Introduction
Children with developmental language disorder (DLD) have no hearing disabilities and show no evidence that their language difficulties associated with a known biomedical aetiology (such as cerebral palsy) [
1]. Some 7–10% of children in the UK enter school with impaired language abilities [
2].
Notwithstanding the absence of neurological abnormalities and cognitive deficits, children and adolescents with histories of DLD do show a heightened risk of various other developmental difficulties. For example, as a group, they tend to manifest higher levels of conduct disorder and hyperactivity than do typically developing peers [
3,
4]. They are prone to greater difficulties in peer relations and friendships [
5‐
7]. They also have higher levels of mental health difficulties, such as anxiety, fearfulness, depressive symptoms and panic [
8,
9].
One area of particular vulnerability for children and adolescents with DLD is emotional regulation. Compared to typical peers, these young people are almost twice as likely to show clinical levels of emotional difficulties [
5,
10]. A meta-analysis of existing evidence suggests that, on average, children with DLD are above the 70th percentile on severity of emotional difficulties [
11]. With the exception of very early childhood, between the ages of 4 and 7 years [
12], longitudinal studies have found higher levels of emotional difficulties in DLD not only across childhood but into young adulthood [
3,
11]. The accumulating evidence indicates a clinically important connection between DLD and the development of emotional difficulties.
The studies available to date are informative of the overall trajectory of emotional difficulties in DLD. Comparisons of results across studies indicate that trajectory of emotional difficulties in DLD appear stable across time, with a modest increase in difficulties with age. Such a trajectory of emotional difficulties is consistent with those found in general population studies [
13,
14]. It is important to note, however, that some investigations that have examined childhood baseline levels of emotional difficulties and later emotional outcomes in DLD have not found stability.
Some investigators have reported longitudinal increase in symptomatology [
8], whilst others have found amelioration/resolution of difficulties [
15] and still others have reported curvilinear patterns, i.e., decrease followed by increase [
16]. Although such inconsistencies are likely to reflect, at least in part, differences in the samples studied and methodological differences with respect to participants’ ages and measures used, they may also indicate individual differences. There may be groups of children with DLD that experience different developmental trajectories of emotional difficulties. DLD is known to be heterogeneous; different children manifest different areas and/or combinations of language difficulties in respect of expression, comprehension, and pragmatic performance [
1,
17].
We also know that there is variability in the ways in which DLD is associated with developmental difficulties in other domains of functioning, such as behaviour or social interactions [
3]. In the social domain, Mok et al. [
7] have documented clear differences in the development of difficulties with peer interactions. One group of children with DLD in that study experienced problems with peers from childhood through adolescence (persistent). Another group had peer difficulties in childhood that appeared to resolve in adolescence (childhood limited). Another group experienced an increase in peer problems from early adolescence (adolescent onset). Other children experienced relatively modest peer difficulties throughout the same period (low/no problems). In the present study, we ask whether similar trajectories are identifiable in respect of emotional difficulties in children with DLD and whether the trajectories followed in respect of emotional difficulties are aligned with those identified in respect of peer relations: that is, do problems in one of these areas invariably signify that problems are likely in the other?
There is some evidence to indicate that emotional and peer problems are associated in childhood and adolescence in general [
18,
19], and this has been reported in DLD populations in particular [
7]. Mok et al. [
7] found that, with respect to peer problems, children in the childhood-onset persistent problems group and those with adolescent-onset problems showed higher levels of emotional symptoms than those with low/no problems. On this evidence, then, it appears that these difficulties are interwoven. What is less clear is how they are interrelated across development. For example, a relatively straightforward expectation could be that difficulties in each domain develop in parallel, due either to one type of problem precipitating the other (e.g., children with emotional difficulties are less able to form and maintain successful peer relations), or because the variables are linked bi-directionally (i.e., each problem type exacerbating the other over time: emotional difficulties impact on peer relations and vice versa), or they share common etiological factors which affect growth of both emotional and peer problems. A more complex possibility is that different children show different patterns of joint trajectories. That is, some may manifest parallel developments across peer relations and emotional regulation, while others may show divergent trajectories. Relatively little research has been conducted into co-occurring developmental trajectories, but the issue is crucial to advancing our understanding of developmental relations and to informing diagnosis and clinical interventions [
18,
19]. Hence, a principal purpose of this investigation was to determine whether these two areas of problematic development occur together over time.
Another aim of this study was to examine potential factors associated with developmental trajectories of emotional and peer problems from childhood to adolescence.
One possibility involves the consequences of facing adolescence with the burden of persisting language difficulties. We examined expressive, receptive and pragmatic language skills and hypothesized that severity of language disorder would be associated with increased difficulties in adolescence. This is because research with children with DLD suggests that language skills, and in particular pragmatic skills, are associated with how well children comprehend emotions and emotional descriptions, how well they self-regulate their own emotions [
20,
21] and whether they engage in successful peer relations and friendships [
3,
7]. We also anticipated that social abilities are likely to play a role in the progress of emotional difficulties. Problems with peer interactions have been shown to be associated with increasing levels of emotional difficulties [
7,
22] whilst prosociality is positively associated with emotional adjustment [
23,
24]. Hence, we expected that lower prosociality in later childhood would be associated with less favourable joint trajectories, namely persistent problems in emotional and peer relations throughout childhood into adolescence and adolescent-onset problems, i.e., increasing problems in these domains during adolescence.
Other factors are known to bear on vulnerability to emotional difficulties which may also bear on social adjustment. These include gender [
14,
25] and parental history of mental health difficulties [
26]. Population studies have revealed that an increase in emotional difficulties in adolescence is more pronounced in girls [
13,
14]. On this basis, we predicted that there would be a larger proportion of girls with DLD with adolescent-onset emotional difficulties. Parental mood and anxiety disorders are known to be associated with increasing levels of emotional, social and behavioural difficulties in their offspring [
22,
27,
28]. Hence, we expected an association between parental mental health difficulties and increasing symptomatology, such that higher indications of parental mental health difficulties would be associated with the less favourable joint trajectories of emotional and peer problems.
Discussion
To the best of the authors’ knowledge, this is the first study to examine joint longitudinal trajectories of emotional difficulties and peer relation problems in children with DLD. The findings reveal five distinct patterns of development: (1) low levels of problems in both domains throughout the period studied; (2) childhood onset of problems in both, which remained persistent throughout; (3) adolescent onset in both; (4) low levels of emotional difficulties throughout, alongside increasing peer problems; and (5) emotional difficulties relatively high in childhood and resolving into adolescence, while peer problems were relatively low throughout. This qualifies previous findings based on data aggregated across whole samples [
3,
49] and, importantly, reveals that the two areas of difficulty do not invariably occur together.
Slightly over half of the sample did show parallel developments. These were the first three groups listed above. For these children, then, to the extent that there are problems in one of these two aspects of development, there will be problems in the other. This is consistent with the possibility that onset of difficulties in one area promotes difficulties in the other, or with assumptions of bidirectional causality, or with the possibility that a third variable (e.g., underlying common etiological factors, such as genetic factors) explains developments in both areas. These are familiar explanations in developmental psychopathology: it is often the case that children with problems in one area of development have additional problems [
50].
The presence of two other groups (together amounting to 46% of the sample), however, complicates the overall picture. In one case, despite relatively high peer problems which increased into adolescence, emotional difficulties were low throughout. For at least some children with DLD, then, peer problems do not precipitate emotional difficulties, and a ‘third variable’ cannot be so straightforwardly attributed responsibility if one domain is seemingly unaffected. Possible interpretations are that these children had sufficiently robust emotional self-regulation or self-efficacy to enable them to withstand emotional problems or that other sources of social support, such as parents, bolstered them against emotional difficulties [
16]. In the final group above, peer problems were relatively low throughout, but emotional difficulties were relatively high in childhood and decreased into adolescence. A possible interpretation is that, for these young people, positive peer relations provide a context that, over time, is conducive to the moderation of emotional difficulties [
51,
52].
Taken together, these findings lend support to arguments that development in children with DLD is heterogeneous—not only in respect of their language disorder but also in terms of how these are associated with other important aspects of personal and social adjustment. This is important from a theoretical perspective, because it suggests that no one explanation—at least, as currently formulated—can account for all manifestations of DLD and its concomitants [
1].
What variables are associated with differing patterns of development of personal and social adjustment in individuals with DLD? We did not find that either comprehension or expressive language difficulties differed among the five joint trajectory groups. It is important to stress that the absence of differences among these groups (all with histories of DLD) does not mean that comprehension or expressive abilities are irrelevant to emotional and peer difficulties [
3,
6]. What the present findings do suggest is that, among children with DLD, whatever comprehension or expressive difficulties they have as measured by the instruments used in this study, do not strongly influence which joint trajectory group they fall into.
One aspect of linguistic ability, however, that does appear to be associated with trajectory group membership is pragmatic competence. Children who followed a persistent trajectory, with high levels of emotional and peer problems from childhood to adolescence, had significantly lower pragmatic scores than most of the other groups, and the increasing peer problems group had the second lowest pragmatic scores. More profound limitations in the ability to handle the functional, interpersonal nuances of pragmatic language may put a young person with DLD at a greater risk of following the less favourable joint emotional–peer trajectories. Skills such as making inferences, appropriate conversational turn taking, and tuning into the facial expressions of others are likely to affect emotional recognition [
53] and emotional self-regulation [
21]. Pragmatic language difficulties are not always apparent to co-locutors, particularly in interaction with peers in childhood. In adolescence, pragmatic difficulties may well be more salient [
54]. Adolescents with poor pragmatic skills may thus encounter “demands that exceed capacity” [
55]. Adolescents with DLD are likely to experience difficulties processing input from peers about feelings and emotional management, which in turn could lead to feelings of frustration, worry and fearfulness. This argument is further supported by our finding that the children who did not fall into the trajectories defined by peer problem skills (i.e., those in Resolving Emotional and Low Level) and those with peer problems emerging later (i.e., adolescent onset) did not have lower pragmatic competence. It remains for future research to examine whether peer problem-free childhood affords the development of pragmatic skills to a competent level.
We did not obtain clear evidence of a gender imbalance associated with particular trajectory groups. Of particular interest, the findings did not support expectations that proportionally more girls would follow the adolescent-onset trajectory. Population studies report higher levels of depressive symptomatology among teenage girls [
25], and we expected that this pattern would be reflected in terms of higher levels of emotional and peer difficulties emerging in adolescence among our female participants. Certainly, many of our participants did show increasing levels of emotional difficulties over time, but this was not a gender-specific outcome. However, it should be acknowledged that, as in most samples of children with developmental language disorder, the proportion of females here was small (24%); future researchers might consider over-recruitment of females to provide more information on the relationship between gender and emotional and peer difficulties in young people with DLD.
The findings with respect to prosociality were also significant. Consistent with expectations, the two least favourable joint trajectory groups (persistent and increasing peer problems in adolescence) did have the lowest mean prosocial scores, and post hoc comparisons between each of these groups and the other joint trajectory groups were statistically significant. Thus, the data not only suggest that lower prosociality accompanies problems in emotional and peer relation domains, but that prosociality is strongly associated with the type of pattern of emotional and peer difficulties that will be followed from childhood to adolescence. We note, however, that these findings are based on the Manchester Language Study (MLS) sample. MLS participants included children with identified developmental language disorders who were receiving support and intervention in language units in childhood. We also note that previous research with the MLS demonstrates that individuals with DLD had continued to develop their expressive and receptive language skills during early adolescence into young adulthood [
32]. The early identification of language difficulties coupled with the context of early, intensive language support received in educational contexts such as language units may have nurtured socialisation processes and the development of emphatic concern, which in turn may have influenced the development of prosociality in individuals who participated in the MLS. Indeed, research with the MLS sample suggests that young people with DLD are prosocial and exhibit stable developmental trajectories of prosociality throughout adolescence [
56]. It is also important to note, however, that more individual differences in prosociality have been found by other researchers. Lindsay and Dockrell [
57], for example, found more individual differences in prosociality in their sample of children with DLD drawn from a variety of schools with different educational provisions in the UK. They found prosocial scores improved between 8 and 12 years of age but worsened by 16 years. Further research with other samples of individuals with DLD, such as community samples or samples of individuals with unidentified DLD would help to unpick the complex relations among these variables over time.
We report preliminary thought-provoking findings that raise the possibility that parental mental health difficulties may be associated with their offspring’s personal and social adjustment. The persistent problems trajectory group had the highest mean score on a measure of parental self-report of their own histories of mental health problems during childhood and adulthood as well as the highest proportion of both parents reporting issues with their mental health. This is consistent with evidence from studies in the general population showing that poorer parental mental health is a predictor of emotional difficulties in children and adolescents [
26]. What this paper adds is that, in the context of DLD, this factor may also be associated with concomitant, persistent peer problems. There are a range of potential mechanisms by which parental mental health may be associated with child’s mental health which may be involved in context of DLD. Goodman and Gotlib [
59] suggest three mediating and transactional pathways (bio-developmental; psychosocial and contextual) regarding postnatal distress and child emotional and behavioural development which may be worth investigating in future research in this area. It needs to be noted, nonetheless, that in this study we did not have standardised clinical measures of parental mental health with known validity and reliability and the differences observed were preliminary and indicative (see also [
58]). Thus, the present finding in this regard should be interpreted with caution. Given the possibility that parental mental health bears on the important aspects of child development in this vulnerable population, the present results warrant further research.
In the same vein, further research could also address some of the limitations present in this study. This investigation used different measures at different ages which may have introduced measurement variability which future research could control for using instruments which span the period of development examined. In addition, minimising attrition so that the same children can be followed across development and maximising completeness of data gathered on associated factors could also be addressed in future longitudinal investigations.
The pattern of findings is important from a clinical perspective. The fact that over half of the sample showed parallel trajectories in emotional and peer domains suggests that diagnosis and monitoring of children with DLD should include examination of much more than language skills. The fact that a large part of the sample showed divergent trajectories across the two domains also warns, however, against assuming that identification of one problem area has clear implications for others; instead, strengths and difficulties need to be identified on an individual basis and potential factors associated with worse outcomes in adolescence. The findings of this investigation also suggest that clinicians should also be sensitive to the possibility that young people experiencing sustained difficulties in both emotional and peer domains may be living in families where there are higher than average levels of parental mental health problems. Furthermore, the difficulties of children with either emotional or peer problems may be less evident than children with both difficulties and professionals need to be vigilant in identifying these needs. In turn, clinical interventions need to take into account the potential breadth of a child’s difficulties, individual areas of robustness/resilience that can be built upon in therapy as well as the potential need for whole family approaches to intervention.
The evidence obtained in this investigation does also offer some positive news concerning emotional and peer difficulties in at least some individuals with DLD. Approximately, 11% of the participants had low levels of difficulties in both domains throughout childhood and adolescence. An additional subset, approximately 24% of the total sample, had emotional problems in childhood that appeared to be resolving during adolescence. These children had low levels of peer problems throughout and also tended to have better pragmatic language scores. Thus, there are encouraging indications not only that some children with DLD do experience relatively favourable trajectories but also that we can identify a particular area of language skills that may be amenable to improvement, with the potential for broader benefits for these young people’s adjustment.