Parenting Preterm Children: Emotional Availability At Preschool Age

Graphical abstract Prematurity is a particular challenge for parents that severely affects the interaction and relationship dynamics between children and their caregivers. Previous studies mostly focused on the infancy of preterms but did not examine the development process. This study highlights whether preterm children and their main caregivers still differ in their emotional availability (EA) at preschool age. Caregiver-child interactions during the transition period of school entry were measured with the Emotional Availability Scales (EAS). N = 25 preterm (25–33 weeks of gestation) and n = 23 term children at an age of 5 years and their main caregivers were observed while playing a construction game that was eliciting either a low or moderate stress level. Caregivers of preterm children showed higher intrusiveness scores compared to caregivers of term dyads. Thus, it seems that parents of preterm children behave intrusively independently of their children's cognitive performance.


Introduction
The parent-infant relationship is shaped by parental caring behavior and the infant's signals and states as well.While term neonates and infants have a high capacity to alter their state and movement patterns according to the caregiver's actions, parents intuitively adapt their behavior according to the infant's signals.Premature birth (gestational age of 37 weeks or lower) may interfere with the development of parent-infant synchronicity.Worries and fears about their children's health and development, as well as the stressful birth, are known risk factors that may lead to deviations in parent-child interaction and hinder intuitive parental skills (Lilienfeld et al., 2012;Sarimski, 1999).Mothers of preterm infants more often report anxious, depressive, and post-traumatic stress symptoms (Davis et al., 2003;Gray et al., 2013;Jotzo, 2012;Karatzias et al., 2007;Korja et al., 2008;Shaw et al., 2006;Vigod et al., 2010).Postpartum anxiety and stress, as well as the perception of the baby's vulnerability, can result in altered parental caring.In contrast to mothers of term infants, mothers of preterms with high post-traumatic stress symptoms exhibit a more controlling and less sensitive interaction style with lower effective structuring (Allen et al., 2004;Feeley et al., 2011;Forcada-Guex et al., 2011).In addition, special challenges for parental skills are that preterms may send ambiguous behavioral signals in the first months after birth that are harder to perceive, interpret, and appropriately respond to, and that preterms show a lower social responsiveness.These challenges are accompanied by the neurological immaturity of preterms and a particular sensitivity to environmental influences (Bugental et al., 2008;Feldman & Eidelman, 2006;Miller et al., 2006;Pickler et al., 2010).

Emotional Availability in Preterm Parent-Child Interactions
Although some studies found mothers of preterm infants to be less sensitive in the first months postnatal (Feldman & Eidelman, 2007;Korja et al., 2012;Muller-Nix et al., 2004;Stack et al., 2019), others failed to discover significant differences (Agostini et al., 2014;Bilgin & Wolke, 2015).Jaekel et al. (2012) report less sensitive behavior of mothers of very low birth weight (\1500 g) or very preterm (\32 weeks of gestation) children at 6 and at 8 years of age.There are a few studies that describe mothers to have a lower ability to structure the interaction, as they show more inappropriate and restrictive structuring behaviors toward their preterm infants (Feldman & Eidelman, 2006;Lilienfeld et al., 2012).In comparison to mothers of term infants, mothers of preterm infants have been described as more intrusive with a more direct, active, and controlling interaction style and increased overstimulation (Agostini et al., 2014;Feldman, 2007;Forcada-Guex et al., 2006, 2011;Korja et al., 2012;Neri et al., 2015).They express more negative behavior in interaction with their preterm infants at the age of 36 months (Lilienfeld et al., 2012).In their first 3 years premature babies, as the other part of the dyad, were perceived as more passive, less responsive, and more negatively involved during interaction with their mothers (Feldman & Eidelman, 2007;Korja et al., 2012;Muller-Nix et al., 2004;Provenzi et al., 2017;Salerni et al., 2007).After 6 months of age, very low birth weight preterms lagged behind a full-term comparison group in their levels of responsiveness and involvement in their interaction with their mothers.These effects remained significant after controlling for maternal EA (Stack et al., 2019).

Cognitive Development Influences on Preterm Parent-Child Interactions
It is still unclear how parenting behavior affects the developement of preterm infants.To examine this link we focus on the association between the cognitive development of preterm children and caregiver-child interaction.Preterms have an increased risk for neurological and cognitive deficits.A meta-analysis reveals that in all age groups, they may suffer from cognitive impairments with a strong association between very preterm birth and intelligence.Associations were found between preterm birth and executive functioning as well as between preterm birth and processing speed.In children from 4 to 10 years, birthweight was associated with intelligence, and gestational age was associated with executive function (Brydges et al., 2018).Cognitive developmental delays were found in 16.9% (Pascal et al., 2018), and intellectual disabilities in 24.3% of very preterm children (Wolke et al., 2001).However, even moderately to late preterm children have a slightly increased risk for cognitive performance impairments (Chan et al., 2016).Particularly, high cognitive impairments of preterm children are attributed to pre-and perinatal stress, such as bilateral hippocampal injuries, ventricular bleeding, and hypoxemic-ischemic damage (Gadian et al., 2000;Isaacs et al., 2000;Linder et al., 2003).Also, invasive procedures during neonatal care contribute to long-term abnormalities in white matter microstructure and lower IQ (Vinall et al., 2014).
If we assume that parenting behavior differs according to the experience of prematurity, the question arises whether this is due to the stressful birth situation (which may or may not be conducive to the development status of the child) or due to the fact that parents unconsciously or intuitively adapt their behavior to the specific needs of their children.The theory of compensatory care assumes that parents may increase caregiving behavior to sick or high-risk infants to attenuate the effects of potentially harmful events, like preterm birth.It proposed that a homeostatic mechanism exists within the caregiver-infant dyad, entailing that caregivers with a sick or fragile infant provide more parental care in order to compensate for and reduce negative effects for the infant (Beckwith & Cohen, 1978).
Up to now, only a few studies have shed light on the link between variations in child cognitive development outcomes and parenting interaction behavior.But, different patterns of maternal postpartum behavior may have an impact on infant neurobehavioral maturation (Feldman & Eidelman, 2003) and child cognitive growth (S.E. Cohen & Beckwith, 1979;Feldman et al., 2004).These include eye contact, vocalization, and emotional touch.Less intrusive maternal behavior as well as smallfor-gestational age and birth weight were associated with cognitive outcomes of 2-year-old preterms (Feldman & Eidelman, 2006).Maternal sensitivity positively affected very low birth weight and very preterm children with cognitive deficits at the time of school entry; with higher maternal sensitivity associated with higher task persistence (Jaekel et al., 2012).High sensitivity again had an equally positive effect on the cognitive development of very low birth weight preterm children and term children at preschool age (Hoff et al., 2004).Treyvaud et al. (2016) describe associations of higher levels of parental-child synchronicity, parent facilitation, sensitivity, and positive affect at 2 years with better cognitive outcomes at school age in very preterm children.Additionally, they found higher levels of parental intrusiveness and negative affect correlated with poorer cognitive outcomes.

The Present Study
Most of the recent studies investigated the interaction of preterm children and their parents during the first 3 years and show that preterm dyads differ in their interaction patterns as compared to term dyads.So far, little research has been done on the interaction beyond the first years of life.The year before starting school and the school entry itself is accompanied by developmental tasks including challenges for the parental skills and the regulatory abilities of the children.To close this gap, the aim of the present study is to address whether the interaction behavior of preterm dyads changes when children grow older.For this purpose, we have examined 5-yearold children and their parents at the time of school entry.The interaction and relationship dynamics were investigated during a play situation based on the concept of EA (Biringen et al., 2014).Furthermore, we considered whether differences in interaction patterns are an adequate adaptation to the special needs of the preterm infant or a negative influence on the child's autonomy and developmental outcome.To answer this question, differences in EA between preterm and term caregiverchild dyads were assessed.

Participants
We originally recruited N = 54 caregiver-child dyads.Data from five dyads were excluded from the statistical analyses because of missing data.Additionally, data from one dyad were excluded because the main caregiver was absent.The final sample included N = 48 dyads, each consisting of the child and its main caregiver (43 mothers, five fathers).The identification of the main caregiver was based on the parents' self-report.Among the children, there was a mean age of M = 67 months (SD = 4.6) and an equal distribution in terms of gender (48% female).All preterm subjects were participating in a standardized overall developmental follow-up assessment before school entry in the Social Pediatric Center (SPC) of the Children's Clinic in Dortmund, Germany.For the recruitment of the term children, posters in surrounding kindergartens, SPC, Ruhr University Bochum, and University of Health Sciences Bochum were used.Parents gave written informed consent for themselves and their children to participate in the study.They received test results of the standardized developmental assessment in written form.Preterm children that met the following criteria were included in the study: (a) born \37 gestational age, (b) not exposed to illicit drugs prenatally, (c) no neurological or psychiatric disorders, and (d) IQ .70. Inclusion criteria (b) to (d) were also applied to the control group.
The local ethics committee of the Psychological Faculty of Ruhr University Bochum approved our study.The study was in accordance with the recommendations according to the World Medical Association (revised Helsinki Declaration).

Measurements
Emotional Availability.The quality of the caregiver-child interaction was investigated in line with the concept of EA (Biringen et al., 2014).According to this concept, EA is seen in a dyadic relationship and interactional dynamics.It evaluates a mutual awareness of emotions, perception, and expression within an interaction (Biringen, 2005).The EAS (Biringen, 2008) are the only global system that provides a dyadic/relational perspective on caregivers and their child.EA was surveyed in a play situation, during which the child and its main caregiver were instructed to reconstruct a chair and a bridge with wooden blocks (117 mm length, 23.4 mm width, 7.8 mm height; Kabouter Plankjes, Kapla) according to a pictorial model.Parents were instructed that they may help their child as they normally do and that it is not necessary that the picture is rebuilt completely.
The EA scales include four dimensions of the caregiver's EA (sensitivity, structuring, non-intrusiveness, non-hostility) and two dimensions of the child's EA (responsiveness, involvement).In the caregiver-child interaction that is to be observed, points between one (minimum value) and seven (maximum/optimal value) can be awarded on each of the six dimensions.Sensitivity consists of the behaviors and the emotions used by a caregiver to create and maintain a positive emotional connection with the child.Structuring refers to the capacity to support and guide the child's learning and permits autonomy, and non-intrusiveness to the ability of a caregiver to follow the child's lead during play without interrupting.Non-hostility is the ability to regulate own negative emotions in interactions with the child and child responsiveness focuses on the child's emotional and social responsiveness to the caregiver.Finally, child involvement refers to the child's ability to involve the caregiver in his or her play or activity in general, thus including the adult in the interaction (Biringen, 2005;Biringen et al., 2014).
The observed and video-taped interaction lasted at least 7 min (M = 11.99,SD = 3.21).When a dyad completed the chair model in less than 7 min, the experimenter presented the bridge model and asked them to reconstruct it as well.The task was chosen ageappropriately to be challenging but manageable and to cause a low to medium stress level (e.g. when the construction or parts of it collapsed).Assessment of EA followed the fourth edition of the EA Scales Manual, Infancy / Early Childhood Version (Biringen, 2008).The EA Scales are quite flexible in application across variations in child age, characteristics of the adult, and setting or context of observation (Biringen et al., 2014).The validity and reliability were rated as good in various studies (Altenhofen et al., 2013;Biringen et al., 2012;Bornstein et al., 2006;Din et al., 2009;Easterbrooks et al., 2012;Waters & Deane, 1985;Ziv et al., 2000).The videos were rated by two independent and trained raters.The mean value of the ratings was used for statistical analysis.The inter-rater reliabilities for the dimensions of EAS were ICC = .76for sensitivity, ICC = .66for structuring, ICC = .86for non-intrusiveness, ICC = .86for non-hostility, ICC = .86for responsiveness, and ICC = .71for involvement.To improve congruency between raters, a practice phase, and an exchange about observations, deviations, and understanding of the items took place.
Cognitive Performance.Children's IQ was assessed with the established Wechsler Preschool and Primary Scale of Intelligence-III German Version, 3rd, revised, and expanded edition (WPPSI -III; Petermann et al., 2014).
Psychosocial and Neonatal Data.Caregivers completed questionnaires on demographic and neonatal data and the child's development.The assessments included questions regarding pregnancy, birth complications, the infant's achievement of developmental milestones, socioeconomic status, and families' home situation.The Family Adversity Index (FAI; Rutter & Quinton, 1977; German adaptation by Voll & Allehoff, 1982) was used as a cumulative risk index for assessing the quantitative psychosocial burden in the family system.The Eltern-Belastungs-Inventar (EBI; Tro¨ster, 2011) serves to record the perceived parenting stress.

Descriptive Statistics
Neonatal and Psychosocial Sample Characteristics.Table 1 displays neonatal and psychosocial data for the preterm (n = 25) and term group (n = 23).The caregivers of term and preterm children reported low current family stress, and a low overall psychosocial risk of their children, and do not differ significantly in their level of education but as expected, preterm, and term children differ significantly in all neonatal characteristics.Sensitivity Analysis.Given a = .05, 1 -b = .8,and group sizes of n = 25 (preterm group) and n = 23 (term group), differences in EA of size d = 0.73 (i.e., between medium [d = 0.5] and large [f = 0.8] effects, cf.J. Cohen, 1988) could be detected (Faul et al., 2007).

Emotional Availability
Figure 1 shows the differences in EA between the preterm and the term group.Term caregiver-child dyads scored higher on all six dimensions of EA than preterm caregiver-child dyads.We used Bonferroni corrections to avoid type I error (Howell, 2002).As such, statistical significance was set at p \ .008.

Non-intrusiveness and IQ
Table 1 shows that preterm children had significantly lower IQ scores than term children.This is in line with prior study findings (Brydges et al., 2018).Assuming that the correlation between birth status and intrusiveness is just due to the IQ of the children, the correlation should disappear when the IQ is extracted.For investigation, we have calculated a partial correlation that shows a significant correlation after controlling for IQ as a covariate, r = .26,p = .04.

Discussion
The present study examined the interaction of preterm children and their main caregivers during the transition to primary school compared to term dyads.On the one hand, the child's organism matures, catching up on individual developmental delays under certain circumstances, including an increase in the child's abilities, and skills (e.g., Poehlmann et al., 2011).On the other hand, parents must be able to perceive and react to their child's development.We have systematically investigated four dimensions of the caregiver's and two dimensions of the child's EA.Since this is-to the best of our knowledge-the first study to look at this, it was unclear from the outset whether there were any significant differences at all, and if so, on what dimension.
There was a clear and significant difference in one dimension.Main caregivers of preterm children behaved more disruptively and intrusive than caregivers of term children.Differences in the intrusive behavior of parents were observed in the early childhood of preterms as well (Agostini et al., 2014;Feldman & Eidelman, 2007;Forcada-Guex et al., 2006;Salvatori et al., 2016).Adult non-intrusiveness refers to qualities such as the absence of overstimulation, interference, overdirection, or overprotection, and hence, is a way to encourage ageappropriate autonomy while maintaining the relationship connection.An excessive level of intrusiveness may be suboptimal for the child's development (Feldman, 2007).High intrusive behavior of caregivers may lead to lower selfconfidence in children, limited self-efficacy experiences, and may have a long-term negative effect on a child's self-awareness.Especially in the transition from preschool to school age, it becomes more important for children to be able to act independently and self-confidently because their skills are more challenged.A supportive and child-oriented interaction from early childhood to adolescence promotes a high expectation of self-efficacy, for example, the subjective conviction that difficult situations can be handled well (Hohm et al., 2017).A longitudinal study, examining preterm infants and their mothers at several time points, found that less intrusive maternal behavior predicted more secure attachment at 16 months, better effortful control at 24 months, and fewer motherreported externalizing behavior problems at 72 months (Poehlmann et al., 2014).
In line with the theory of compensatory care (Beckwith & Cohen, 1978) we tried to clarify whether the observed intrusive behavior of parents may be beneficial for the development of preterm children.It is implied that parental intrusiveness is a conscious or unconscious strategy to compensate for the developmental delays of their preterm children.Thus, we took the child's IQ into account.Our findings support this assumption, showing that caregivers of preterm children had significantly lower scores on the non-intrusiveness scale even when IQ was controlled.This suggests that caregivers do not interact more intrusively because of the lower cognitive performance of their preterm children.They rather seem to behave intrusively perhaps as a remnant of the particularly demanding and challenging situation during and shortly after birth.It could be possible that intrusive behavior develops with the experiences of an abrupt end of pregnancy and prenatal bonding process, and the stressful birth situation (Forcada-Guex et al., 2011;Neri et al., 2015).In line with attachment theory, situations that trigger feelings of helplessness may disorganize the Caregiving System (Bowlby, 1982).For instance, mothers reported feeling insecure and anxious about interacting with their infants in the early periods after discharge, and babies were perceived as sleepy and unresponsive (Nicolaou et al., 2009).The stressful context of having a prematurely born infant can also limit the perspective on the child.Yatziv et al. (2018) described that mother's executive functions were associated with lower misinterpretations of their infant's minds during interactions (nonattuned mind-mindedness), only among mothers of term infants, but not among mothers of preterm infants.They concluded that stress leads to a more automatic processing instead of controlled processing.But maternal psychological stress caused by preterm birth does not necessarily indicate low-quality parenting (Hall et al., 2017).Besides, due to the early exposure to the extrauterine environment, preterms expressive needs, and signals are harder to interpret accurately and respond correctly to, which consequently may be a starting point for the development of non-promoting interaction dynamics (Feldman & Eidelman, 2006;Pickler et al., 2010).
Furthermore, it must be considered that different interaction behavior of parents can particularly be beneficial for the development of preterm children in their early months of life but does not prove to be beneficial when their abilities increase.A longitudinal study from 18 to 30 months that examined preterm maternal-child interactions found an increase in maternal sensitivity, structuring, non-intrusiveness, and child responsiveness over time (Salvatori et al., 2016).Other studies also show that high sensitivity is meant to be conducive to child development (Bakermans-Kranenburg et al., 2003), while parental control should decrease as competencies and self-regulatory abilities increase (Dallaire & Weinraub, 2005;Grolnick & Farkas, 2002;Landry et al., 2000).However, a combination of high maternal sensitivity and control may benefit the development of children at risk (Forcada-Guex et al., 2006;Landry et al., 1997).
One limitation of the present study is that the differentiation between mothers and fathers is missing.They were both included if they considered themselves as the main caregiver.It has been shown that mothers and fathers may differ in their perception of the child's specific needs and, as a result, in their parental interaction pattern.However, in our study mothers and fathers did not differ significantly on the non-intrusiveness scale.Nevertheless, it is known that there are significant differences in the interaction styles of parents regardless of birth status.Fathers will often show more exciting and animated engagements, associated with higher arousal (Feldman, 2003;Forbes et al., 2004;Kokkinaki & Vasdekis, 2015), more physical stimulations, and a tendency to more often challenge the abilities and selfconfidence of their children (Kindler & Grossmann, 2008).To date, there is a very small set of studies investigating father-infant interaction in the context of prematurity.While mothers show higher control, fathers show higher unresponsiveness in interaction with their threemonths-old preterm infants (Neri et al., 2017).However, the present results do not seem to be simply due to the inclusion of caregiver-child dyads with fathers.We conducted additional analyzes in which we excluded all caregiver-child dyads with fathers, but this did not change any of the statistical results.The difference on the dimension non-intrusiveness remains significant (U = 110.5,p \ .008),and there are still no differences on all other dimensions (p ..008).Nevertheless, it would be interesting to look at differences between mothers and fathers in caregiver-child interactions in future studies with equally sized groups, which would allow more meaningful comparisons.
Further studies could also explore the influence of the family's cultural background.In the present study, families with and without a migration background were unequally distributed which is not ideal.Families from different cultures may prioritize different parenting approaches.Research suggests that constructs used to capture differences in parenting may not apply universally across cultures.In socio-centric cultural contexts, the concept of sensitive maternal care includes the expectation that the mother guides, structures, and controls the child's behavior (Carlson & Harwood, 2003;Harwood et al., 1999;Keller, 2008).Intrusive-insensitive mothering at an age of 2½ years was associated with poorer subsequent reading achievement for African Americans but not Mexican American children, suggesting qualitatively different parenting constructs for both cultural groups (Caughy et al., 2017).In West Cameroon NSO farmer families, responsive and controlling behavior was identified as the optimal handling of infants, which from a Western perspective can be misunderstood as intrusive (Yovsi et al., 2009).It thus seems possible that the unequal distribution of families with and without a migration background could have affected our results as well.To see if this was the case, we conducted an additional analysis in which we excluded families with a migration background.There was still a significant difference in non-intrusiveness, U = 66.0, p \ .005,demonstrating that our results were not simply due to the unequal distribution of families with and without a migration background.
We conclude that an intrusive parent-child dynamic is associated with preterm birth.That means that a lower EA of parents of preterms seems to exist independently of lower child IQ in middle childhood and in the critical developmental phase of the transition to school as well.As mentioned earlier, children learn to act independently and self-confidently during the transition from preschool to school age which requires a supportive and child-oriented interaction from the parents.Therefore an early preventive intervention, for example, the short-term Ulm Modell (Ziegenhain, 2004), which integrates the concept of self-regulatory behavior (Als, 1982;Brazelton, 2009) and attachment theory, may be exceedingly helpful to support the parents in coping with early potential stress, and in noticing and interpreting preterm infant signals in order to be able to respond appropriately.Further research on the possible specific effects of this intervention for this group may be needed.

Conclusion
In this article, we compared interactions of caregivers with preterm children and caregivers with term children during the transition to primary school compared to term dyads.Even at this age, behavior still differed significantly, especially in terms of the intrusiveness of the caregivers.Caregivers of preterm children were more intrusive compared with caregivers of term children in a typical, everyday situation (playing together).These findings suggest that in the future we should pay attention not only to the short-term but also to the long-term psychological effects of preterm birth.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Figure 1 .
Figure 1.Mean differences in emotional availability for caregivers and infants in the preterm and term group.

Table 1 .
Neonatal and Demographic Information (Mean, SD) for Preterm Infants (PT) and Full-Term Infants (FT).
a chi-square-test b t-test c U-test; 1 = missing data.