The current study reveals that the EBF rate, as the measurement of exclusively breastfed infants from birth to six months, is 71.3% among the mothers of infants aged 7-12 months within our maternal health WeChat groups. However, considering that the participants were only from the WeChat groups, further comparative study is needed to confirm the role of WeChat groups in promoting breastfeeding.

Health workers have played an important role in breastfeeding initiation and continuation [4,5,6,7,8,9]. In the traditional model of maternal health care, women are required to attend antenatal and postnatal clinics or classes regularly, and health workers are needed to follow up the mothers by telephone calls or home visits after their hospital discharge. This could decrease the compliance of women and increase the workload of health workers, which in turn diminishes the effectiveness of maternal health care. Because WeChat is now universally used by almost all women of childbearing age in our area, we adopt the establishment of maternal health WeChat groups as part of routine maternal health care. To facilitate management, we have established seven WeChat groups according to the administrative regions, namely, one WeChat group is responsible for the perinatal women of a county or district. In this way, health workers can provide regular breastfeeding education for all women and individual counseling for those who are experiencing feeding problems. The higher awareness rate (86.1%) of all three breastfeeding WHO recommendations in this study also reflects the effectiveness of this model of health education. In addition, the mothers within the group can also share their breastfeeding experiences with each other, thereby enhancing their breastfeeding confidence.

We acknowledge that other factors could also contribute to this higher rate of EBF up to six months in this study. In particular, most mothers gave birth at the BFHI certificate hospitals which adopted the “Ten Steps to Successful Breastfeeding” launched by WHO [13]. Therefore, it is not surprising that a higher proportion of mothers, even for those with cesarean delivery, reported that they had skin-to-skin contact immediately after birth, rooming-in with their infants, breastfeeding initiation within the first hour and EBF during hospitalization in our study. These interventions have all been shown to be strong contributors to establishment and continuation of breastfeeding [7, 14,15,16,17].

In line with previous studies [4, 5, 18,19,20], we found that there were significant differences in maternal age, BMI, employment status, mode of delivery, breastfeeding initiation, and perception of insufficient breast milk between mothers with and without EBF based on the univariate analyses. However, multivariable regression analysis showed that only maternal age and perception of insufficient breast milk were statistically associated with EBF up to six months, especially that perceived insufficient breast milk significantly decreased the odds of EBF from birth to six months. Regarding the relationship between maternal age and EBF, data in the literatures are conflicting. Some studies observed that older mothers were associated with lower EBF compared to younger ones [3, 5]; in contrast, other studies showed that older mothers were more likely to practice EBF than younger ones [19, 21]. In this study, maternal age was found to be negatively but marginally associated with EBF, indicating a limited effect.

The lack of associations of employment status and breastfeeding initiation with EBF up to six months may be due to the fact that the majority of surveyed mothers were employed and had breastfeeding initiation within the first hour. As for mode of delivery, previous studies suggested that women with cesarean delivery were less likely to EBF than those with vaginal delivery [3, 14]. Delayed onset of lactation, disrupted mother-infant interaction, inhibited infant suckling and poor pain relief may mediate the effects of caesarean delivery on breastfeeding [22]. In the current study, although caesarean delivery was less common in EBF mothers than in non-EBF mothers, it was not statistically associated with EBF up to six months after adjustment for other confounding variables. This result is consistent with the study by Ruan et al [5], and suggests that if the mothers receive adequate breastfeeding support during hospitalization, caesarean delivery is not necessarily a barrier to EBF. This may be especially important considering that a relatively higher caesarean delivery rate in China [23]. In addition, we did not found the differences in mothers’ educational level, monthly family incomes, skin-to-skin contact, and room-in between mothers with and without EBF, those have been identified as the factors related to EBF in previous studies [2, 3, 8, 17]. Again, sample characteristics and hospital practices may explain the discrepancies between this study and other studies.

With regard to the reasons for non-EBF up to six months, no or insufficient breast milk is the foremost one reported by the non-EBF mothers. This is consistent with the results of multivariable regression analysis and of other studies [3, 5, 6]. However, in fact, only few mothers have physiological insufficient milk supply and most mothers can produce enough breastmilk to meet their infant’s demand [24]. As such, this result may imply the inadequate education and guidance provided by health workers on this issue. The next main reasons are those work related factors, including inability to breastfeed their infants as needed after return to work and lack of flexible breaks at work. Notably, although fewer mothers (6.8%) stated no breastfeeding room or refrigerator at workplaces as the reason for non-EBF, 72.7% of the employed mothers reported there was no breastfeeding room at their workplaces. Other reasons are various, including infant crying or mother feeling tired or troubled with breastfeeding, nipple and breast problems, the concern about breast milk alone being not sufficient for infant’s needs, and perceived inconveniences or discomfort of breastfeeding in public. Fortunately, most reasons listed above can be amended through education and interventions. For example, health workers can guide mothers how to tell the difference between physiological and perceived insufficient breast milk, prepare mothers for tiredness and fatigue, improve mothers’ ability to soothe their infants, and eliminate their concern about insufficient breast milk nutrition. Returning to work before six months is still the common reason of early weaning breastfeeding for working mothers [5, 7, 9]. Thus, breastfeeding-friendly work policies and environments are needed for improving EBF among those mothers. For example, a relatively long maternity leave can extend breastfeeding duration for working mothers [7, 18, 25]. Hence, government may consider a longer paid maternity leave, guarantee frequent and flexible breaks at work, and encourage the provision of an independent breastfeeding room with a refrigerator at the workplaces. In Jiaxing, women can now have a paid maternity leave of 128 days for vaginal delivery and 143 days for cesarean delivery.

This study has some limitations. First, due to the nature of cross-sectional design, we cannot establish causal relationships between EBF and associated factors. Second, this study only enrolled the mothers within our maternal health WeChat groups, so the sample would be underrepresented. It is possible that women who joined in the WeChat groups had stronger intention to breastfeed exclusively than those outside groups. Further study, such as prospective cohort study, is therefore needed to explore the effect of WeChat groups on breastfeeding and women’s perceptions of WeChat groups by comparing women within WeChat groups with those outside. Such a study should also use the WHO definition of EBF in order to be comparable with other studies. Third, the response rate was estimated to be about 75.3%. It is possible that mothers who did not practice EBF were less willing to respond than those who did, which may lead to overestimation of the EBF rate. Fourth, because the mothers completed the questionnaire six months after their delivery, the recall bias could not be avoided. Also the self-report nature of the study may cause reporting bias. Finally, the questionnaire was designed to be relatively simple in order to increase the participation rate. There were some important factors that failed to be measured, such as mothers’ intention and attitude to breastfeeding, the time when mothers introduce complementary foods and stop breastfeeding, supports of husbands and families, etc. These variables have previously been reported as the factors affecting EBF [7, 25] and may provide more information for future breastfeeding education and interventions.