Severe prematurity is associated with long-term developmental difficulties, and other factors such as heredity, health problems, and environmental conditions influence the extent of problems experienced [1
]. Previous studies have documented that temperamental difficulties reported in preterm born children (hereafter named preterms) affect later outcomes such as children’s self-regulatory capacities [3
], behavior problems, social skills [4
], and language development [8
]. In addition, a difficult temperament is associated with more parenting stress [10
]. Overall, it seems evident that high reactivity, negative emotionality, weak effortful control, and low self-regulation capacity are risk factors that put preterms at increasing risk across childhood [4
Most conceptualizations of temperament agree that it is biologically based and has relatively stable features, even though influenced by children’s maturation, learning, and environmental transactions [14
]. A child’s temperament covers the unique way emotions are expressed, characterized by latency, intensity, and duration of responses [16
]. The number of temperamental constructs and the terms used vary, depending on definitions and measures [17
There have been calls for studies that investigate how challenging temperamental expressions among preterms may be met by environmental adjustments [4
]. Some studies have focused on neonatal factors (e.g., less pain exposure) [18
] and the risk of altered brain maturation, since preterms may be exposed to non-nutritional stimuli in a hospital unit for months [20
]. Others point out that sensitive and supportive parenting is particularly important for children with challenging temperament dispositions [3
] as it is for preterms [23
]. Lastly, it is reported that factors such as the child’s age and gender and maternal factors (e.g., age and education) may influence temperamental development [26
According to the transactional model of development, child and parent characteristics and their dyadic and triadic interactions develop over time due to mutual influences [28
]. Several studies have focused on preterms’ susceptibility to parenting [11
]. Furthermore, parenting is described as more challenging after a preterm birth, because of the abruptness of the preterm delivery [30
], the parent–child separation [31
], unfamiliar early surroundings [32
], and these children’s often poorly regulated behavior that may seem incomprehensible to new parents [22
]. In families with very preterm children, parent–child dyads have previously been described as having less frequent symmetric co-regulation patterns and less positive and more neutral affective intensity of both infants and mothers, compared to families with full-terms [34
]. Thus, prematurity, children’s temperament, and parents’ capacity to make suitable adaptations to their children represent a mix of factors that may influence long-term transactional patterns. Early interventional effects on parenting stress have been reported [35
], and parenting stress is assumed to affect parents’ capacity to be sensitive [36
]. Mothers of preterms are reported to be equally sensitive as mothers of full-terms [38
], but preterms seem more negatively affected by low parental sensitivity than full-terms [24
]. Thus, parenting stress reported in infancy may be a predictor of parents’ perception of preterm born children’s temperament [21
This paper describes part of a randomized clinical trial (RCT) that aimed to determine whether an early, structured parental guidance program could improve developmental outcomes in a group of very preterm born children [41
]. Temperament was reported at similar levels in the intervention and control groups in infancy [42
]. One early finding was that the program positively influenced the association between child regulatory behavior and maternal stress reported at six months and one year [42
]. The current focus is on how parents’ perception of children’s temperament has developed across toddler and preschool years. This study explores whether dimensions of emotionality, activity, sociability, shyness, and soothability differ between an intervention and a control group from two to seven years, and to what extent differences may be associated with parental participation in an early parent training program. It is hypothesized that the modified Mother–Infant Transaction Program (MITP-m) empowered parents, in addition to decreasing parenting stress reported by mothers at all ages. Mothers in the MITP-m group reported a higher perception of competence in their parenting role at children’s ages of one, two, three, five, and seven [35
]. This may have enhanced their ability to adapt to their child’s temperament more successfully than the control parents.
The following question will be addressed: Is the development of temperamental constructs (emotionality, activity, sociability, shyness, and soothability), as reported by parents of preterms, positively associated with participation in an early parent training program?
This study explores parental perception of temperamental development among preterms across childhood. It investigates to what degree parents’ participation in an early, structured parent training program influences their perception of temperamental patterns from two to seven years. The main finding is that preterms’ emotionality and soothability were positively associated with participation in the MITP-m, while preterms’ shyness, sociability, and activity showed similar development in both groups. Mothers’ and fathers’ perceptions of preterms’ shyness and sociability were reported with high mean group stability across early childhood, while children’s levels of activity decreased from toddlerhood until early school age. On the other hand, as explained below, mean levels of children’s emotionality and soothability as reported by fathers formed somewhat different trajectories, depending on the group of the families. This created some group differences that persisted until age seven.
Emotionality, as defined by the EAS questionnaire, consists of questions that capture children’s tendency to react with strong negative emotions [15
]. Thus, while some degree of emotionality will characterize most children, higher scores may be difficult for parents to handle [24
]. PI mothers reported less emotionality in their children than did PC mothers at the age of two, and this difference persisted at later follow-ups. On the other hand, PI and PC fathers reported emotionality at similar levels in toddlerhood, but while PI fathers reported less emotionality with age, PC fathers continued to rate their children’s emotionality as high until age seven (Figure 1
In the MITP-m program, the concept of temperament was particularly addressed in the third home visit. This was approximately one month after discharge from the hospital, at a time where parents had a more nuanced perception of their child’s unique behavior. Even though many PI fathers could not join the sessions before discharge, due to long distances from home to hospital, older siblings, or work, they were more frequently present for the home visits. PI fathers may have had a catch-up effect from the four home visits, because each visit included reflections about the child’s signs and needs, which may have changed since the previous meeting. Although PI fathers rated their children’s emotionality at similar levels as PC fathers in infancy, they may have acquired knowledge and skills that helped them to cope more successfully in the long term. Thus, the intervention may have strengthened PI parents’ ability to adapt to their children’s emotional expressions and needs early in childhood. This assumption corresponds with previous results from this study. Olafsen et al. found a strong negative correlation between maternal stress and children’s regulatory competence reported by PI mothers at six months, while PC mothers reported a similar association at age one [42
]. The authors presumed that the intervention had changed the relationship between maternal stress and children’s temperament already in infancy. The current study indicates that this alteration of transactional patterns may have continued across the following early childhood years. It may also have increasingly influenced fathers’ perception of child temperament, as fathers became more involved in the care of their child after the first year. The Norwegian parental/maternity leave at that time was mostly used by mothers.
The findings also concur with another study that focused on how an early intervention was especially positive for preterms with negative emotionality in infancy [23
]. Those children showed a clinically meaningful decrease in social and cognitive problems three years after an intervention, compared to children with less negative emotionality [23
]. Enabling parents to manage their infants’ negative emotionality therefore seems to be a core element in early interventions. Negative emotionality challenges parents’ coping abilities, as their efforts to interact with the child frequently do not seem to work. Many spontaneous parent−child interactions may fail, and some parents may need more knowledge and coping strategies to understand why their child behaves in a particular way. A significant association between fathers’ report of emotionality and years of paternal education indicates that this may have been particularly challenging for fathers with fewer years of education.
How negative emotionality could best be approached was addressed in a recent experimental study [49
]. Mothers were asked to soothe a distressed infant simulator in early infancy, and this was compared with the mothers’ report of negative affectivity in their own child in later infancy. Specific behaviors, such as greater use of soothing touch and maternal vocalization in the simulation test were associated with less negative emotionality and fear in the mothers’ children [49
]. The importance of touch and vocalization were repeatedly mentioned and praised during the entire MITP-m program. The association discussed here confirms the need for a focus on temperament in interventional research and implementation [23
PI fathers rated their children with higher soothability than PC fathers did at age two, and this difference persisted across follow-ups until children’s age of seven. Fathers’ perception of their children’s soothability depends to a large extent on paternal experiences, knowledge about different strategies, and how each strategy may fit their child’s needs in different situations. Father−child interactions have also been described as more physical, stimulatory, and playful than mother−child interactions [50
]. In the current study, PI fathers received guidance in the MITP-m home visits about their child’s limited capacity to engage in intense activities. This enabled them to adopt milder ways to soothe their child and adapt this to new interactions as the child grew older. PC fathers, on the other hand, were limited to the strategies they naturally had available based on cultural traditions and their own experiences. This may have been challenging, especially for those caring for a prematurely born child with limited regulatory capacities. PC fathers may to a greater extent have perceived their child as being/having a problem, compared to PI fathers, who may have been empowered by the MITP-m intervention. This suggestion is supported by a previous study which showed more child-related stress reported by PC fathers than by PI fathers until children’s age of five, while paternal stress related to the parenting role was reported at similar levels by all fathers at all ages [35
]. By contrast, PC mothers reported more parenting stress than PI mothers at all follow-ups [35
], and maternal stress at children’s age of one was a significant covariate in the analysis of maternal perception of children’s soothability (Table S1
, additional resources).
A child’s availability for parental co-regulation and support is to some extent covered in the soothability dimension in the questionnaire used, identical with the framing in the previously mentioned CCTI questionnaire [46
]. Children’s soothability has been reported to moderate associations between negative child temperament and maternal sensitivity [52
], between child reactivity and maternal reports of co-parenting quality [53
], and described as an important predictor of elevated levels of comorbid internalizing and externalizing behavior problems in referred children [54
]. The present findings illustrate that child soothability may be strengthened by a program that probably empowers parents, increases their knowledge, and decreases their sense of stress in parent−child interactions.
Children’s sociability, shyness, and levels of activity were similar in the two preterm groups, after adjusting for parental level of education and parenting stress in infancy. This study thus confirms the high stability of several temperamental dispositions recently reported in a large sample of American children [26
]. In an early work by Buss and Plomin, both emotionality and sociability were described as “superfactors” that seemed to “pervade temperament measures in infancy and early childhood” [45
]. Even though these dimensions may be “superfactors”, our study has revealed a potentially important difference between them, namely that preterms’ sociability, shyness, and activity seem unaffected by early environmental changes, while programs such as the MITP-m and others may modify preterms’ emotionality and soothability [23
]. This difference is confirmed by the fact that no studies reporting interventional effects on children’s sociability or shyness have been detected despite thorough literature searches.
Strengths of this study are the RCT design, high participation rates across all follow-ups, and the statistical analysis that allowed all data to be included. Previously validated questionnaires have been used, even though the items originate from two different questionnaires. One limitation is that the original questionnaires were translated without back-translation and one question changed its meaning and was therefore excluded, leaving only four items to cover the shyness dimension. Despite this, a reliability analysis of data reported on each dimension by both mothers and fathers showed moderate internal consistency. In addition, this paper presents results from ten separate analyses, which might imply a possibility of Type I errors in the presentation of results. However, we chose not to perform a Bonferroni correction of the results because it seemed too conservative. To test the robustness of the current findings, this study should be replicated.