Authors: Na Wu, Yijing Wang, Lijing Ding, Yiting Xu, Xuyan Jia, Keying Huo, Xiao Yao, Wenyan Zhang, Yue Wan, Peimin Xie, Lin Ye, Rong Huang
Categories: Article, Infant, Preterm, Breastfeeding, Trajectory, Longitudinal study, Mothers
Source: International Journal of Nursing Studies Advances
Authors: Na Wu, Yijing Wang, Lijing Ding, Yiting Xu, Xuyan Jia, Keying Huo, Xiao Yao, Wenyan Zhang, Yue Wan, Peimin Xie, Lin Ye, Rong Huang
Breast milk is the nutrient of choice for preterm infants. The World Health Organization has set a target of ensuring by 2030 that 70 % of infants younger than age 6 months are exclusively breastfed. The breastfeeding status of mothers of preterm infants is far below that target. Preterm infants are a special group, and there are many difficulties in and obstacles to adhering to breastfeeding. It is urgent to understand the current situation of breastfeeding among mothers of preterm infants, the change in their breastfeeding behaviors over time, and factors influencing breastfeeding in this population during different postpartum stages.
We aimed to explore potential breastfeeding behavior trajectories and analyze modifiable influencing factors to provide a reference for health care staff to identify risk groups of breastfeeding attrition at an early stage and develop targeted intervention measures.
We conducted a prospective, single-center, longitudinal study involving a cohort of mothers of preterm infants from July 2023 to May 2024. We administered the Breastfeeding Behavior Questionnaire, Breastfeeding Intention Questionnaire, Breastfeeding Prediction Scale, Breastfeeding Knowledge Questionnaire, and Edinburgh Postnatal Depression Scale. Data were collected at four before discharge (T0) and at 6 weeks (T1), 4 months (T2), and 6 months postpartum (T3). A mixed model of latent variable growth was used to fit the trajectory of maternal breastfeeding behavior, and multiple logistic regression analysis was used to detect factors influencing the trajectory of breastfeeding behavior.
Three types of heterogeneity tracks were a "high-level rising" group (36.5 %), "medium-level declining" group (26.0 %), and "low-level declining" group (37.5 %). Household income, spouse's educational background, smoking in the spouse, breastfeeding patterns before discharge (T0), perceived breastfeeding control and breastfeeding intention (T1), perceived breastfeeding control, breastfeeding intention, and breastfeeding knowledge (T2), and perceived breastfeeding control, breastfeeding positive attitude, and breastfeeding negative attitude (T3) were predictors of different profiles.
Breastfeeding is of great importance to promote the growth and development of preterm infants. However, the breastfeeding situation among mothers of preterm infants 6 months after delivery is not encouraging. The rate of exclusive breastfeeding is low and needs to be improved. Clarifying those variables related to the Theory of Planned Behavior, such as breastfeeding positive and negative attitudes, perceived breastfeeding control, breastfeeding knowledge, and postpartum depression score, can provide a theoretical basis for the early identification and screening of mothers at risk of exclusive breastfeeding attrition. Interventions should be delivered at scale to benefit all new mothers and infants.
What is already known•The breastfeeding status of mothers of preterm infants is below the recommended levels.•Research on the developmental trajectory of breastfeeding behavior is in its infancy, and the number of studies on behavior trajectories is relatively small.
What this paper adds•We found that maternal breastfeeding self-efficacy facilitated adherence to good breastfeeding behavior.•Breastfeeding intention was an important protective factor, and breastfeeding negative attitudes may put women at greater risk of weaning at 6 weeks postpartum.•Three types of heterogeneity development tracks were "high-level rising" (36.5 %), "medium-level declining" (26.0 %), and "low-level declining" (37.5 %) group.•We have provided a reference for health care workers to identify risk groups early and develop intervention measures.
In preterm infants, the respiratory, circulatory, nervous, and other bodily systems and related organs are insufficiently mature. Complications such as neonatal respiratory distress syndrome, neonatal hypoglycemia, feeding intolerance, and necrotizing enteritis are more likely to occur. There is also a risk of long-term neurodysplasia and chronic digestive diseases (Vaidya et al., 2022; Wang et al., 2024). The mother's breast milk is the best choice for preterm infant feeding. Researchers have shown that the protein content of breast milk is much higher in mothers of preterm infants than that of mothers of full-term infants because it contains the appropriate calcium and phosphorus ratio to meet the rapid growth needs of preterm infants (Kemp et al., 2023). Breast milk from these mothers also has a certain therapeutic effect, helping to improve short-term and long-term quality of life in preterm infants (Sahin et al., 2020). The World Health Organization and the United Nations Children's Fund advocate that all infants, including preterm infants, should be breastfed as soon as possible within 1 hour of birth and exclusively breastfed for 6 months after birth (WHO, 2003). The goal is to exclusively breastfeed 50 % of infants worldwide up to 6 months of age by 2025 and 70 % of infants up to 6 months of age by 2030 (WHO, 2008).
The outlook for breastfeeding during hospitalization and after discharge in preterm infants is not optimistic, both in China and abroad. Preterm infants have more practical difficulties in breastfeeding than full-term infants because they have a longer hospital stay, immature organ development, weak sucking ability, and separation from their mothers. Researchers in a cohort study of 3217 preterm infants from 11 European countries showed that the rate of exclusive breastfeeding at the time of discharge was 38 % in 2018 (Bonnet et al., 2019). In a longitudinal study of 270 preterm infants in Australia (Keir et al., 2022), the rates of exclusive breastfeeding at discharge, 6 weeks, 3 months, and 6 months (corrected age) were 74 %, 41 %, 35 %, and 29 %, respectively. The current situation of breastfeeding in preterm infants urgently requires further scientific research and guidance support.
Researchers have pointed out that lower gestational age and small for gestational age are associated with decreased rates of exclusive breastfeeding (Nejsum et al., 2023; Oliveira and Valle Volkmer, 2021). Breastfeeding behavior refers to the feeding of infants with breast milk exclusively, breast milk partially, or with breast milk substitute; the indicator of the choice regarding breastfeeding is the feeding pattern during the 24 hours prior to survey (WHO, 2021). At present, the definition of breastfeeding behavior is a relatively broad summative definition of breastfeeding modes and technologies, as an innate practice and acquired behavior (Azimi and Nasiri, 2020). Breastfeeding of preterm infants affects maternal breastfeeding experiences and may influence maternal breastfeeding decisions in mothers of preterm infants (Jonsdottir et al., 2022). Researchers in one study pointed out that mothers of preterm infants had lower breastfeeding self-efficacy than full-term mothers, and breastfeeding self-efficacy at 5 months postpartum was positively correlated with exclusive breastfeeding behavior among mothers (Ahmed and Rojjanasrirat, 2021). Factors that affect breastfeeding include individual factors (self-reported insufficient milk; parents and health professionals frequently misinterpret typical, unsettled baby behaviors as signs of milk insufficiency or inadequacy). Structural barriers that undermine the breastfeeding environment (Ringheim, 1993) include gender inequities, harmful sociocultural infant feeding norms, economic growth and urbanization, corporate marketing practices (Neves et al., 2020), and political activities that weaken breastfeeding protection policies (Pérez-Escamilla et al., 2023). In terms of breastfeeding attitudes, mothers of preterm infants who have positive attitudes toward breastfeeding after delivery are more likely to breastfeed during their hospital stay (Niela-Vilén et al., 2016). Breastfeeding intention is an important protective factor of exclusive breastfeeding for preterm infants at 6 weeks and at 3 and 6 months after delivery (Huang et al., 2023). Breastfeeding self-efficacy among mothers of preterm infants is low and is positively correlated with exclusive breastfeeding behavior in the mother at 5 months postpartum (Ahmed and Rojjanasrirat, 2021). However, a theoretical framework has rarely been used to explore the factors influencing breastfeeding in mothers of preterm infants, and relevant studies are not systematic or comprehensive. Furthermore, most researchers have adopted a cross-sectional survey design and have explored the influencing factors only at a certain point during the postpartum period, which is not conducive to identifying the trajectory of breastfeeding and influencing factors at each stage in preterm maternal risk groups with different breastfeeding interruptions.
Current researchers in the field of breastfeeding cover a variety of theoretical frameworks. The application of Self-Efficacy Theory in breastfeeding research mainly focuses on the evaluation of mothers' confidence in their own ability to breastfeed and the influence of this confidence on breastfeeding behavior by enhancing breastfeeding self-efficacy (Brockway et al., 2017). Researchers can study this area to promote positive breastfeeding behavior and provide theoretical guidance for formulating intervention strategies and support measures.
The Transtheoretical Model (Kloeblen et al., 1999) is used in the field of breastfeeding to understand the changes and progression of mothers' breastfeeding behavior. Using this model, it is argued that behavioral change is a gradual, phased process, with individuals going through multiple stages from non-preparation to maintenance.
The Theory of Planned Behavior has been widely used in breastfeeding research, which helps with improving understanding regarding the influence of individual intention, attitudes, subjective norms, and perceived behavioral control on breastfeeding behavior (Zhang et al., 2021). The Theory of Planned Behavior includes three independent determinants (behavioral attitudes, subjective norms, and perceived behavioral control), which influence actual behavior through behavioral intention. The subjective norm of breastfeeding refers to the support and opinion of people who are important, as perceived by the mother, such as family members, friends, and medical staff (Wen et al., 2021). The subjective norm of breastfeeding also reflects the psychological and social support felt by mothers. Bajoulvand et al. (2019) found that perceptual behavioral control could explain 65 % of mothers’ intention to breastfeed, and both the intention to breastfeed and perceptual behavioral control could jointly predict 79 % of mothers' changes in exclusive breastfeeding behavior.
Behavioral attitude refers to an individual's evaluation of the pros and cons of a certain behavior. Researchers have shown that breastfeeding attitudes have a greater impact on breastfeeding intentions than perceived behavioral control and subjective norms. Additionally, breastfeeding attitudes are not static, fostering mothers' positive attitudes toward breastfeeding, which is conducive to promoting breastfeeding behavior and prolonging the duration of breastfeeding (Guo et al., 2016).
The degree of perceived behavioral control refers to the perceived difficulty in performing a certain behavior, reflecting past experiences and expected obstacles (Parker et al., 2020). Application of the Theory of Planned Behavior in the field of breastfeeding behavior research is appropriate for investigating the individual, environmental, and structural influences of breastfeeding and has shown a relatively good ability to explain, predict, and improve breastfeeding behavior (Khani Jeihooni et al., 2022; Yazdanpanah et al., 2022).
In longitudinal studies of mothers of preterm infants, researchers have focused mainly on investigation of the status quo regarding exclusive breastfeeding during different postpartum periods. Breastfeeding trajectory of postpartum breastfeeding behavior is in a relatively new stage of exploration. Therefore, we applied a longitudinal design based on the Theory of Planned Behavior, and a mixed model of latent variable growth was adopted to explore the developmental trajectory of breastfeeding behavior among mothers of preterm infants at 0–6 months after delivery. We also analyzed predictors of the developmental trajectory of breastfeeding behavior to provide a basis for the formulation of intervention measures to improve breastfeeding of preterm infants.H1 There are different breastfeeding trajectories of mothers of preterm infants within 6 months after deliveryH2 There are different influencing factors of breastfeeding for mothers of preterm infants with different stages in the 6 months after delivery
We followed a prospective longitudinal design using breastfeeding behavior questionnaires as core indexes to measure the level of breastfeeding behaviors among mothers of preterm infants. A mixed model of latent variable growth was used to fit the trajectory of maternal breastfeeding behavior, and regression analysis was used to explore potential factors influencing the development trajectory of maternal breastfeeding behavior.
The normal initiation of milk production, which is the basis for the woman's implementation of breastfeeding behavior, occurs within 24–72 h postpartum during the hospital stay of the postpartum woman. Therefore, the first survey time point was set during the postpartum woman's hospital stay before discharge(T0). The postpartum period lasts for 42 days, during which the mother is in the critical phase of recovery. Breastfeeding attrition occurs in 32 % to 58 % of breastfeeding mothers within the first 6 weeks. Therefore, the second survey was conducted 42 days after delivery, as this period is crucial for establishing breastfeeding behavior and breastfeeding continuation (T1). When babies add complementary foods has been a topic of current interest for pediatricians to discuss; usually the time ranges from 4 to 6 months of age. Some women as early as 4 months postpartum add complementary foods, which may reduce the infant's exclusive breastfeeding rate, affecting the maternal exclusive breastfeeding behavior. Therefore, the time point of the third survey was set at 4 postpartum months(T2). The Global Infant and Young Child Feeding Strategy formulated by World Health Organization and United Nations International Children's Emergency Fund (WHO, 2003)advocates exclusive breastfeeding within 6 months after birth, so the time point of the fourth survey in this study was set at 6 months postpartum(T3).
The study hospital has two campuses with a total of nearly 25,000 births annually, covering an area of 60 acres with a building area of 71 m^2^. The east campus has more than 300 beds in the obstetrics department, with 1500 preterm infants birth per year.
This was an observational study. According to the principle of sample size estimation in multivariate statistical analysis, the sample size was 5–10 times the observed variables (Althubaiti, 2023). The variables explored in this study general data (24 items), one dimension of the Breastfeeding Behavior Questionnaire, one dimension of the Breastfeeding Knowledge Questionnaire, one dimension of the Breastfeeding Intention Questionnaire, four dimensions of the Breastfeeding Prediction Scale, and one dimension of the Edinburgh Postnatal Depression Scale. A total of 32 variables were addressed in this study, and the required sample size was 160–320 participants. Considering a 20 % rate for loss to follow-up, the required sample size was 192–384 mothers of preterm infants.
A convenience sampling method was used to recruit participants who met the exclusion and inclusion criteria. This study was conducted at a Grade 3 Class A obstetrics and gynecology hospital in Shanghai from July 2023 to May 2024. The inclusion criteria (1) mothers aged ≥20 years who gave birth to a single live infant and (2) gestational age 28 to <37 weeks. The exclusion criteria (1) participants with acute, chronic infectious diseases or with other physical conditions making breastfeeding unsuitable, (2) participants serious mental and psychological diseases; (3) participants who refused to breastfeed; and (4) infants with congenital deficiencies (e.g., cleft lip and palate) who could not be breastfed.
This study was approved by the Hospital Ethics Committee (approval number KS22343). We followed the ethics rules regarding voluntary participation and confidentiality, as well as providing benefit and no harm to the participant. Written informed consent was obtained before the start of the survey. Each participant was informed that their participation was voluntary and that they had the right to withdraw from participation at any point during the data collection period.
The Breastfeeding Behavior Questionnaire developed by Alami Ali (2014) is a standardized questionnaire based on the Theory of Planned Behavior theory. The questionnaire includes breastfeeding intention (three items) and breastfeeding behavior (four items). The content validity range of the questionnaire is 0.65–0.99, Cronbach's α coefficient is 0.79, and the intra-group correlation coefficient is 0.81, showing moderate internal consistency. The four items are as follows. ① Do you feed your baby only with your breast milk? ② Do you feed your baby only with formula milk? ③ Do you feed your baby with animal milk (cow/goat milk)? ④ Do you feed your baby only with sugar water or rice soup? A 5-point Likert scoring method was adopted, with scores of 1 and 5 indicating “strongly disagree” and “strongly agree,” respectively. Participants’ actual breastfeeding behavior was assessed using these scores.
The Infant Feeding Intention Scale was designed and developed by Laurie Nommsen-Rivers (Nommsen-Rivers and Dewey, 2009) to measure the willingness to breastfeed, with a Cronbach's α coefficient of 0.90. In this study, we used a portion of the scale, consisting of the following three items. Do you plan to feed your baby formula only (no breastfeeding at all)? Do you plan to try to breastfeed at least once? Do you plan to exclusively breastfeed for 6 months after delivery? A 5-point Likert scale was adopted, with scores of 0 and 4 indicating “strongly disagree” and “strongly agree,” respectively. The total score was calculated according to the sum of the average score of items 1–3, and intention to breastfeed was assessed using the score level. The content validity was 0.92.
This scale (Janke, 1992) was developed following the Theory of Planned Behavior. The Breastfeeding Attrition Prediction Tool has 52 items, including four the Breastfeeding Positive Attitudes Scale (14 items), Breastfeeding Negative Attitudes Scale (15 items), Social and Professional Support Scale (13 items), and Breastfeeding Self-Control Scale (10 items). The original scale uses a 6-point Likert scoring method, with a Cronbach's α coefficient 0.79–0.85. In this study, we adopted the Chinese version of the Breastfeeding Prediction Scale (Chenting, 2023), which contains four dimensions and a total of 48 items. The four dimensions are breastfeeding positive attitudes (12 items), breastfeeding negative attitudes (15 items), social and professional support (11 items), and breastfeeding self-control (10 items). A 5-point Likert scale was adopted for the Chinese version of the Breastfeeding Attrition Prediction Tools scale. The Cronbach's α coefficient of this scale is 0.903, and the Cronbach's α coefficient of each dimension is 0.878–0.931.
This questionnaire (Ouyang et al., 2012) was developed to evaluate the mastery of knowledge about breastfeeding. There are 18 items on the questionnaire. The Cronbach's α and content validity of the questionnaire are 0.82 and 0.87. If an answer is wrong, 0 points are given. The score is the total for the scores on each item, with a total score of 18 points. The higher the score, the greater the knowledge reserve and the better the mastery of breastfeeding knowledge. The KR-21 (Kuder-Richardson formula 21) used to assess internal consistency was 0.779.
The Edinburgh Postnatal Depression Scale (Cox et al., 1987) was developed to screen for postpartum depression and is currently the preferred scale for evaluating postpartum depression. The total score on the scale ranges from 0 to 30 points. The higher the score, the greater the level of depression or the greater the risk of depression. The criteria for screening postpartum depression are scores ≤5, which indicate no or very mild postpartum depression; scores of 6–9 indicate mild postpartum depression, and ≥10 points indicate moderate to severe postpartum depression.
Data were collected between July 2023 to May 2024. Researchers with backgrounds in nursing education were recruited and trained in the research objectives, informed consent, and research process. If participants raised questions during the process of completing the questionnaire, standard language was used to explain the meaning of any questionnaire items. The survey included sociodemographic characteristics, as well as the Breastfeeding Behavior Questionnaire, Infant Feeding Intention Scale, Breastfeeding Attrition Prediction Tools, Breastfeeding Knowledge Questionnaire, and Edinburgh Postnatal Depression Scale. Data were collected in four waves, including the day after delivery during hospitalization, and during three follow-up visits conducted at 6 weeks and 4 and 6 months after delivery via WeChat or phone.
In descriptive analysis of general demographic data and breastfeeding-related variables, mean ± standard deviation was used for the statistical description of continuous variables conforming to a normal distribution. Median and interquartile range was used for continuous variables not conforming to a normal distribution, and frequency and percentage was used for categorical variables. Scores on the Breastfeeding Behavior Questionnaire, Infant Feeding Intention Scale, Breastfeeding Attrition Prediction Tools, Breastfeeding Knowledge Questionnaire scale, and Edinburgh Postnatal Depression Scale measured in the four stages of parturiency were statistically described using mean ± standard deviation for continuous variables, and one-way repeated measures analysis of variance (ANOVA) was performed for scores of breastfeeding behaviors at the different stages. If the p value of the Mauchly spherical test was <0.05, the Greenhouse–Geisser method was used for correction.
Mplus 8.0 software was used to build a mixed model of latent variable growth to fit the track of breastfeeding behavior for participants. First, it was assumed that the model had one potential category. Then, we built a model with one potential category, followed by a model with two potential categories, and then a mixed model of latent variable growth with three and four potential categories. Univariate and regression analysis were used to explore potential factors influencing the development trajectory of maternal breastfeeding behavior.
Participants completed four surveys in total. The first was a baseline survey after delivery during hospitalization. (T0). The second survey was conducted at 6 weeks postpartum (T1). The third survey was conducted at 4 months postpartum (T2). The fourth survey was conducted at 6 months postpartum (T3). If the woman did not respond after five attempts to contact her, she was considered lost to follow-up. If a participant explicitly refused to participate in the follow-up, she was also considered lost to continuing during the next three follow-ups. During follow-up, if serious organ dysfunction or death occurred in a preterm infant, follow-up with the mother was stopped (Supplementary Material-follow-up flow chart).
A total of 206 participants completed all questionnaires during the entire study process. Table 1 shows the sociodemographic characteristics of study participants. A total of 46 participants were lost to follow-up; the follow-up rate was 81.7 % (206/252). The final loss to follow-up rate was 18.25 % (46/252). There were no significant differences in sociodemographic characteristics between participants included in this study and those lost to follow-up.Table 1Socio-demographic characteristics of the participants(N = 206).Table 1VariableFrequency (Mean ± Standard Deviation)Proportion (%)Age (years)32.33 ± 3.9020–295024.330–3914771.4≥4094.4EthnicityHan20197.6other52.4MarriageYes20499.0No21.0BeliefYes18790.8No199.2EducationJunior high school104.9High school199.2Undergraduate13063.1Postgraduate and above4722.8ResidenceCity15776.2Town3014.6Urban199.2Household income (Yuan)≤500083.95001∼10,0006431.110,001∼15,0005124.8>15,0008340.3EmployeeInstitution209.7Government2914.1Liberal profession3918.9Company11857.3Way of paymentBasic health insurance17785.9self-financing2914.1Maternity leave≤158 days14168.4159∼173 days4320.9≥174 days2210.7Return to workYes18187.9No2512.1Primary caregiverMother3617.5Mother-inlaw6933.5Nanny5828.2Confinement center*178.3Other2612.7Spouse age (Year)33.84 ± 4.4820–293316.030–3915072.8≥402311.2Spouse educationJunior high school83.9High school2311.2Undergraduate12460.2Postgraduate and above5124.8Spouse smokingYes6230.1No14469.9Gestational week (day)≤223188.7224⁓2373215.5238⁓25815675.7First sucking time after delivery30 min10.5With 1 h62.9With 2 h31.52 h after delivery5828.2Maternal separation13867.0Breastfeeding times<815374.3≥85325.7Breastfeeding patterns before dischargeExclusive breastfeeding3316.0Mixed feeding11756.8Formula feeding5627.2Breastfeeding plan after dischargeExclusive breast milk12761.7Mixed feeding7536.4Formula feeding41.9Duration of breastfeeding planTwo months41.9Four Months146.8Six Months7235.0One year7034.0Two years41.9Spontaneous cessation of breastfeeding3918.9Other31.5Cracked nippleYes178.3No18991.7Baby-care rooms in public placesYes13766.5No6933.5Had given birthYes6029.1No14670.9Note: Confinement A facility that typically offers professional postpartum recovery services, also referred to as confinement services, for new mothers. These services include providing confinement meals by professional nutritionists to aid in the mother's rapid physical recovery, imparting feeding knowledge, and having professional nurses to care for the babies.N = The number of participants.
The maternal breastfeeding behavior score showed a gradual increasing trend in T0 and T1, reaching the highest at T1; the scale score was 16.33 ± 3.07. Then, the score showed a gradual decreasing trend from T1 to T2 and T3, reaching the lowest score at T3; the scale score was 13.43 ± 4.99. The average score of each item was basically consistent with the total score in the Breastfeeding Behavior Questionnaire during the different periods. The results showed that the data did not meet the spherical hypothesis (W = 0.84, p < 0.05). There were significant differences in the scores for breastfeeding behaviors between the four periods (F = 34.11, p < 0001).
Repeated measures ANOVA was performed for each variable during the different periods. The results of one-way repeated measures ANOVA showed that with the change in postpartum time, there were statistically significant differences in breastfeeding intention, breastfeeding knowledge, breastfeeding positive attitudes, breastfeeding negative attitudes, perceived breastfeeding control, and social and professional support for preterm infants during the different periods (p < 0.05). With increased postpartum time, the breastfeeding intention of participants was weakened, and social and professional support showed a downward trend. The results are shown in Table 2.Table 2Maternal TPB-related variables at different periods, breastfeeding knowledge and postpartum depression scores (Mean ± standard deviation).Table 2VariableT0T1T2T3Intergroup factorF(pη²*)p*Breastfeeding intention6.58 ± 1.396.52 ± 1.496.07 ± 1.83——8.186(0.038)0.000Breastfeeding positive attitude250.57 ± 49.08251.36 ± 61.22249.71 ± 67.51225.21 ± 64.1212.879(0.059)0.000Breastfeeding negative attitude197.14 ± 58.88189.93 ± 57.00196.25 ± 71.72227.05 ± 89.5110.236(0.131)0.000Social and professional support163.61 ± 61.25156.77 ± 50.71149.97 ± 51.81137.33 ± 50.0411.271(0.143)0.000Perceived breastfeeding control38.33 ± 6.7439.20 ± 6.3639.33 ± 7.0737.30 ± 8.275.047(0.024)0.003Breastfeeding knowledge12.31 ± 3.2914.59 ± 3.0814.94 ± 3.1014.27 ± 3.5937.897(0.156)0.000Postpartum depression scores7.80 ± 3.967.78 ± 4.057.37 ± 3.437.53 ± 3.780.957(0.005)0.406Note:** T0 = postpartum hospitalization stage, T1 = 42 days after delivery.T2 = 4 postpartum months, T3 = 6 months postpartum.TPB= Theory of Planned behavior.
We established a total of four alternative growth hybrid models with different potential categories. We finally concluded that the growth mixture model (GMM) with three potential categories had the best fitting index and the best effect. First, when the number of potential categories in the GMM model increased from 1 to 3, the AIC (Akaike)value, BIC(Bayesian)value, and aBIC (adjust Bayesian) value all decreased gradually, however, when the number of potential categories increased from 3 to 4, the AIC value, BIC value, and aBIC value all increased. The GMM model with three potential categories had the smallest AIC, BIC, and aBIC values. Second, the model fit well when entropy was ≥0.80, and the GMM model with three potential classes had the highest entropy value of 0.975. Third, if the Bootstrap Likelihood Ratio Test or Vuong-Lo-Mendell-Rubin likelihood ratio test showed p < 0.05, it was considered that the model had a good fitting effect; only GMM models with three potential categories met the criteria. We finally determined that the mixed model of latent variable growth with three potential categories had the bestfitting effect on the trajectory of postpartum breastfeeding behavior in participants; detailed results are shown in Table 3.Table 3GMM model for four alternative potential categories of breastfeeding behavior trajectories in 206 women.Table 3CategoryLLAICBICaBICEntropyBLRTVLMRCategory probability (%)1−2160.8354339.6704369.6214341.105————2−2129.8184283.6364323.5704285.5500.8920.00000.005740.8/50.23−2086.7184203.4364253.3544205.8280.9750.00000.000037.5/26.0/36.54−2084.7974205.5954265.4974208.4660.8660.33330.606636.6/21.8/25.4/16.2Note: GMM: Mixed Growth Model. LL: Loglikelihood. AIC: Akaike. BIC: Bayesian. aBIC:adjust Bayesian. BLRT: Bootstrapped. VLMR: Vuong-Lo-Mendel1-Rubin
Three change curves could be seen on the fitting trajectory chart of maternal breastfeeding behavior development. In other words, there were three potential categories of maternal breastfeeding behavior trajectories, as shown in Fig. 1, supporting H1. According to changes in the initial scores of maternal breastfeeding behaviors and scores at different time points, the three potential categories of breastfeeding behavior trajectories were named respectively as follows. The first curve (category 1, Class 1): breastfeeding behavior became increasingly subpar as breastfeeding behavior scores declined rapidly to low levels over time, with a lower mean value of each dimension score; therefore, potential category 1 was named the “low-level declining group.” The second curve (category 2, Class 2): with increased time, the score of breastfeeding behavior was at a moderate level and decreased slowly; breastfeeding behavior worsened over time; therefore, potential category 2 was named the “medium-level declining group.” The third curve (category 3, Class 3): breastfeeding behavior was at a high level and gradually increased; therefore, category 3 was named the "high-level rising group.”Fig. 1Three potential categories of maternal breastfeeding behavior trajectories.Fig. 1Note: T0 means postpartum hospitalization stage, T1 means 42 days postpartum, T2 means 4 months postpartum, T3 means 6 months postpartum
The developmental trajectory of postpartum breastfeeding behavior in participants was divided into three potential categories. Based on these categories, single factor analysis was conducted according to the general demographic characteristics of participants, variables related to the Theory of Planned Behavior, breastfeeding knowledge, and postpartum depression scores, to explore factors influencing the developmental trajectory of postpartum breastfeeding behavior.
Using the chi-squared test or F-test, univariate analysis of general demographic factors showed that monthly household per capita income, spouse's educational background, and smoking in the spouse had an impact on the development trajectory of breastfeeding behavior among participants. There were significant differences in the distribution of monthly household income per capita, spouse's education, spouse's smoking, and feeding patterns before discharge among the three groups.
Based on the Theory of Planned Behavior, we explored factors influencing the development trajectory of breastfeeding behavior in participants. Therefore, relevant variables of the Theory of Planned Behavior were combined with maternal breastfeeding knowledge and postpartum depression levels to construct the influencing factor system of the theory for breastfeeding behavior in participants. Variables related to the Theory of Planned Behavior (breastfeeding intention, breastfeeding positive attitude, breastfeeding negative attitude, social and professional support, perceived breastfeeding control), breastfeeding knowledge, and level of postpartum depression measured at different periods were analyzed using one-way ANOVA. In the T0 period, we found that no variables were statistically significant across the three categories of maternal breastfeeding behavior trajectories. In the T1 period, there were significant differences in the breastfeeding intention, breastfeeding attitude, social professional support, perceived breastfeeding control, and postpartum depression score. In the T2 period, there were significant differences in breastfeeding intention, breastfeeding positive attitude, breastfeeding negative attitude, social professional support, perceived breastfeeding control, and breastfeeding knowledge. In the T3 period, there were significant differences in breastfeeding positive attitude, breastfeeding negative attitude, social professional support, perceived breastfeeding control, breastfeeding knowledge, and postpartum depression score (Table 4), supporting H2.Table 4Single factor analysis of TPB correlation variables, BKQ and EPDS scales on potential categories of breastfeeding behavior trajectories.Table 4VariablesThe low-level decline group (N = 78)Medium level decline group (N = 53)High level rise group (N = 75)F(pη²)pMultiple comparisonsT0Breastfeeding intention6.31 ± 1.336.72 ± 1.466.75 ± 1.382.286(0.022)0.104Breastfeeding positive attitude249.50 ± 54.74247.34 ± 44.80253.96 ± 46.090.310(0.003)0.734Breastfeeding negative attitude198.18 ± 65.59194.08 ± 49.96198.21 ± 58.010.096(0.001)0.909Social and professional support166.42 ± 59.01158.00 ± 63.95164.64 ± 62.150.313(0.003)0.732Breastfeeding control37.73 ± 6.0738.30 ± 7.6538.99 ± 6.760.662(0.006)0.517Breastfeeding knowledge12.55 ± 3.1912.36 ± 3.4312.01 ± 3.310.519(0.005)0.596Postpartum depression score8.05 ± 3.978.13 ± 4.497.31 ± 3.530.925(0.009)0.398T1Breastfeeding intention5.87 ± 1.576.44 ± 1.337.25 ± 1.1619.251(0.159)<0.051 < 2<3^a^Breastfeeding positive attitude230.81 ± 68.22250.66 ± 57.41273.23 ± 47.909.988(0.090)<0.051 < 3^a^Breastfeeding negative attitude190.10 ± 67.64196.77 ± 50.30184.91 ± 49.010.671(0.007)0.512Social and professional support145.23 ± 48.54154.40 ± 50.30170.44 ± 50.614.989(0.047)<0.051 < 3^a^Breastfeeding control36.19 ± 6.5739.47 ± 6.3142.15 ± 4.5719.899(0.164)<0.051 < 2<3^a^Breastfeeding knowledge14.53 ± 2.8214.32 ± 2.9514.85 ± 3.440.490(0.005)0.613Postpartum depression score8.46 ± 3.938.38 ± 4.856.64 ± 3.304.818(0.045)<0.051&2>3^a^T2Breastfeeding intention5.08 ± 1.875.92 ± 1.437.21 ± 1.3534.624(0.254)<0.051 < 2<3^a^Breastfeeding positive attitude214.92 ± 66.74248.70 ± 57.01286.60 ± 55.2327.037(0.210)<0.051 < 2<3^a^Breastfeeding negative attitude226.56 ± 90.53188.91 ± 54.23169.92 ± 44.2813.842(0.120)<0.051>2&3^a^Social professional support127.03 ± 50.02150.81 ± 47.70173.23 ± 46.2017.680(0.148)<0.051 < 2<3^a^Breastfeeding control34.82 ± 6.7239.32 ± 6.1844.04 ± 4.5647.085(0.317)<0.051 < 2<3^a^Breastfeeding knowledge13.83 ± 3.5215.68 ± 2.1015.57 ± 2.938.617(0.078)<0.051 < 2&3^a^Postpartum depression score7.78 ± 3.317.68 ± 4.266.72 ± 2.782.152(0.021)0.119T3Breastfeeding negative attitude177.64 ± 34.27219.23 ± 36.57278.91 ± 62.4289.358(0.468)<0.051 < 2<3^a^Social professional support301.14 ± 93.42191.00 ± 37.18175.47 ± 52.2274.768(0.424)<0.051>2&3^a^Breastfeeding control109.69 ± 31.20131.32 ± 38.03170.33 ± 54.5139.275(0.279)<0.051 < 2<3^a^Breastfeeding knowledge31.12 ± 5.9036.87 ± 6.5844.03 ± 6.0484.875(0.455)<0.051 < 2<3^a^Postpartum depression score13.55 ± 3.8513.81 ± 3.7915.35 ± 2.885.608(0.052)<0.051&2 < 3^a^Breastfeeding negative attitude8.36 ± 2.878.19 ± 4.806.20 ± 3.457.814(0.071)<0.051&2>3^a^Note: a = Bonferroni test, b = Tamhani test,In the multiple comparisons, 1 = the low-level decline group.2 = the medium-level decline group, 3 = the high-level rise group.T0 = postpartum hospitalization stage, T1 = 42 days after delivery, T2 = 4 postpartum months, T3 = 6 months postpartum.TPB = Theory of Planned Behavior, BKQ = Breastfeeding Knowledge Questionnaire.EPDS = Edinburgh Postnatal Depression Scale.
We established unordered multi-class logistic regression models at different postpartum periods, with the dependent variable being the three potential categories of breastfeeding development trajectories and the independent variables being statistically significant variables identified in the single-factor analysis. Taking the low-level declining group as the reference category, we explored the factors influencing the development of breastfeeding behavior among participants in the high-level rising group and medium-level declining group and established a multi-classification logistic regression model with three different categories of maternal breastfeeding trajectory as dependent variables. Table 5 shows that household income, spouse's education, spouse's smoking, breastfeeding patterns before discharge, variables related to the Theory of Planned Behavior, breastfeeding knowledge, and postpartum depression scores all influenced the trajectory of breastfeeding behavior in participants.Table 5Multivariate logistic regression analysis of predictors of maternal breastfeeding behavior development trajectory at stage T0-T3Table 5VariablesRegression coefficientStandard errorWald χ^2^pOR (95%CI)*Medium level decline group (T0)****Household income (Yuan)≤5000-0.4651.0830.1840.6680.628 (0.075, 5.250)5001-100000.4940.5130.9270.3361.639 (0.600, 4.479)10001-15000-1.4490.5856.1370.0130.235(0.075,0.739)>15000元Spouse educationHigh school-0.3870.7280.2820.5960.679 (0.163, 2.833)Undergraduate-1.4100.5147.5210.006*0.244 (0.089, 0.669)Postgraduate and aboveSpouse smokeYes-1.1920.5035.6200.0180.304 (0.113, 0.813)NoBreastfeeding patterns before dischargeExclusive breastfeeding0.1980.6840.0840.7721.219 (0.319, 4.658)Mixed breastfeeding0.5120.4521.2840.2571.669 (0.688, 4.046)Formular breastfeeding High level rise group (T0)****Household income (Yuan)≤5000-0.5510.9220.3570.5500.576 (0.095, 3.509)5001-10000-0.2970.4510.4330.5110.743 (0.307, 1.799)10001-15000-1.0950.4485.983**0.0140.334 (0.139, 0.804)>15000Spouse educationHigh school-0.1980.6930.0810.7760.821 (0.211, 3.195)Undergraduate-0.0370.4850.0060.9390.964 (0.373, 2.492)Postgraduate and aboveSpouse smokeYes-0.5100.3911.7010.1920.600 (0.279, 1.293)NoBreastfeeding patterns before dischargeExclusive breastfeeding1.5530.5557.8220.0054.725 (1.591, 14.030)Mixed breastfeeding0.7150.4063.1080.0782.045 (0.923, 4.530)Formular breastfeeding *Medium level decline group (T1)****Breastfeeding intention0.1930.1881.0540.3051.213 (0.839, 1.753)Breastfeeding positive attitude0.0000.0040.0070.9341.000(0.992,1.009)Social and professional support0.0020.0050.1210.7281.002(0.993, 1.011)Perceived breastfeeding control0.1000.0464.6190.0321.105 (1.009, 1.210)****Postpartum depression score0.1120.0573.8470.0501.118 (1.000, 1.250)Household income (Yuan)≤5000-1.1991.1591.0700.3010.301(0.031, 2.925)5001-100000.2450.5490.1980.6561.277 (0.435, 3.748)10001-15000-1.6200.6256.7130.010****0.198 (0.058, 0.674)>15000Spouse educationHigh school-0.6990.7790.8050.3700.497 (0.108, 2.290)Undergraduate-1.4860.5457.422**0.0060.226 (0.078, 0.659)Postgraduate and aboveSpouse smokeYes-1.3190.5356.074**0.0140.267 (0.094, 0.763)NoBreastfeeding patterns before dischargeExclusive breastfeeding0.1800.7430.0580.8091.197 (0.279, 5.139)Mixed breastfeeding0.5480.4911.2480.2641.730 (0.661, 4.526)Formular breastfeeding High level rise group (T1)****Breastfeeding intention0.5450.1947.870**0.0051.724 (1.178, 2.523)****Breastfeeding positive attitude0.0010.0040.0910.7631.001 (0.993, 1.009)Social and professional support0.0030.0050.5780.4471.003 (0.995, 1.012)Perceived breastfeeding control0.1410.0488.793**0.0031.152 (1.049, 1.265)****Postpartum depression score0.0070.0610.0130.9081.007 (0.894, 1.134)Household income (Yuan)≤5000-2.0861.0543.916**0.0480.124 (0.016, 0.980)5001-10000-0.9530.5383.1350.0770.386 (0.134, 1.107)10001-15000-1.5560.5398.3300.0040.211 (0.073, 0.607)>15000Spouse educationHigh school-0.3430.7750.1960.6580.710 (0.155, 3.239)Undergraduate-0.0060.5570.0000.9910.994 (0.333, 2.962)Postgraduate and aboveSpouse smokeYes-0.7540.4542.7570.0970.470 (0.193, 1.146)NoBreastfeeding patterns before dischargeExclusive breastfeeding1.2780.6473.895**0.0483.589 (1.009, 12.765)Mixed breastfeeding0.6650.4841.8930.1691.945 (0.754, 5.020)Formular breastfeeding *Medium level decline group (T2)****Breastfeeding intention0.1000.1540.4230.5161.105 (0.818, 1.494)Breastfeeding positive attitude-0.0060.0051.3600.2440.994(0.983, 1.004)Breastfeeding negative attitude-0.0020.0030.2750.6000.998 (0.992, 1.005)Social and professional support0.0030.0060.2860.5931.003 (0.992, 1.014)Perceived breastfeeding control0.0800.0482.7900.0951.083 (0.986, 1.189)Breastfeeding knowledge0.2370.1025.4450.0201.268 (1.039, 1.548)****Household income (Yuan)≤5000-1.2981.1661.2400.2650.273 (0.028, 2.682)5001-100000.2680.5510.2360.6271.307 (0.444, 3.848)10001-15000-1.2610.6284.030**0.0450.283 (0.083, 0.971)>15000Spouse educationHigh school-0.6250.7770.6460.4210.535 (0.117, 2.457)Undergraduate-1.4500.5556.8200.0090.235 (0.079, 0.696)Postgraduate and aboveSpouse smokeYes-1.0390.5393.7100.0540.354(0.123, 1.018)NoBreastfeeding patterns before dischargeExclusive breastfeeding0.4990.7720.4180.5181.648 (0.363, 7.488)Mixed breastfeeding0.3260.4860.4490.5031.385 (0.534, 3.591)Formular breastfeeding *High level rise group (T2)****Breastfeeding intention0.4450.1716.7720.0091.560 (1.116, 2.182)Breastfeeding positive attitude0.0020.0050.1770.6741.002(0.992, 1.013)Breastfeeding negative attitude-0.0070.0052.3820.1230.993 (0.984,1.002)Social and professional support0.0080.0062.3340.1271.008(0.998, 1.019)Perceived breastfeeding control0.1720.05111.2530.0011.187 (1.074, 1.313)Breastfeeding knowledge-0.0670.0880.5790.4470.936(0.788, 1.111)Household income (Yuan)≤5000-1.8871.1142.8710.0900.151 (0.017, 1.344)5001-10000-0.3960.5720.4790.4890.673 (0.219, 2.065)10001-15000-0.5540.6090.8300.3620.574 (0.174, 1.893)>15000Spouse educationHigh school-0.6130.8480.5220.4700.542 (0.103, 2.855)Undergraduate-0.0730.5980.0150.9030.930 (0.288, 3.005)Postgraduate and aboveSpouse smokeYes-0.4580.5190.7790.3770.633(0.229, 1.748)NoExclusive breastfeeding1.6020.7584.465**0.0354.961 (1.123, 21.911)****High level rise group (T2)****Breastfeeding patterns before dischargeMixed breastfeeding0.5200.5460.9090.3401.683 (0.577, 4.905)Formular breastfeeding Medium level decline group (T3)****Breastfeeding positive attitude0.0140.0083.1910.0741.014 (0.999, 1.029)Breastfeeding negative attitude-0.0160.00417.947**0.0000.984(0.977, 0.991)Social and professional support-0.0000.0070.0010.9771.000(0.987, 1.013)Perceived breastfeeding control0.0540.0491.2440.2651.056 (0.960, 1.161)Breastfeeding knowledge0.0280.0650.1900.6631.029 (0.905, 1.169)Postpartum depression score0.0780.0671.3610.2431.081 (0.948, 1.232)****Household income (Yuan)≤5000-1.1011.4460.5790.4470.333 (0.020, 5.664**)5001-100000.7840.6901.2910.2562.189 (0.567, 8.460)10001-15000-1.3640.7193.5980.0580.256 (0.062, 1.046)>15000Spouse educationHigh school-0.2740.9560.0820.7740.760 (0.117, 4.953)Undergraduate-1.3630.6794.0310.0450.256 (0.068, 0.968)Postgraduate and aboveSpouse smokeYes-0.8590.6521.7390.1870.424(0.118, 1.519)NoBreastfeeding patterns before dischargeExclusive breastfeeding-0.1870.9410.0390.8430.829 (0.131, 5.251)Mixed breastfeeding0.5970.6050.9760.3231.817 (0.555, 5.946)Formular breastfeeding High level rise group (T3)****Breastfeeding positive attitude0.0270.0099.574**0.0021.028 (1.010, 1.045)Breastfeeding negative attitude-0.0210.00614.1220.0000.979(0.969, 0.990)Social and professional support0.0010.0080.0130.9101.001 (0.986, 1.016)Perceived breastfeeding control0.1870.0638.6640.0031.205(1.064, 1.365)Breastfeeding knowledge-0.0050.0810.0040.9470.995 (0.848, 1.166)0.0070.0750.0080.9291.007(0.869, 1.166)Household income (Yuan)≤5000-3.1111.5783.885**0.0490.045 (0.002, 0.983)5001-10000-0.3170.7790.1650.6850.729 (0.158, 3.357)10001-15000-1.4470.7913.3480.0670.235 (0.050, 1.108)>15000High level rise group (T3)****Spouse educationHigh school0.1591.1080.0210.8861.172(0.134,10.282)Undergraduate0.1730.7910.0480.8271.189 (0.252, 5.603)Postgraduate and aboveSpouse smokeYes-0.6360.7150.7920.3740.529(0.130, 2.148)NoBreastfeeding patterns before dischargeExclusive breastfeeding0.7401.0060.5410.4622.096(0.292, 15.053)Mixed breastfeeding0.5890.7110.6870.4071.803 (0.447, 7.262)Formular breastfeeding *Note: The low-level decline group was taken as the reference category. * Control groupT0=hospital stay before discharge, T1= 42 days after delivery, T2=4 postpartum months, T3=6 months postpartum.
This work followed a prospective longitudinal study design using the Breastfeeding Behavior Questionnaire as the core index to measure the level of breastfeeding behavior in participants. A mixed model of latent variable growth was used to fit the developmental trajectories of breastfeeding behavior in participants and to identify the developmental trajectories of breastfeeding behaviors in this population.
We found that the rates for exclusive breastfeeding of preterm infants during hospitalization and at 6 weeks, 4 months, and 6 months after delivery were 16 %, 33.5 %, 43.2 %, and 37.4 %, respectively; these rates were far from the breastfeeding targets proposed in China and abroad, indicating that the current situation of preterm infant breastfeeding is suboptimal. In our study, the rate of exclusive breastfeeding at 6 weeks postpartum was slightly lower than that in another survey (Keir et al., 2022). The reason for this difference may be that mothers of preterm infants investigated by Keir et al. mainly had late preterm infants, in whom the ability to swallow is stronger. The rate of exclusive breastfeeding at 6 months postpartum was 37.4 % in that study. Most women in our study were older first-time mothers with higher education levels, who might pay greater attention to breastfeeding their premature infant and have good breastfeeding compliance.
Maternal breastfeeding behavior after delivery was at moderately low levels, and participants had poor breastfeeding status after discharge, which is consistent with another study (Méio et al., 2018). Postpartum breastfeeding behavior scores were highest at 6 weeks postpartum. The reasons for this may be that during hospitalization, most preterm infants were admitted to the neonatal intensive care unit and separated from their mothers. When the biological activity of prolactin is not high, lactation initiation in mothers is relatively delayed (Ostrom, 1990). With a short hospitalization time and the mother's ability to secrete breast milk, the needs of newborns in the hospitalization stage differ, so the rate of exclusive breastfeeding is low. From the time of hospitalization to 6 weeks postpartum, feeding behavior scores increased, which meant that participants were getting better at breastfeeding.
The first category was the low-level declining group, which accounted for 37.5 % of all participants. The initial level of breastfeeding behavior in this group was the lowest among the three categories. With increased postpartum time, breastfeeding remained at a low level and showed a rapidly declining trend, which meant that 37.5 % of participants had breastfeeding behavior at a stably low level and exhibited increasingly poor behavior in the first 6 months postpartum. The second category was the medium-level declining group, accounting for 26.0 % of all participants. The initial level of breastfeeding behavior in this group was the highest among the three categories; however, during postpartum follow-up, participants maintained a stably medium level and showed a slow declining trend. The third group was the high-level rising group, accounting for 36.5 % of all participants. The initial level of breastfeeding behavior in this group was second only to the medium-level declining group in the second category. However, the level of breastfeeding behavior increased significantly during follow-up and remained at a high level. This meant that the level of breastfeeding behavior in this group of women increased rapidly over time, maternal breastfeeding behavior became increasingly better, and there was less risk of stopping breastfeeding.
There are few reports of similar investigations among mothers of preterm infants, so it is difficult to compare and discuss our findings with those of other studies. We found results similar to those of a study on the development trajectory of infant breastfeeding intensity in the United States (Whipps et al., 2019). Those researchers analyzed the follow-up data of mothers' breastfeeding practices and termination of breastfeeding in the first year after delivery, established a group trajectory model, and identified four potential categories of breastfeeding intensity in the first year the breastfeeding group, late weaning group, early weaning group, and formula feeding group. The fourth category group in Whipps et al. (2019) was the breastfeeding group (27.2 %), which was characterized by high initial breastfeeding intensity, a sustained high intensity up to 6 months after delivery, and a decrease after 10 months postpartum. The trajectory of breastfeeding intensity in the 6 months after delivery was similar to that of the third category in our study, the high-level rising group. We found that medical staff should pay greater attention to mothers of preterm infants in the low- and medium-level declining groups.
Monthly household income influenced the breastfeeding behavior of participants in the first 6 months postpartum. During hospitalization and at 6 weeks, 4 months, and 6 months after delivery, participants with monthly household per capita income of 10,001–15,000 Yuan (a mid-to-high income) were more likely to be in the low-level declining group. This was consistent with Mortazavi et al. (2015),who reported that mothers with higher monthly per capita income were more likely to have higher maternal income. The introduction of formula during the breastfeeding process influenced breastfeeding behavior in these mothers and led to the cessation of exclusive breastfeeding. Among participants from middle- and high-income families, the income level of the household had a certain impact on their breastfeeding decisions. Therefore, to promote breastfeeding, social institutions should strive to increase the publicity about the benefits of breastfeeding and avoid excessive promotion of powdered milk for those high-income families. Women from low-income families may have less education and limited access to information; this may diminish their emphasis on breastfeeding and lower their level of breastfeeding behavior. Healthcare staff should reinforce education on breastfeeding and offer timely professional guidance and assistance when low-income families encounter difficulties in breastfeeding, aiming to enhance the breastfeeding rate among premature infants.
During postpartum hospitalization and at 6 weeks and 4 months after delivery, women who chose exclusive breastfeeding before discharge were more likely to be in the high-level rising group. Thus, exclusive breastfeeding before discharge was a protective factor, which was consistent with past survey results (Henderson et al., 2022). Lactation initiation was completed among participants who achieved exclusive breastfeeding during hospitalization, and milk production gradually increased. Neonatal intensive care units should aim for family-integrated care to establish appropriate breastfeeding behavior among mothers as early as possible.
Having a spouse with a bachelor's or other college degree or a spouse who smoked were risk factors for postpartum breastfeeding behavior among mothers of preterm infants; those women were more likely to be in the low-level declining group, which was consistent with previous study results (Blazé et al., 2024). It is speculated that smoking in the spouse may cause maternal psychological pressure and anxiety, which may interfere with breast milk production and breastfeeding behavior. Therefore, healthcare staff should prioritize women whose spouses hold a bachelor's or smoke, provide targeted breastfeeding health education to their family, encourage them to provide emotional and physical breastfeeding support, aiming to promote breastfeeding for these infants. At 6 months postpartum, participants with highly-positive attitudes about breastfeeding were more likely to be in the high-level rising group, with better breastfeeding behavior. Breastfeeding women with high negative attitude scores were more likely to be in the low-level declining group and have poor breastfeeding behavior. This was consistent with past findings (Niela-Vilén et al., 2016) that breastfeeding positive attitude is the strongest predictor of a longer duration of breastfeeding. Mothers with breastfeeding negative attitudes may lack knowledge and awareness about breastfeeding, as well as support from family members and healthcare staff, and they may find it difficult and feel helpless when experiencing obstacles to breastfeeding. Healthcare staff should promptly screen mothers' attitudes toward breastfeeding after delivery, identify those with negative attitudes, and provide education and technical support to help mothers form positive attitudes toward breastfeeding to promote breastfeeding in preterm infants. Healthcare providers in urban settings can implement workplace breastfeeding support programs, such as flexible working hours and designated lactation spaces. In rural communities, peer-led breastfeeding education workshops conducted by local health volunteers can enhance awareness and reduce cultural barriers.
Compared with the low-level declining group, at 6 weeks, 4 months, and 6 months postpartum, women with high scores of perceived breastfeeding control were particularly likely to be in the high-level rising group and to have good breastfeeding behavior. Perceived breastfeeding control reflects the control and self-efficacy of mothers regarding breastfeeding behavior. This was consistent with published findings(Gerhardsson et al., 2018; Kronborg et al., 2024) from studies investigating mothers of preterm infants. Perceived breastfeeding control is a predictor of early postpartum breastfeeding behavior and the duration of breastfeeding for mothers of preterm infants. Participants with high scores for breastfeeding intention were more likely to be in the high-level rising group, with better breastfeeding behavior at 6 weeks and 4 months after delivery. This was similar to past reports (Keir et al., 2022; Wang and Chang, 2023) that the breastfeeding intention of mothers of preterm infants was an important predictor of the behavior and duration of breastfeeding during the first 6 months after delivery. It may be that breastfeeding intention is a factor directly affecting breastfeeding behavior, with mothers being more likely to adhere to the goal of exclusive breastfeeding. Therefore, healthcare staff should conduct regular follow-up contact to evaluate their breastfeeding intention, provide timely social and professional support and health education to mothers with lower breastfeeding intention, and thus increase their motivation to adhere to breastfeeding behavior.
At 4 months postpartum, women with low breastfeeding knowledge scores were more likely to be in the low-level declining group than the medium-level declining group. Consistent with other researchers (Zhang et al., 2018), low level of breastfeeding knowledge is a risk factor for poor breastfeeding behavior of premature infants after discharge, with mothers who lack knowledge of breastfeeding being more likely to choose mixed feeding or formula feeding after discharge. Strengthening breastfeeding knowledge and skills among mothers of preterm infants can improve their breastfeeding intention and perceived breastfeeding control to promote breastfeeding positive attitudes. In addition, publicity materials and videos should be tailored according to the individual conditions of mothers of preterm infants, in order to meet the practical needs.
Univariate analysis showed that social and professional support at T1, T2, and T3 was significantly different among potential categories of the maternal breastfeeding behavior development trajectory (p < 0.05). This was in line with a study by Abrmanová et al. (2023), in which they found that the degree of social and professional support for mothers of preterm infants had an impact on breastfeeding behavior for premature infants. We found statistically significant differences between the potential categories of postpartum depression scores in the development trajectory of maternal breastfeeding behavior at T1 and T3 (p < 0.05), which was in line with past research (Gila-Díaz et al., 2020).Researchers have shown that women who choose formula or mixed feeding are more likely to experience depression than those who are exclusively breastfeeding. Postpartum depression would reduce the rate of exclusive breastfeeding and affect maternal breastfeeding behavior. According to Rollins et al. (2016), the promotion of breastfeeding practice is achieved through the interaction of individual (the mother and her infant), environmental (healthcare institution, community, family, and workplace) and structural (sociocultural and market–environmental) factors. Focusing on maternal emotions and early identification of mothers of preterm infants with depressive tendencies may be an effective intervention measure to reduce the incidence of postpartum depression.
We used self-reports that included questions regarding breastfeeding behavior; social desirability bias may have led to over-reporting, as participants were aware of the optimal behavior as reported by study investigators(Huang et al., 2023). We recommend instead observational studies of breastfeeding behavior. Sample representativeness and generalizability of the findings could be biased, as the study was conducted in one hospital in Shanghai; multi-center studies in the northwest region should be increased in the future.
Breastfeeding is of great importance to promote the growth and development of preterm infants. However, the breastfeeding situation among mothers of preterm infants within 6 months after delivery is not encouraging. The rate of exclusive breastfeeding among study participants increased from the period of hospitalization to 4 months after delivery and then declined rapidly up to 6 months after delivery. Overall, the exclusive breastfeeding rate was low and must be improved. Improving understanding of variables related to the Theory of Planned Behavior, such as breastfeeding positive and negative attitudes, perceived breastfeeding control, breastfeeding knowledge, breastfeeding intention, and postpartum depression can provide a theoretical basis for the early identification and screening of new mothers at risk of exclusive breastfeeding attrition. Interventions such as counseling programs aimed at boosting maternal perceived breastfeeding control should be delivered at scale to benefit all mothers and their preterm infants.
Na Wu: Writing – review & editing, Writing – original draft, Investigation, Data curation, Conceptualization. Yijing Wang: Investigation, Formal analysis, Data curation, Conceptualization. Lijing Ding: Writing – original draft, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Yiting Xu: Software, Investigation, Data curation. Xuyan Jia: Investigation, Formal analysis, Data curation. Keying Huo: Investigation, Formal analysis, Data curation. Xiao Yao: Investigation, Formal analysis, Data curation. Wenyan Zhang: Investigation, Data curation. Yue Wan: Software, Formal analysis, Data curation. Peimin Xie: Writing – review & editing. Lin Ye: Software, Formal analysis, Data curation. Rong Huang: Writing – review & editing, Supervision, Methodology, Funding acquisition, Data curation, Conceptualization.
The article is our original work and has not received prior publication and is not under consideration for publication elsewhere. All authors of this paper have read this version of the article and approve the submission and there are no conflicts of interest with respect to this publication.