OBJECTIVES: To determine whether maternal report of infant behavioral dysregulation at 6 months is associated with a higher prevalence of behavioral concerns at 5, 14, and 21 years of age; and to assess the extent to which maternal and social factors may affect reported child behavior outcomes.
METHODS: From the Mater-University of Queensland Study of Pregnancy cohort of 7223 live singleton babies, information on dysregulation was available for 6389 children at 6 months. Of those children, behavior data were available for 4836 at 5 years by using a modified Child Behavior Checklist, 4746 at 14 years by using a full Child Behavior Checklist and Youth Self-Report, and 3558 at 21 years by using a Young Adult Self-Report. Of the youth with dysregulation data at 6 months, 2308 completed the Composite International Diagnostic Interview at 21 years. Potential confounding and mediating factors were analyzed by using logistic regression.
RESULTS: Maternal-reported behavioral dysregulation at 6 months was associated with a significantly higher prevalence of maternal-reported behavior problems at 5 and 14 years (P < .001), but not youth self-reported problems at 14 or 21 years, or Composite International Diagnostic Interview–Diagnostic and Statistical Manual diagnoses at 21 years. The strength of association between infant dysregulation and maternal-reported behaviors was greater at 5 years than at 14 years, and was substantially reduced by adjusting for maternal, social, and infant factors, especially potentially the mediating factors of maternal anxiety and depression.
CONCLUSIONS: Infant behavioral dysregulation was a risk factor for maternal-reported behavior concerns at 5 and 14 years, although was unrelated to young adult mental health.
- CBCL —
- Child Behavior Checklist
- CIDI —
- Composite International Diagnostic Interview
- MUSP —
- Mater-University of Queensland Study of Pregnancy
- YASR —
- Young Adult Self-Report
- YSR —
- Youth Self-Report
What’s Known on This Subject:
Infant behavioral dysregulation is a common concern, involving irritability, excessive crying, and problems with feeding and sleep. Previous research into its behavioral outcomes has been limited by small cohorts and short follow-up, and findings have been contradictory.
What This Study Adds:
Long-term follow-up of a large cohort showed that infant behavioral dysregulation was a risk factor for maternal-reported behavior concerns at 5 and 14 years, but was unrelated to young adult mental health outcomes.
Infant behavioral dysregulation is a common concern for parents and health care providers. The term dysregulation may be defined as a response that is poorly modulated and falls outside of the normal range. Dysregulated infants show symptoms of increased irritability, excessive crying, and problems with feeding and sleep.1–3 A limited number of studies have examined the behavioral sequelae of infant dysregulation in early childhood. The definition of behavioral dysregulation varies among studies. Some researchers4–6 used Wessel et al’s criteria7 which define excessive infant crying, whereas other researchers8,9 determined cases based on parental complaint in interview or questionnaire. One system of classification10 described regulatory problems of infancy as difficulties with self-soothing, settling to sleep unaided, or overcoming neophobia with foods, with “multiple regulatory problems” encompassing all of those. Examining associations with excessive crying in infancy, von Kries et al11 showed a higher prevalence of eating and sleeping difficulties in children with crying persisting beyond 6 months of age. In all previous studies of the outcomes of dysregulation, infants were <6 months old at the time of case selection.
Research findings are somewhat contradictory regarding the outcomes of infant behavioral dysregulation. Wolke et al4 followed infants with excessive crying, and at 8 to 10 years of age reported increased rates of hyperactivity, conduct problems, negative emotionality, difficult/demanding behavior, and eating problems, as well as impaired school performance and adaptability. Other authors, such as Desantis et al,12 Rautava et al,8 and Forsyth and Canny,9 also described increased behavioral problems in children with a history of infant dysregulation. In contrast, St James-Roberts et al5 and Elliott et al6 found no significant differences between dysregulated infants and controls at 15 months and 2 to 4 years of age, respectively. A number of maternal and infant risk factors, such as poverty or low birth weight, are associated with higher likelihood of infant dysregulation,13 whereas breastfeeding at 5 months has been shown to be a protective factor for regulatory problems of feeding at that age.14 Some studies suggest that infant dysregulation in isolation is a minor risk factor for later behavioral outcomes when compared with psychosocial risk factors in the family.15,16 Such studies also question whether the focus should be on the dysregulated infant or the mother-infant relationship. A recent meta-analysis17 reported more behavioral problems in children with a history of dysregulation when followed-up to a maximum of 10 years, especially those with multiple family risk factors. There has been no research to date on infant dysregulation as a predictor of behavioral outcomes in adolescence and adulthood. Previous individual studies have also been limited by small case numbers.
In this study, we assess whether infant dysregulation at 6 months of age is an independent risk factor for behavioral concerns at 5, 14, and 21 years. We also aimed to determine the extent to which maternal and social factors influence perceptions of dysregulation and later behavioral outcomes.
The Mater-University of Queensland Study of Pregnancy (MUSP) is a longitudinal cohort study that commenced in 1981–1983 at the Mater Misericordiae Mothers’ Hospital in Brisbane, Australia. Of the 8556 women who were invited to participate at the first pregnancy visit, 99% accepted. Women under the care of private obstetricians and those transferred from other hospitals were excluded.18 The study cohort comprises a total of 7223 live singleton babies, as described by Najman et al.19
Follow-up was by questionnaires at the first prenatal visit and at 6 months, 5 years, 14 years, and 21 years, and individual assessments at 5, 14, and 21 years. Questionnaire data regarding dysregulation were available for 6389 (88.5%) of the children at 6 months. Both dysregulation and Child Behavior Checklist (CBCL) data were available for 4836 (67%) of the children at 5 years, 4746 (66%) at 14 years, and 3558 (49.3%) youth at 21 years. There were 3078 participants with complete data at all phases. At 21 years, 2308 of the 3558 youth also completed the Composite International Diagnostic Interview (CIDI).20 In the MUSP trial, loss to follow-up at different phases has been reported as primarily related to measures of social disadvantage.19 For this study, a comparison of the characteristics of participants and nonparticipants at 5-, 14-, and 21-year follow-up is shown in Supplemental Table 6. At all follow-up phases, those not included were more likely to have birth weight of <2500 g, and had mothers who were less well educated, younger, unmarried, of lower recorded income, and more likely to have had symptoms of anxiety and depression in the neonatal phase. Those not included at 21 years were also more likely to have been born at a gestational age of <37 weeks. There was no significant difference in rates of dysregulation between those participants seen and not seen at 5-, 14-, or 21-year follow-up (P values .52, .61, and .18, respectively).
Classification of Dysregulation
At 6 months, mothers completed questions regarding the frequency of their infants having “colic,” “sleeplessness,” “feeding problems,” or “overactivity.” These items were not further defined, and were administered as described. Possible responses included “almost every day,” “a few times a week,” “a few times a month,” “rarely,” and “never.” These responses were then allocated a numerical score, from 1 to 5 in that order, and were summed to give a total score ranging from 4 to 20. Children whose scores fell within the lower 10% of total scores were classified as “dysregulated.” The 4 scored questions had a Cronbach α coefficient of 0.625. From the total group at 6 months, between 5% and 10% of infants had symptoms reported almost every day for each item individually. For each item, ∼70% of infants had symptoms reported rarely or never.
At 5 years, mothers completed a modified version of Achenbach’s CBCL,21,22 with the externalizing problems scale consisting of 11 items with a Cronbach α of 0.84, and the internalizing problems scale comprising 10 items with a Cronbach α of 0.76. The validity of this abbreviated form of the CBCL compared with the complete CBCL was assessed in 76 selected mothers of 5-year-old children who completed the full CBCL. Correlations between the scales were 0.96 for the total problem score, 0.94 for the externalizing scale, and 0.89 for the internalizing scale.23,24 At 14 years, a full CBCL and Youth Self-Report (YSR) were completed, by using a rating scale similar to the 5-year questionnaire. At 21 years of age, youth completed a Young Adult Self-Report (YASR). For all behavior questionnaires, those scoring in the highest 10% of scores were classified as having a behavior problem. At 5 years of age, restriction in the number of items used in the modified CBCL precluded the use of subscales.
At 21 years, the CIDI, apart from the Impulse Control scales, was administered and recorded on a computer program, by using Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria to assess the presence of Axis I diagnoses, with any lifetime anxiety, affective, alcohol, or nicotine disorder reported in the analysis. As the CIDI was administered face-to-face, researchers were unable to interview participants residing farther than 100 km outside the metropolitan area; thus, fewer young adults completed the CIDI than completed the YASR.
Confounding and Mediating Factors
Factors considered on the basis of the literature and availability in the database included level of maternal education (incomplete high school, complete high school, and post high school), maternal age (13–19 years, 20–35 years, and >35 years), previous births (nil and ≥1), marital status at 6 months, whether the mother’s partner at 6 months was the same as at the first pregnancy visit, and family income at 6 months (<$10 400 as the lowest quartile, or ≥$10 400). Child factors examined included birth weight (<2500 g or ≥2500 g), gestational age at delivery (<37 weeks or ≥37 weeks), Apgar score at 5 minutes (<8 or ≥8), breastfeeding duration (nil, <4 months, or ≥4 months), and admission to a NICU. Maternal anxiety or depression at birth, both measured by using the 7-item anxiety and depression subscales from the Delusions Symptoms State Inventory,25 were examined separately because of their role as potential mediators of the association between dysregulation and behavioral outcomes.
Strength of mother-infant attachment was assessed at 6 months through 6 multiple-choice items, comprising “Caring for my baby is very satisfying,” “I feel so angry that sometimes I could smack my baby,” “My baby makes me too tired,” “My baby is so good I hardly know he/she is there,” “I sometimes feel like hitting my baby,” and “I feel fed up looking after my baby all day.” The resulting mother-infant attachment score, labeled “Positive about caring for baby,” has a Cronbach α coefficient of 0.77.
The associations among social, family, and infant confounders and mediators and all behavioral outcomes are shown in Supplemental Tables 7 and 8.
SPSS 15 (SPSS Inc, Chicago, IL) was used for the analysis. For categorical variables, a χ2 test was applied, with a 2-tailed P value of <.05 used to indicate statistical significance, with relationships between infant dysregulation and summary behavior outcomes shown in Table 1.
Logistic regression was used to analyze potential confounding and mediating factors. The first 9 potential confounding factors shown in Table 2 were initially included simultaneously in multivariable models together with dysregulation (model 1). Maternal anxiety and depression at 6 months were then added as possible mediators to the effect (model 2). The strength of association was expressed as odds ratios and precision of the estimate by using 95% confidence intervals. Outcomes were examined for the total group, and separately for males and females. Study numbers shown in tables for 5-year outcomes (Table 3), 14-year outcomes (Table 4), and 21-year outcomes (Table 5) are for those with complete information on all variables.
Ethical approval was obtained from the ethics committees of both the Mater Health Services and the University of Queensland before initial data collection and at each stage of follow-up.
The association between dysregulation at 6 months and Internalizing, Externalizing, and Total problem score at 5, 14, and 21 years is shown in Table 1 for the total group and for the 3078 participants seen at each phase of follow-up. Participants with reported infant dysregulation at 6 months had poorer reported behavioral outcomes in CBCL scales at 5 and 14 years. These associations were significant for both genders. There was no significant association between infant dysregulation and youth self-reported behavioral concerns at 14 years of age on YSR or at 21 years of age on YASR. For the 3078 participants seen at each phase, findings were mildly attenuated.
The 8 specific behavior subscales of the CBCL and YSR at 14 years and the YASR at 21 years were similarly examined. For the total group at 14 years, all associations with the CBCL subscales were significant. Gender-specific CBCL subscale results at 14 years were generally similar for males and females, although Withdrawal was not significant in males, nor was Delinquent Behavior or Attention Problems in females. No significant associations were present, however, either for the total group or separately for boys or girls, for any of the YSR scales at 14 years. Apart from YASR Aggressive Behavior at 21 years in females (P = .03), there were no significant associations between infant dysregulation and the YASR scales or the CIDI-Auto scales at 21 years.
Those factors examined as potential confounders and mediators, and significantly related to dysregulation, are shown in Table 2. Table 3 shows unadjusted and adjusted odds ratios for the associations between dysregulation at 6 months and these behavioral outcomes at 5 years. Unadjusted estimates for the total study group (n = 4836) were similar, although slightly higher. For the total group and separately for males and females, all behaviors were significantly associated with dysregulation and remained significant though attenuated in model 1. With the addition of maternal anxiety and depression at 6 months (model 2) the effect size reduced, although remained significant.
For the CBCL total group at 14 years, all associations with the composite scales were significant in the unadjusted analysis and remained significant although reduced in the adjusted analysis. When examined separately by gender, the only significant associations were for Total and Externalizing scales in boys. For the 8 CBCL subscales at 14 years (Table 4), relationships for all scales were significant in the unadjusted analysis. For the total group at 14 years (n = 4746) estimates were similar, although slightly increased. In model 1 the strength of the relationships lessened, with only 5 subscales remaining significant. These were Aggressive Behavior, Attention Problems, Thought Problems, Social Problems, and Anxiety-Depression. When examined separately by gender, for boys findings were similar to the total group in model 1, although Delinquent Behavior was now also significant and the association with Attention Problems strengthened. Aggressive Behavior and Anxiety-Depression were the only subscales significant for girls. The addition of maternal anxiety and depression (model 2) reduced the magnitude of all associations. For the total group, Attention Problems was no longer significant, and Aggressive and Delinquent Behavior in boys and Anxiety-Depression in girls were also no longer significant. None of the 8 YSR subscales at 14 years were significantly related to infant dysregulation in either the unadjusted or adjusted models.
The association between infant dysregulation and 21-year YASR scales and lifetime CIDI diagnoses are shown in Table 5. Apart from YASR Aggressive Behavior in females, which was significant only in the unadjusted analysis, no significant associations were present.
Behavioral dysregulation at 6 months of age was associated with a higher prevalence of maternal-reported behavioral problems at 5 and 14 years. There was no association, however, between infant dysregulation and youth-reported problems at 14 or 21 years of age. The strength of association for maternal-reported behaviors was greater at 5 years than at 14 years, and was substantially attenuated by adjusting for social confounding and mediating factors.
Previous studies examining the effects of infant dysregulation on later behavior have been somewhat variable in their findings and have been limited by only short- to medium-term follow-up. Wolke et al,4 Desantis et al,12 Rautava et al,8 and Forsyth and Canny9 all demonstrated increased behavioral concerns in children with a history of infant dysregulation, when followed to a maximum of 10 years of age. In particular, Forsyth and Canny9 noted that children with a history of dysregulation, followed to 3 years of age, were more likely to be viewed by their mothers as vulnerable and with behavioral problems, but had no actual measurable difference in attention span, behavior regulation, or sociability. St James-Roberts et al5 on the other hand, reported that infant dysregulation or persistent crying were not predictive of either maternal or researcher measures of infant behavior problems, temperament, or negativity at 15 months of age. Elliot et al6 found no differences between infants with and without “excessive crying,” when followed to 2 to 4 years of age. Our results, from long-term follow-up over 21 years, support the view that infant dysregulation is a risk factor for maternal-reported behavioral concerns in childhood and adolescence, although not adolescent-reported behavioral concerns or young adult mental health.
Some research15,16 has suggested that infant dysregulation is a minor risk factor for the development of later behavioral problems when compared with psychosocial factors. Although our results did show attenuation of the strength of association between infant dysregulation and later maternal-reported behavioral problems when adjusted for maternal personal and social factors, the association still remained significant.
The incongruence between maternal and youth reports is not entirely unexpected. There is generally an inherent discrepancy between self-report and proxy or surrogate report.26 A meta-analysis by Achenbach et al27 found a low correlation, in the range of 0.2 to 0.3, among reports from different types of informants. It is unclear which of the CBCL and YSR is more valid or accurate, although one could reasonably expect self-report to be the more valid indicator of child mental health, as the child is likely to be more knowledgeable about his or her internal feeling state than is the mother.28 Furthermore, as a surrogate reporter, the mother is able to report on only a subset of the child’s behaviors: those that she observes. A study by Najman et al23 evaluating early childhood screening for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition mental health problems at 21 years in the MUSP cohort, found that the YSR at 14 years had a higher positive predictive value than the CBCL at either 5 or 14 years.
A number of possible explanations exist for the difference observed in maternal- and youth-reported behaviors at 14 years. First, it may be that infant dysregulation is an early marker for behavioral problems in childhood, which improve by adolescence and young adulthood. Second, it is possible that there is a persistent distortion of the early mother-child relationship, without other relationships or the child’s self-perception being affected. Finally, it may reflect a bias in maternal reporting of infant dysregulation and/or later behavioral problems. Although a broad range of potential confounders was examined, it is possible that residual confounding exists.
The lack of association of infant behavioral dysregulation with adult outcomes may be explained by the theory of a developmental cascade, by which differences in behavior or cognition maximally affect the next most proximate phase of development, which in turn affects the following phase and so on.29 The attenuation of this effect with time is not unexpected, with research by McCall30 demonstrating that the closer the measurement points, the more likely it is that associations will be found.
A number of authors31–33 have researched the influence of temperament on behavioral adjustment. Temperament refers to the style of behavior, rather than the content. There is some commonality between temperament and behavior literature, although this study does not seek to define participants’ temperament, but instead examines a particular set of infant behaviors at 6 months of age.
Subject attrition is a common source of potential bias in any longitudinal cohort study. In this study, 88.5% of the original cohort completed maternal questionnaires for enrollment at 6 months of age and, of those, 76% were followed-up at 5 years with the CBCL, 74% at 14 years with both the CBCL and YSR, and at 21 years, 56% completed the YASR and 33.4% completed the CIDI. Those participants attending inconsistently or lost to follow-up had higher levels of social risk factors34 and were therefore more likely to have poor behavioral outcomes and stronger relationships with dysregulation, consistent with the findings comparing the full sample to the 3078 seen on every occasion. The inconsistency between the CBCL and YSR at 14 years is likely to be unchanged. Research has shown that although a large nonparticipant rate may affect prevalence estimates, it does not necessarily bias the associations between the variables.35
Another possible limitation of this study is the allocation of cases to dysregulated and nondysregulated groups. Criteria for case selection have varied in different studies.4–6 Previous studies on behavioral outcomes of infantile dysregulation have selected cases before 6 months. Wessel et al7 found that most “fussy” infants had resolution of symptoms by the end of the second month of life, with only around 20% continuing to be classified as fussy after 3 months of age. By selecting our cases of dysregulation at 6 months of age, it is possible that we have selected more severe cases of dysregulation. Indeed, Hemmi et al17 reported that persistent regulatory problems in infancy, lasting beyond the first 3 months of life, were more strongly associated with childhood behavioral problems.
Infant behavioral dysregulation was a risk factor for maternal-reported behavioral concerns at 5 and 14 years of age, although had no significant impact on young adult mental health. Although maternal, social, and infant factors may partly explain associations at 5 and 14 years, maternal concerns are nevertheless real to the mothers. Where able, clinicians should offer support, education, and reassurance to parents of dysregulated infants, being mindful of possible contributing factors. As a multifactorial issue, the support of a multidisciplinary team may be beneficial. By facilitating early referral to appropriate professionals, such as public health nurses, family therapists, psychologists, and social workers, clinicians may aim to improve not only behavioral outcomes in childhood and adolescence, but also parents’ perceptions of their children and the needs of the parents themselves.
The authors acknowledge the mothers and children who have participated in the MUSP, as well as the MUSP research and data collection teams.
- Accepted July 2, 2012.
- Address correspondence to Rebecca Hyde, BSc, MBBS, FRACP, Mater Children’s Hospital, Raymond Terrace, South Brisbane, Queensland, Australia 4101. E-mail:
Dr Hyde developed the study aims, drafted the literature review, and wrote the first draft of the manuscript; Drs Hyde and O’Callaghan were responsible for study design, data analyses, and interpretation of findings; Drs O'Callaghan, Najman, Williams, and Bor are responsible for the conceptual development and continued management of the Mater-University of Queensland Study of Pregnancy and its outcomes, and take responsibility for the integrity and accuracy of the data analysis; and Drs Hyde, O'Callaghan, and Najman edited drafts of the paper and contributed to the discussion and conclusion. All authors contributed to and approved the final version of the paper.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: The core Mater-University of Queensland Study of Pregnancy was funded by a grant from the National Health and Medical Research Council of Australia. Funding for the current research project was provided by the Golden Casket Foundation. The views in this article are those of the authors and do not necessarily reflect the views of any funding body.
- von Hofacker N,
- Lehmkuhl U,
- Resch F,
- Papousek M,
- Barth R,
- Jacubeit T
- Wolke D,
- Rizzo P,
- Woods S
- Wessel MA,
- Cobb JC,
- Jackson EB,
- Harris GS Jr,
- Detwiler AC
- Rautava P,
- Lehtonen L,
- Helenius H,
- Sillanpää M
- Forsyth BW,
- Canny PF
- Emde RN,
- Egger H,
- Fenichel E,
- Guedeney A,
- Wise BK,
- Wright HH
- Hemmi MH,
- Wolke D,
- Schneider S
- Keeping JD,
- Najman JM,
- Morrison J,
- Western JS,
- Andersen MJ,
- Williams GM
- Najman JM,
- Bor W,
- O’Callaghan M,
- Williams GM,
- Aird R,
- Shuttlewood G
- ↵World Health Organization. Composite International Diagnostic Interview (CIDI): version 2.1. Geneva, Switzerland: WHO; 1992
- Achenbach T
- Bedford A,
- Foulds G
- Lawlor D,
- Mishra G
- Bornstein MH,
- Hahn C-S,
- Bell C,
- et al.,
- ALSPAC Study Team
- McCall R
- Prior M,
- Sanson A,
- Smart D,
- Oberklaid F
- Kagan J
- Martikainen P,
- Laaksonen M,
- Piha K,
- Lallukka T
- Copyright © 2012 by the American Academy of Pediatrics