Abstract
We explored parental feeding practices, feeding issues during the first 2 yr of life, and the relationship between feeding issues and levels of maternal distress in preterm infants. Four hundred twenty mothers (239 with preterm infants, 181 with full-term infants) participated in the study. The Behavior-Based Feeding Questionnaire for Infants With Premature History and the Parenting Stress Index–Chinese Version were used as the two outcome measures. The results indicated that preterm infants had different feeding experiences compared with their full-term counterparts. They tended to start solid food later in life and had limited experiences in food variation. Parenting stress was prevalent in parents with preterm infants and was associated with the frequency of feeding issues, especially at later ages when supplementary foods were being introduced.
Feeding problems are one of the most common developmental issues observed in preterm infants after their discharge from the neonatal intensive care unit (NICU). Researchers have reported a high cooccurrence of feeding problems in preterm children with low birthweight (Gouyon et al., 2012; Laptook, 2013; Leone et al., 2012). An estimated 20%–45% of preterm infants experience feeding problems in the first 2 yr of life (Crapnell et al., 2013; Laptook, 2013; Rommel et al., 2003; Thoyre, 2007), and these feeding-related issues or difficulties may persist into the preschool age (Howe et al., 2010; Ross & Browne, 2013).
Despite a high prevalence of feeding problems in infants with prematurity, relatively little research in the extant literature has examined preterm infants’ feeding behaviors after their hospital discharge (DeLucia & Pitts, 2006; Migraine et al., 2013). First, little is known about parental feeding practices in the preterm population after infants are discharged from their initial NICU stay. Clinicians need to understand how parents with preterm infants progress with infant feeding to provide relevant recommendations. However, in most studies, researchers have focused on feeding behavior in very young infants, who feed exclusively on milk (both formula and breast milk), or in children age 2 yr and older, who mainly feed on solids (van Dijk et al., 2012). Few studies have reported on feeding practices during the transition from milk to solid food or the introduction of supplementary foods (Spiegler et al., 2015). Chung et al. (2014) reported that parents with premature infants tended to introduce solid food earlier than recommended because they were dissatisfied with their child’s growth. Premature infants who are not developmentally ready at the time they are introduced to solid food may develop avoidant feeding behaviors, which, in turn, may induce unnecessary parental distress related to infant feeding patterns (Chung et al., 2014).
Second, clinicians currently have little evidence to determine or predict whether feeding issues of preterm infants observed during early infancy improve over time and whether these feeding issues will become problematic over the course of the child’s life. Studies on preterm infants’ feeding problems after hospital discharge have been mostly generated from a tightly defined preterm infant group treated in special feeding clinics (DeMauro et al., 2011; Lutz, 2012). Observations from this subpopulation may present a skewed picture of preterm infants’ feeding behaviors. Instead of describing feeding behaviors that range from transient, relatively minor mealtime problems, the extreme end of spectrum may be overemphasized. Therefore, more research in nonclinical samples during early childhood is essential.
Last, little is known about the relationship between parents’ perception of their child’s feeding problems and parenting stress through the first 2 yr of parenthood (Lutz, 2012). Previous studies have reported that infant feeding problems might be associated with parents’ anxiety and maladjustment in infants with different medical conditions (Powers et al., 2002; Tallandini et al., 2015). These findings led us to hypothesize that the same phenomenon would be observed in the preterm population. We anticipated that the families of preterm infants with feeding difficulties would be characterized by higher levels of maternal psychological distress and that this parenting stress may continue during early infancy (Howe et al., 2014).
The purpose of this study was threefold. First, we explored the parental feeding practices in preterm infants to understand what, when, and how mothers introduced foods to their infants; specifically, we examined the timing of meals and types of food that mothers chose for their children. Second, we examined feeding issues of preterm infants during the first 2 yr of their live; specifically, we identified infants’ feeding issues perceived by their mothers at different stages of development and compared these feeding issues with those from a group of full-term infants. Third, we examined the relationship between mothers’ reported feeding issues and the levels of maternal distress in both preterm and full-term groups. On the basis of previous studies (Chung et al., 2014; Spiegler et al., 2015), we hypothesized that our sampled mothers of preterm infants would introduce supplementary food to their children earlier than recommended time in comparison with mothers of full-term infants. Finally, we hypothesized that mothers who perceived more feeding problems in their children would experience higher levels of maternal distress. Moreover, mothers with preterm infants would express higher distress more often than mothers with full-term infants.
Method
Design
This was an exploratory study with a nonexperimental, cross-sectional design. Between 2012 and 2013, parents attending the developmental follow-up for their preterm infants or in the well-baby clinic at a large urban hospital in the southern region of Taiwan were approached to join the study during their visits. With their consent, they were asked in person to complete two questionnaires: the Behavior-Based Feeding Questionnaire for Infants With Premature History (BFQ; Howe & Ho, 2009) and the Parenting Stress Index–Chinese Version (PSI–C; Weng, 2003). The hospital’s ethics committee granted permission for the study.
Participants
For the preterm infant–parent pairs, we recruited parents with children who were less than age 2 yr, with gestational age less than 37 wk and birthweight ≤1,500 g. The full-term infant–parent pairs were recruited from outpatient well-baby clinics for comparison with their preterm counterparts. The inclusion criteria were comparable with the studied group, except for the infants’ gestational age, which was ≥37 wk. We excluded 85 infant–parent pairs from the original sample for this study because the mothers did not complete the feeding survey questionnaires. A total of 420 infant–mother pairs, 239 pairs in the preterm group and 181 pairs in the full-term group were included in the final analysis.
Outcome Measures
Behavior-Based Feeding Questionnaire for Infants With Premature History.
The BFQ is a nonstandardized, criteria-based assessment that was used to assess infant’s feeding behaviors from the perspective of the primary caregiver. In this questionnaire, participating mothers were asked to report their observations of their child’s feeding behavior by answering 33 questions regarding feeding-related issues grouped into 6 categories: endurance, gastrointestinal-related issues, muscle tone, oral motor function, respiration, and sensory regulation. In addition, participating mothers were also asked in the questionnaire to describe their feeding practices, including frequency and duration of feeding and the types of food that they introduced to their infants. The face and content validity of the questionnaire was established via expert opinions (Howe & Ho, 2009). The BFQ has been used with Taiwanese population to examine relationships between perceived feeding issues and parental stress (Howe et al., 2014).
Parenting Stress Index–Chinese Version.
The PSI–C was used to assess parental perceptions of their degree of stress in relation to different dimensions of parenting roles. This 94-item, 5-point Likert scale questionnaire was standardized to be used with Taiwanese parents of children less than age 12 yr (Weng, 2003). Normative data were collected from 1,362 mothers of children in Taiwan. It provides a total stress score and two summary domain scores for parent and child. The parent domain reflects the parents’ perception of their own characteristics that may affect their parenting ability, and the child domain measures the parents’ perception of their child’s characteristics. Adequate psychometric properties, including Cronbach’s α and confirmatory factor analysis, were reported in the test manual (Weng, 2003).
Data Collection and Analysis
The results obtained from the BFQ and PSI–C were used as two outcome measures. Demographic information extracted from medical records of all participating infants was used as independent variables for the study. The demographic information included infants’ gestational age, birthweight, severity of medical complications represented by the Neonatal Medical Index (Korner et al., 1993), mothers’ level of education, and family income.
We performed all statistical analyses using IBM SPSS Statistics (Version 20; IBM Corp., Armonk, NY). We performed mean comparison of feeding questionnaires between preterm and full-term groups using analysis of covariance (ANCOVA) with adjusted age and maternal education year as covariance. We considered p < .05 to be significant. In addition, we explored group differences among infants with or without experience in supplementary foods. We divided both full-term and preterm infants into two subgroups for analyses on the basis of their diet (milk vs. supplementary food) and then compared group differences in parent-perceived feeding issues and parental stress. The milk group represents infants who had milk or formula as their only source of nutrition, and the supplementary foods group represents infants who had other nutrient sources in addition to milk or formula. Infants who were on 100% solid foods were also included in the supplementary foods group.
Results
Characteristics of Infants and Mothers in the Preterm Group
The demographic characteristics of infants and mothers of both preterm and full-term groups both as a whole and in subgroups (milk and supplementary food groups) are presented in Table 1. Preterm and full-term groups are described as a whole in detail elsewhere (Howe et al., 2014). The descriptions of the subgroups are as follows. There were 82 mother–infant pairs included in the milk group. It consisted of 24 preterm infants (16 boys and 8 girls), with a mean adjusted age of 49.0 ± 31.3 wk, and 58 full-term infants (27 boys and 31 girls), with a mean chronological age of 12.1 ± 7.3 wk. There were 322 mother–infant pairs included in the supplementary foods group. The sample of preterm infants in the supplementary foods group consisted of 199 infants (113 boys and 86 girls), with a mean adjusted age of 59.5 ± 31.3 wk, whereas the full-term group consisted of 123 infants (59 boys and 64 girls), with a mean chronological age of 56.4 ± 23.8 wk. The skewness and kurtosis values of the distributions of continuous variables are all less than ±1.96. These values suggest that the distributions are reasonably symmetrical.
Characteristics of Infants in Both Full-Term and Preterm Groups as a Whole and in Subgroups (Milk and Supplementary Food Groups)
Note. Characteristics include infants’ birthweight, gestational age, gender, and severity of medical complications as classified by the Neonatal Medical Index (NMI): Levels I–V, with V indicating the highest severity. The value of NMI is presented only in the preterm group. M = mean; NT = New Taiwan Dollar; SD = standard deviation.
Introduction of Supplementary Foods
The timing of the first introduction of supplementary foods in the preterm infant group was much later than their full-term counterparts. Full-term infants started supplementary feeding as early as 2.83 mo, whereas the earliest preterm infants began at 5.79 mo. At age 6 mo, 100% of full-term infants in our study had supplementary foods added to their diet, whereas only 30.4% of preterm infants had similar experiences. About half of the preterm infants studied had experienced supplementary foods at age 12 mo (53%), and the percentage increased to 69.1% by 18 mo. By 24 mo, most of the studied preterm infants (98.3%) had experience with supplementary foods. There were 10 preterm participants who had not yet experienced any supplementary foods by age 24 mo.
The reported supplementary foods included congee, fruits, vegetables, and food cluster (mixed foods, table foods, and meat). In this study, congee was the most frequently reported supplementary food, followed by fruits. Congee is a type of rice porridge. When it is used to feed young infants in Taiwan, it is often cooked with protein (e.g., chicken, pork, fish, tofu, egg) or vegetables until all ingredients are soft. The full-term group appeared to experience more variations of supplementary food than the preterm group. A majority of mothers from both groups reported feeding their infants homemade foods (i.e., mashed starch-based foods, fruits, and vegetables) over store-bought, processed baby food.
Mothers’ Perceived Feeding Issues
We conducted an analysis to test our hypothesis that mothers of preterm infants may perceive more feeding issues. We used ANCOVA with adjusted age and level of maternal education as covariance to compare the differences between the milk and supplemental food subgroups of preterm and full-term infants in total feeding scores and the six areas of feeding issues identified by the BFQ. Significant differences were found in all areas of perceived feeding issues between preterm and full-term groups who had supplementary food in their diet (supplementary food groups); specifically, in these groups mothers with preterm infants perceived significantly more feeding-related issues in all six subcategories compared with mothers with full-term infants (Table 2). No significant differences were found in the groups of infants in which milk or formula was the only source of nutrient (milk groups; see Table 2).
Results of Parent-Perceived Feeding Issues of Infants in Both the Milk and Supplementary Food Groups
Note. Analysis of covariance with adjusted age and maternal level of education as covariance was used to compare the differences between preterm and full-term groups. M = mean; SD = standard deviation.
Relationship Between Perceived Feeding Issues and Maternal Distress
The relationship between mother-perceived feeding issues and maternal distress was explored in both preterm and full-term groups. The correlations among PSI–C total score, child domain, parent domain, and total feeding score and its subscales were calculated to examine the strength of the linear association among them. The Pearson’s correlation coefficient was computed pairwise, and the resulting correlation matrix is shown in Table 3 for preterm and full-term groups in both the milk and the supplementary food groups.
Correlations Between Parental Distress and Feeding Issues in Both Preterm and Full-Term Groups
Note. PSI–C = Parenting Stress Index–Chinese Version.
p < .05. **p < .01.
Milk Groups.
Significant associations between perceived feeding issues and maternal distress were observed in the mothers with full-term infants. The total score of the parent-perceived feeding issues was positively related to PSI–C total score (r = .355), child domain (r = .366), and parent domain (r = .298). In addition, feeding-related endurance issues had significant positive correlations with PSI–C total score (r = .316), child domain (r = .290), and parent domain (r = .298). No significant relationship was found between parent-perceived feeding issues and maternal distress in the group of preterm infants who had milk or formula as their only nutrient source.
Supplementary Food Groups.
Significant associations between parent-perceived feeding issues and maternal distress were found in infants of both preterm and full-term groups who consumed solid food as part of their diet. In our preterm infant sample who was fed supplementary food, the total number of parent-perceived feeding issues was positively related to PSI–C total score (r = .166) and parent domain (r = .187). In addition, feeding-related oral motor issues and feeding-related regulation issues were found to be the two major sources of parenting stress in this group. Both had significant positive associations with PSI–C total scores and the child and parent domains (see Table 3).
In the full-term supplementary foods group, the associations between perceived feeding issues and parenting stress were similar to their preterm counterpart. The total scores of parent-perceived feeding issues were positively related to PSI–C total score (r = .320), child domain (r = .318), and parent domain (r = .263). Feeding-related muscle tone issues were the major source of stress for parents with full-term infants in this group. They had a significant positive correlation with PSI–C total score (r = .412), child domain (r = .264), and parent domain (r = .453). In addition, both feeding-related gastrointestinal issues and feeding-related regulation issues showed significant correlations with maternal distress, which was reflected in the PSI–C total score and child domain (see Table 3).
Discussion
Most studies to date have reported feeding behaviors either in young infants who feed exclusively on milk or in children 2 yr and older who mainly feed on solid foods. Feeding behavior in the period between and the initial introduction of solid food have largely been neglected (Spiegler et al., 2015; van Dijk et al., 2012). In this study, we explored the timing and progression of the introduction of supplementary foods in preterm infants using a group of full-term infants as a comparison. The results show that preterm infants tended to start their supplementary foods much later in life than their full-term peers, and those who had supplementary foods added to their diet had a much more restricted selection than their full-term counterparts. Although both groups reported starch-based foods, such as cereals and congee, as starter supplementary foods, full-term infants received many more variations of high-protein foods introduced alongside congee at an earlier age. For example, mothers with full-term infants reported feeding their children congee mixed with meat, eggs, and vegetables as early as 2 mo old.
Previous studies reporting on feeding outcomes in preterm infants after discharge claimed that parents often introduce their infants to solids before the 6-mo age range and often earlier than their full-term reference group (Chung et al., 2014; Spiegler et al., 2015). However, our findings differ from studies conducted in Western countries (Morgan et al., 2006; Ross & Browne, 2013; Spiegler et al., 2015). We found that the timing and progression of supplementary foods introduced to preterm infants was much later than the full-term comparison group in our study sample in Taiwan. Our mothers of preterm infants did not introduce supplementary foods to their preterm infants until they were 6 mo old. Even though this timing and progression of supplementary foods adheres more closely to the American Academy of Pediatric recommendations, it does not reflect the common feeding practice of early supplementary food introduction in the preterm population or in different cultural backgrounds (Morgan et al., 2006; Woo et al., 2013).
We suspect that in addition to preterm infants’ reported delayed oral motor skills (Ross & Browne, 2013), the discrepancy might also be the result of the advice that mothers received on infant food progression. Mothers of preterm infants may have diligently followed the advice of their pediatricians or therapists who were mostly trained in Western medicine, whereas mothers of full-term infants may have leaned toward traditional child care practices, such as following family advice (i.e., grandmothers of the participating children or social networks of mothers) for food progression. Caution should be observed before concluding that the findings of early introduction of different supplementary foods in this study are an illustration of cultural differences in feeding practice. Further investigations on the pros and cons of various feeding practices affecting the preterm population are warranted.
In our study, we found that mothers in both preterm and full-term groups perceived a similar amount of feeding issues when milk or formula was the only nutrient for infants. However, mothers with full-term infants reported experiencing higher stress associated with feeding-related issues. We offer possible explanations for this observation. First, the average age of the full-term infants in the milk group was much younger than the preterm infants (full-term: 2.6 ± 1.73 mo vs. preterm: 11.44 ± 7.30 mo). These results were expected because mothers with younger infants tend to generally experience higher stress, regardless of their infants’ birth status (Gray et al., 2012).
Second, mothers may have different expectations for feeding performance depending on their infants’ birth status. We assumed that mothers with preterm infants may have preconceived ideas that their “fragile” infants require extra care during feeding, and they accept the fact that their feeding will not be as easy as “normal” babies (Forsyth & Canny, 1991; Ross & Browne, 2013). However, mothers with full-term infants may expect their infants to be robust eaters, and consequently they have lower tolerance for their feeding issues. They consequently interpreted their infants’ behaviors of low endurance (i.e., falling asleep in the middle of feeding) as evidence of poor acceptability or a short attention span, and these perceptions could be sources of their distress.
Finally, mothers of preterm infants may receive more support from health professionals than mothers of full-term infants. In Taiwan, preterm infants were followed not only by their pediatricians but also by a team of developmental specialists by means of developmental follow-up clinics. These external support systems may provide just enough assistance to alleviate parental anxiety surrounding possible feeding issues.
Mothers with preterm children perceived significantly more feeding-related issues than mothers with full-term children when supplementary foods were introduced. Oral motor functions and sensory regulation–related feeding issues were found to be the two major sources of stress reported by mothers with preterm infants who had started solid foods, whereas atypical muscle tone was the major source of stress for mothers with full-term infants. DeMauro et al. (2011) reported similar findings on the postdischarge feeding pattern in early and late preterm infants. The participating parents also reported oral motor dysfunction and avoidant feeding behaviors as the most commonly perceived feeding problems in their preterm infants during the first year (DeMauro et al., 2011).
To further examine the hypothesis of maternal perceptions of vulnerability, we conducted an additional analysis to explore whether the level of preterm infants’ medical complications was a significant factor that affected maternal stress and perceived feeding issues. On the basis of previous studies (Maypole et al., 2011; Tallandini et al., 2015), we hypothesized that mothers with preterm infants who had more medical complications may report experiencing higher distress and more perceived feeding problems. However, we did not find significant differences among the groups with different levels of medical complications in their PSI–C score and number of perceived feeding issues. Forsyth and Canny (1991) reported that vulnerability was found to be higher in children with current health or developmental problems. The level of medical complications in our study, as measured by the Neonatal Medical Index, represented infants’ medical status at discharge, not their current health status. Further investigation of the effects of perceived vulnerability on maternal distress and perceived feeding issues is needed.
Limitations to this study warrant consideration. First, the existing discrepancy of sample distribution may limit the generalizability of the findings. The participating mothers were recruited from two different outpatient clinics. Mothers of preterm children were recruited from the developmental follow-up clinics, whereas mothers of full-term children were recruited from well-baby clinics. The age distributions of the two groups of children are different because of the differences in the schedules of follow-up visits. Our participating preterm infants tended to be clustered in more defined age groups (3 mo, 6 mo, 12 mo, 18 mo, and 24 mo), whereas our full-term infants were more equally spread, from 0 to 24 mo old. There were also differences in the hospital catchment area for both groups. Despite all participants being recruited from the same hospital, the demographic profiles of the mothers from both groups were different.
Because of the nature of the clinics (special clinic vs. regular well-baby clinic) and different referral systems, mothers of preterm children hailed from a wider range of demographic catch areas, whereas mothers of full-term children were from the local urban city. This discrepancy may also provide some insight into the socioeconomic differences between the two groups. Stratified sampling instead of the convenient sampling method would be a better approach for future studies. In addition, the smaller sample size in the milk groups might bias the results. Cautious interpretation of the results is suggested.
Second, we asked mothers from both groups to report supplementary foods they had offered to their children in the questionnaire as an open-ended question. The information we gathered was valuable but not comprehensive. A longitudinal study with a food diary and questions of food frequency, and specific strategies used by parents to promote or discourage eating, should be considered for future studies to understand the food consumption patterns of children and parental feeding styles from both groups (de Lauzon-Guillain et al., 2012; Vereecken et al., 2004). This information will also allow researchers to conduct comparison studies among different cultures by investigating potential cultural influences on parent feeding practices.
Third, the BFQ used in the study was originally designed for the preterm population. We acknowledge that it might be biased when used in a healthy population. Because our intention was to explore perceived feeding problems in the preterm population and use the full-term population as comparison, we decided to use it for participants of both groups. Observe caution when interpreting full-term infants’ feeding behaviors.
Our findings have important clinical implications. They demonstrate that parent-perceived feeding issues could be a major source of parenting distress, regardless of whether the infants were born preterm or full term. Mothers reacted to the perceived feeding issues of their infants differently depending on the infants’ feeding issues. Infants who had feeding issues, such as oral motor problems or difficulty in accepting new food, may be interpreted as demanding or less adaptable, and mothers may compound this behavior further by feeling depressed, isolated, and not adequate as a parent. Previous research has focused mainly on attachment and the mother–infant relationship with mothers who have preterm infants with feeding issues. However, we advocate that parenting stress should be examined comprehensively.
In our study, parents of preterm infants reported experiencing more feeding difficulty with their children, and they tended to advance their feeding with much more reluctance, evidenced by the later introduction of supplementary foods and less variety. They may benefit from increased anticipatory guidance about feeding dysfunction before hospital discharge. Practitioners should screen for parental discomfort surrounding infant feeding and discuss adaptations for caring for premature infants. Finally, practitioners should take detailed feeding histories to identify signs of feeding dysfunction and know when to refer at-risk patients for feeding therapy (DeMauro et al., 2011).
Implications for Occupational Therapy Practice
In addition to the clinical implications mentioned earlier, the results of this study have the following specific implications for occupational therapy practice:
When evaluating preterm infants’ feeding progress and experience, practitioners should be aware of the timing of the first introduction of supplementary food and food-related variety experience.
Mothers with preterm infants perceived significantly more feeding-related issues with their children and experienced higher distress compared with parents with full-term children. Practitioners should consider the primary caregiver–infant dyad as the unit when designing a feeding intervention.
Practitioners should be aware of cultural influences on feeding practices and be able to interpret feeding evaluation in context to design and implement culturally relevant feeding interventions.
Conclusion
Parents with preterm infants perceived significantly more feeding-related issues in their children and experienced higher distress compared with parents with full-term children. In addition, differences in parental feeding practices, such as feeding progression and timing, were also observed in this study. This study is the first step in understanding the behaviors of parents of preterm infants related to their feeding practices, the perceived feeding issues of their infants, and the feeding-related factors that contribute to stress. Although this study provides some information to help health professionals conduct more targeted interventions, it also indicates the need for further studies on parental feeding practices and preterm children’s feeding behaviors.
Footnotes
Acknowledgments
We thank the participating families for sharing their experiences with us.