Statistics from Altmetric.com
What is already known on this topic
Developmental delay occurs in 6–11% of preschool children born moderately to late preterm, compared to 4% of those born full-term.
Early detection of the co-occurrence of developmental and behavioural problems may help to identify those moderate preterm-born children most in need of early intervention.
What this study adds
In preschool children born moderately to late preterm, 25–39% of those with developmental delay also have emotional and behavioural problems.
Increased awareness of behavioural and emotional problems is warranted in moderate and late preterm-born children with developmental delay.
Further research is needed to determine the benefits of addressing co-occurring neurodevelopmental problems at preschool age.
Eight out of 10 preterm children are born at 320 to 366 weeks’ gestation, that is, they are moderate to late preterm (MLP) children.1–3 Compared to full-term children, MLP children are at increased risk of neonatal morbidities, such as respiratory problems, hypoglycaemia and hyperbilirubinaemia,4 and of long-term neurodevelopmental problems, such as developmental delay and learning problems.5–10 The risk of developmental delay is approximately twice as high in MLP than in full-term children, with developmental delay occurring in 6–11% of preschool MLP children and in 4% of full-term children, as measured by abnormal total scores on the Ages and Stages Questionnaire (ASQ).8 ,11
Developmental delay frequently co-occurs with emotional and behavioural problems (EBP).12 ,13 For example, many children with motor and coordination problems also have attention, social and/or psychosomatic problems.14–16 In previous research among very preterm children (<32 weeks’ gestation), high rates of co-occurring developmental problems and EBP were found.17 ,18 Up to 50% of very preterm children appeared to have more than one developmental or behavioural disability at the age of 5 years, compared to 8% of full-term controls.17 ,18 However, the co-occurrence of developmental delay and EBP in MLP children of about the same age is unknown.
Our main aim, therefore, was to determine the co-occurrence of developmental delay and EBP in MLP children at preschool age, and to compare the results with full-term controls. We hypothesised that MLP children would have an increased risk of co-occurring problems since intra-uterine brain development is disrupted 3–8 weeks before term gestation, which may cause changes in the functional connectivity processes of neural motor networks.19 Second, we aimed to explore whether other risk factors for co-occurrence could be identified.
We derived the data for this study from the Longitudinal Preterm Outcome Project (LOLLIPOP), a large prospective cohort study in the Netherlands. LOLLIPOP was designed to investigate the growth, development and general health of preterm children, focusing mainly on MLP children. Our participants consisted of a community-based sample of MLP children and a random sample of full-term controls (380 to 416 weeks’ gestation). At the time the study was designed, children born between 32 and 36 weeks’ gestation were referred to as moderate preterms. Therefore, children born at 360 to 366 weeks’ gestation were not included. LOLLIPOP was approved by our institutional medical ethics review board and written informed consent was obtained from all parents.
Participants and procedure
Participants were recruited from 13 randomly selected preventive child healthcare (PCH) centres from urban and rural areas of the Netherlands. In the PCH centres, 95% of children nationally are seen free of charge by well-child care doctors at regular intervals from birth up to 4 years of age.20 The 13 selected PCH centres monitored the physical and mental development of 45 446 children, which at the time represented 25% of all 4-year-old children monitored by PCH centres in the Netherlands. The sample size was based on estimates of the numbers needed to compile growth curves for Dutch preterm children.21
Each of the PCH centres provided a sample of preterm children and full-term controls of the same age range, based on their medical files. After the file of each second preterm child had been selected, the file of the next full-term child served as a control. Children were excluded if they had major congenital malformations or syndromes, if the gestational age could not be verified or was beyond the set range, or if families had moved between sampling and inclusion. An overview of the LOLLIPOP sampling procedure was provided previously.8 The parents of 995 MLP children agreed to participate in the long-term follow-up section of the study, as did the parents of 577 full-term children. Data on both the ASQ and Child Behaviour Checklist (CBCL) questionnaires were available for 903 MLP and 538 full-term children. For these children we were able to determine whether developmental delay co-occurred with EBP or not. In the majority of cases the mother filled out the ASQ and CBCL.
Developmental outcomes were measured using the Dutch version of the 48-month form of the ASQ, which is a validated, parent-completed screening instrument.22 ,23 We computed scores for the five developmental domains of the ASQ: fine motor, gross motor, communication, problem-solving and personal-social skills.23 Each domain consists of six questions on developmental milestones. Parents were asked whether their child had achieved (yes, 10 points), had partly achieved (sometimes, 5 points) or had not yet achieved (no, 0 points) a certain milestone. By taking the mean of the five domain scores, we computed a total score for development. Children who had an ASQ total or domain score >2 SDs below the mean of the Dutch reference group were considered as having developmental delay.22
Emotional and behavioural problems
We measured EBP using the CBCL for ages 1.5–5 years.24 ,25 This checklist has good psychometric properties and is widely used in a variety of healthcare settings and for research purposes25 in various countries including the Netherlands.26 ,27 It consists of 99 problem items and one open-ended item for recording any problems not listed on the form. Each item can be rated as either 0=not true, 1=somewhat or sometimes true, or 2=very true or often true. By summing the ratings for sets of items, we calculated internalising and externalising problems scores, and a total problems score. In accordance with the CBCL manual,24 we set the cut-off for subclinical and clinical problems at the 84th and 90th percentiles, respectively.
Data on background characteristics were collected from (1) a general parental questionnaire that was sent to the parents together with the ASQ and CBCL, and (2) medical files from PCH centres, clinical paediatricians, midwives and obstetricians. The general questionnaire consisted of information about pregnancy, delivery, gestational age, birth weight, medical conditions of the child, family composition, ethnicity, educational level of the parents and family income. Gestational age was calculated by using the last date of menstruation, and was confirmed by early ultrasound measurements in more than 95% of cases. For socioeconomic status we used a composite measure that consisted of education, occupation, and family income, as previously reported.28 Important information from the parental questionnaire, such as gestational age, was cross-checked with information from the medical files. In case of inconsistencies, we checked data against information contained in discharge letters.
First, the baseline characteristics of children with and without co-occurrence were compared, using t tests for continuous data and χ2 tests for categorical data. Second, we assessed rates of subclinical and clinical EBP among MLP and full-term children with and without developmental delay. We did this for the ASQ total score and for each ASQ domain separately. Third, using logistic regression analyses, we assessed the risk of co-occurrence in MLP compared to full-term children, for which we used a composite measure of co-occurrence, that is, an ASQ total or domain score >2 SDs below the mean and a CBCL score >84th percentile on total problems, internalising or externalising problems. In multivariate logistic regression analyses, we added gender, socioeconomic status, maternal age and ethnicity to the univariate model in three consecutive steps to determine whether the effects of MLP birth could be explained by these factors and to explore whether other independent risk factors for co-occurrence could be identified. All MLP and full-term children participating in this study were included in both the univariate and multivariate logistic regression analyses. The full-term children served as the reference group. All statistical analyses were performed in SPSS for Windows V.22.0 (Chicago, Illinois, USA).
Table 1 shows the baseline characteristics of study participants with and without co-occurrence of developmental delay and EBP. In total, 100 out of 1441 children (6.9%) had developmental delay and EBP: 77 out of 903 (8.5%) MLP children and 23 out of 538 (4.3%) full-term children. Gender, maternal age, socioeconomic status and ethnicity were significantly associated with co-occurrence (table 1). Seventy-five of the 100 children with co-occurrence were male and 25 were female.
The prevalence rates of EBP in children with developmental delay are shown in figure 1. Overall, MLP children with developmental delay (figure 1A) had higher rates of EBP than MLP children without developmental delay (figure 1B) or full-term children with or without developmental delay (figure 1C, D), particularly with respect to problems in the clinical range, that is, a CBCL score above the 90th percentile. Abnormal CBCL scores were more prevalent in MLP children with an abnormal ASQ total problems score than in full-term children, particularly regarding externalising problems: 33.8% (95% CI 23.9% to 45.4%), including 22.5% in the clinical range, versus 23.8% (95% CI 10.7% to 45.4%), including 4.8% in the clinical range, respectively. Regarding the five ASQ domains, rates of (sub)clinical EBP varied from 25% to 39% in MLP children (figure 1A). Delay in problem-solving skills frequently co-occurred with externalising problems (36.0%, 95% CI 24.1% to 49.9%; including 30.0% in the clinical range) and delay in personal-social skills with internalising problems (38.7%, 95% CI 26.4 to 52.8; including 26.5% in the clinical range). In supplementary table 1 an overview of abnormal scores on the ASQ is given for this sample of MLP and full-term children.
Table 2 shows that MLP children were more likely to have some type of developmental delay co-occurring with EBP than full-term children (unadjusted OR 2.09, 95% CI 1.29 to 3.37). Fewer girls than boys had developmental delay co-occurring with EBP, and the few girls with co-occurring problems were almost all MLP children (22 out of 25).
We included potential confounders step-by-step in multivariate logistic regression analyses (table 3). Adjustment for gender (model 1) and socioeconomic status (model 2) both mitigated the effect of moderate prematurity on co-occurrence. In the final multivariate model (model 3), the risk of co-occurrence was nearly twice as high in MLP children compared to full-term children (OR 1.86, 95% CI 1.14 to 3.03). Male gender, low family socioeconomic status and young maternal age were independent risk factors for co-occurrence in model 3.
This population-based study demonstrated that a quarter to a third of MLP children with developmental delay had co-occurring EBP at preschool age, as measured by the ASQ and CBCL. In particular, clinical-range externalising problems often co-occurred with overall developmental delay in MLP children, compared to full-term children. After we adjusted for socioeconomic status, gender, maternal age and ethnicity, the risk of developmental delay co-occurring with EBP remained significantly higher for MLP than for full-term children.
To the best of our knowledge, this is the first study to report on the co-occurrence of developmental delay and EBP in MLP children. In general, research on the co-occurrence of neurodevelopmental problems is scarce, particularly among preschool children.29 Prior studies on co-occurrence in very preterm children, born at less than 32 weeks’ gestation, showed that up to 50% had more than one developmental or behavioural disability at 5 years of age.17 ,18 In comparison, we found that 25–39% of MLP children had some form of developmental delay co-occurring with clinical or subclinical EBP, which is lower than the rates found in very preterm children but higher than those in full-term children, suggesting that there is a gradient in degree of co-occurrence by gestational age. Furthermore, we found that co-occurrence of developmental and behavioural problems was much more common among boys, which indicates a strong gender-specific influence. A comparable effect of gender has been described in an article about the prevalence of DAMP, that is, a combination of deficits in attention, motor control and perception problems.30
A number of neurophysiological and psychosocial hypotheses have been proposed to explain why developmental delay and EBP frequently co-occur.14 ,31 First, co-occurrence of developmental delay and EBP may be a consequence of aberrations in certain brain networks.31 ,32 Children with multiple brain network aberrations may be more vulnerable to exogenous or endogenous disturbances, resulting in a higher risk of both developmental and behavioural problems.31 ,32 Neuroimaging studies support this hypothesis, showing that children with both developmental and attention problems exhibit reduced functional connectivity in specific brain regions.19 The brain network aberration hypothesis could also explain why MLP children were more likely than full-term children to have developmental and behavioural problems: MLP children lack 3–8 weeks of intrauterine brain development, which increases the likelihood of disruption in brain development.33 ,34 The final months of pregnancy are characterised by the reorganisation and differentiation of newly formed neural networks, and evidence is increasing that microstructural and neural connectivity processes are disrupted in case of preterm delivery.35 ,36 Consequently, neural networks may differentiate atypically, increasing the risk of neurodevelopmental problems.35 ,36
Second, differences in rates of co-occurring developmental delay and EBP may partly be explained by psychosocial factors, such as parenting style and parent–child interactions. In families with preterm children, for example, worries about the child's health may lead to an overprotective parenting style.37 It is known that both developmental delay and EBP are associated with heightened levels of distress in families, potentially leading to negative long-term effects on child development.13 ,38 Furthermore, because of the bi-directional association between parenting style and behavioural problems in offspring,39 an enduring downward spiral may ensue, and the risk of co-occurring neurodevelopmental problems may increase.
Additionally, it is known that life adversity and stress significantly increase the risk of neurodevelopmental problems in offspring.40 This may partially explain why we identified low family socioeconomic status and young maternal age as independent risk factors for co-occurrence of developmental and behavioural problems. Maternal ethnicity was no longer associated with co-occurrence in our final analyses, which indicates that other socio-demographic factors may be collinear regarding the impact on the co-occurrence of developmental and behavioural problems in early childhood.
The findings of this study need to be considered in the context of some strengths and limitations. A major strength was the large community-based sample of MLP children, without selection of particular high-risk groups. Second, we had access to multiple sources of information. This enabled us to cross-check information from parents with retrospective medical files and discharge letters which strengthened the validity of our findings. The study also had limitations. First, among full-term children the prevalence of any type of developmental delay co-occurring with EBP was low (23 out of 538 children). In particular this limited us in interpreting the rates of EBP per developmental domain for the full-term comparison group. In addition, we used parental questionnaires to assess developmental delay and EBP instead of professional observations. Nevertheless, the added value of parental reports was apparent in the identification of psychosocial problems in PCH centres.41 Furthermore, the prevalence of neurodevelopmental problems in MLP children might be somewhat overestimated because late preterm children born at gestational age 360–366 weeks were not included in the LOLLIPOP study. Lastly, we were unable to take into account all risk factors associated with developmental delay and EBP, such as parental psychopathology and early experiences of emotional neglect.42 We recommend taking these factors into account in further research on co-occurring problems in preterm children.
We have demonstrated that 25–39% of MLP children with developmental delay also have EBP, indicating that increased awareness of EBP is warranted in MLP children with developmental delay. Further research is needed to determine whether early detection of the co-occurrence of developmental and behavioural problems results in better long-term health,12 ,43 ,44 especially for children with multiple risks, such as MLP children with low family socioeconomic status.28 Our findings indicate that much can be gained by addressing the problems of this vulnerable group.
We thank the preventive child healthcare (PCH) physicians EMJ ten Vergert, BM van der Hulst and MJ Broer van Dijk for coordinating the field work, and Dr T Brantsma-van Wulfften Palthe for correcting the English manuscript.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.