastfeeding among mothers regardless of their migration status [27,28]. Given that nipple pain was often attributed to incorrect positioning and attachment [28], health care providers need to find effective measures to help establish proper positioning for mothers, good attachment of the babies to the breast as well as effective sucking. In addition, insufficient milk supply was frequently reported by previous studies as the leading cause of breastfeeding cessation or its exclusivity [29,30]. However, there is limited evidence on how to design interventions to address maternal perceptions or concerns of insufficient milk supply [31,32]. Further studies are needed to identify effective interventions regarding the maternal perception of milk supply.
In addition to the common predictors of EBF, as explained above, there were unique risk factors for migrant and non-migrant mothers, respectively. Among migrant mothers, having premature infants significantly decreased the prevalence of EBF. The EBF rate for migrant mothers with preterm infants was 9.8%, which was vastly lower than that of non-migrant mothers (31.0%). The low EBF rate could be attributable to mother-infant separation in neonatal intensive care units (NICUs) in China [33], which resulted in many difficulties and challenges for mothers with premature infants. The disparity in the EBF rate of premature non-migrant mothers. Even though migrant mothers are more likely to deliver their infants in higher level hospitals in general, the proportion of migrant mothers with premature infants who delivered in hospitals at the municipal or higher levels was 65.9%, lower than that of nonmigrant mothers (73.6%). In China, high-level hospitals are believed to have well-trained health workers and high-quality medical care [34]. An internationally well-known practice such as Kangaroo Care (KC), has only been implemented in some NICUs of high-level hospitals in China as pilot studies, for promoting the development of premature infants [35].
Implementation of KC was hindered by the limits on parental visitation, inconsistent guidelines, and safety fears of the medical staff [35]. Medical practice in NICUs may need to encourage more involvement of parents to care for their infants.
Additionally, migrant mothers working in agriculture-related fields or as casual workers were more likely to practice EBF. This is possibly because when compared with mothers working in other fields, they have more flexible time, and the shortest commuting time of 29-minutes for those who have already gone back to work. When compared with unemployed mothers, these migrant mothers were more likely to adopt health-seeking behaviors like starting to consider how to feed their infants before pregnancy and attending breastfeeding education sessions in hospitals. These findings were in alignment with previous studies, finding that mothers who worked full-time were less likely to initiate or continue breastfeeding [36]. Our results suggested that working was not necessarily a barrier to breastfeeding. Allowing lactating mothers to work with a flexible schedule that may provide support for breastfeeding.
Among non-migrant mothers, EIBF and cesarean section were distinctive predictors of EBF. The rate of EIBF was 12.8% in non-migrant mothers, which was a little higher than the 9% found in a study in the Sichuan province [37], but much lower than the national  [38] and the 59.4% found in a study in central and western China [39]. A large regional disparity of EIBF in China may be due to the delayed process of implementing Early Essential Newborn Care (EENC) in China. Even in places where EENC had been implemented, skin-to-skin contact was often interrupted due to inadequate facilities and early initiation of breastfeeding was difficult to be ensured [40]. Hospitals in China need to speed up their process of adopting EENC recommended practices and national policies are needed to be established to help hospitals and health workers overcome obstacles in implementing these practices.
For obstetrical history-related factors, we also found that the prevalence of cesarean delivery was high and negatively associated with EBF among non-migrant mothers only. The prevalence of cesarean delivery was 39.6% among non-migrant mothers, which was higher than the national average of 36.7% in 2018 [41]. Even though China had made efforts in decreasing cesarean delivery in the past decade[42], the national prevalence is still far higher than the suggested prevalence of 10-15% by the WHO [43]. The government and health facilities need to make effective policies and measures in addressing non-clinical reasons for preferring a cesarean delivery, such as maternal request and perceived convenience [44].
This study derived data from a large-scale cross-sectional survey covering 8 sample sites in urban areas, which is the first and largest study to explore and compare breastfeeding practices and their potential predictors among migrant and non-migrant mothers in China.
Findings in this study filled the knowledge gap in the similarities and differences in the predictors of EBF among migrant and non-migrant mothers, which had important implications for future studies as well as health interventions to promote optimal breastfeeding practices in China.
However, our study still faced the following limitations. First, differences in the predictors of EBF among migrant and non-migrant mothers revealed in this study, may come Future studies need to take a closer look at the differences in health facility choices among migrant and non-migrant mothers and identify potential obstacles faced by the migrants.
Second, mothers covered in this study may not be representative of the mothers in general.
That is because we only interviewed mothers who brought their infants into immunization clinics at the time of the interview, which naturally excluded children who were left behind in their early childhood and may overestimate the prevalence of EBF for both groups [45,46]. A low rate of EBF would be really problematic since it will not only result in nutrition-related harm to the early childhood development of children, but also increase the health risks of mothers in China[1]. Third, when selecting sample sites in the first stage, we considered the executive capacities, and collaboration of the provincial level CDC. Thus, sample sites selected in this study to some extent reflected a higher efficiency of health systems at the county/district level. Caution is needed when interpreting these results and the results disclosed in this study cannot be generalized to the whole urban areas of China. Fourth, responses of the mothers may be influenced by social desirability bias since data collection occurred in community health centers, where mothers usually received education on EBF.