Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
August 2004

Parent Identification of Early Emerging Child Behavior ProblemsPredictors of Sharing Parental Concern With Health Providers

Author Affiliations

From the School of Epidemiology and Public Health, School of Medicine (Ms Ellingson and Dr Briggs-Gowan), and the Department of Psychology (Dr Briggs-Gowan), Yale University, New Haven, Conn; the Department of Psychology, University of Massachusetts, Boston (Dr Carter); and the Department of Psychiatry, Case Western Reserve University, Cleveland, Ohio (Dr Horwitz).


Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004

Arch Pediatr Adolesc Med. 2004;158(8):766-772. doi:10.1001/archpedi.158.8.766

Objectives  To better understand the predictors of parental discussions with pediatric care providers (pediatricians, psychologists/psychiatrists, social workers, early intervention providers, or other medical specialists) regarding early child behavior problems and to suggest strategies for eliciting early identification from parents in health care settings.

Design  A cross-sectional survey of parents of children from a representative healthy birth cohort. The survey included the Infant-Toddler Social Emotional Assessment, measurement of parental worry regarding problematic behavior, and demographic factors.

Setting  Fifteen urban and suburban towns in the northeastern United States.

Participants  The study sample consisted of all parents of 11- to 39-month-olds (n = 269) who exceeded the 90th percentile on 1 or more Infant-Toddler Social Emotional Assessment problem domain scores (representing elevated problematic behavior symptoms) from an original sample of 1278.

Results  Few parents (17.7%) who reported elevated problematic behavior spoke to a provider about such problems. In adjusted models, speaking to a provider was associated with reported worry about behavior (odds ratio [OR], 3.47 [95% confidence interval (CI), 1.74-6.92]) and with low reported child social-emotional competence (OR, 2.68 [95% CI, 1.23-5.84]). In adjusted models, worry was most likely among parents who reported low child competence (OR, 2.18 [95% CI, 1.07-4.22]) and disruption in family routines attributed to the child's behavior (OR, 2.38 [95% CI, 1.31-4.33]).

Conclusions  Parental worry is a robust predictor of help seeking among parents of children with behavioral problems. Further, lags in social competence contribute to both parental worry and help seeking. These findings, in conjunction with previous evidence that child behavior problems amenable to early intervention are often unidentified, suggest that systematic inquiry by health care providers about parental concerns is important in the identification of early emerging behavioral health problems.

Expressed parental concern regarding a child's early emerging behavioral or developmental problems has recently gained attention as an important component in the identification and referral of such problems by pediatric care providers (pediatricians, psychologists/psychiatrists, social workers, early intervention providers, or other medical specialists).1 Current studies reveal that 7% to 24% of 2- to 3-year-old children have social-emotional or behavioral problems.2 However, rates of intervention services for young children with behavioral health needs remain low35 despite the existence of effective early interventions and the concern of federal policy makers.6,7

Children of preschool age are dependent on caregivers (most often parents) for both recognition of behavioral problems and subsequent help seeking.8,9 Those younger than 5 years are especially sensitive to the role of parents in detection and service seeking because they are not in contact with formal school systems, considered the "de facto" mental health system for older children.10 Difficulties may exist in the detection and diagnosis of behavioral health problems in young children because parents may have trouble distinguishing normal from problematic behavior. Thus, examining factors that trigger parental concern and how such concern affects access to care may inform early identification efforts.

Several models of health access obtained from the broader health services literature justify the exploration of parental concern in the help-seeking process. Common to 3 seminal help-seeking models is the importance of need, sociodemographic access variables, and help seekers' perceptions.1113 Demographic factors have been shown to play a role in parental help seeking for child behavioral health services for older children,14,15 as have parental burden and stress.9,16,17 However, the role of parental attitudes and perceptions in help seeking for young children has been largely unexplored.

Because earlier findings showed that only 7.5% of young children with reported behavioral problems received any care and that parental worry was associated with parents' thinking about or talking to a provider for health services in general,18 we were specifically interested in the role and predictors of worry with respect to parental help seeking for behavioral health problems. In this article, we extend prior work by examining the role of worry with respect to parental help seeking for behavioral health problems, as well as the predictors of parental worry among children at elevated risk for behavioral health concerns.


The original healthy birth cohort sample was identified from birth records at the State of Connecticut Department of Public Health, Hartford.2 Children from the New Haven–Meriden standard metropolitan statistical area, Connecticut (1990 census), born at Yale–New Haven Hospital between July 1995 and September 1997, were eligible. Recruiting of participants, baseline data collection, and all subsequent data collection efforts were approved by Yale University's human investigation committee.

The study sample (Figure 1) included children reported to be at or above the 90th percentile (within the representative study population) on the internalizing, externalizing, and/or dysregulation domains of the 166-item parent-reported Infant-Toddler Social Emotional Assessment (ITSEA).19 The externalizing domain measures aggression, defiance, and impulsivity. The internalizing domain addresses depression, withdrawal, anxiety, and inhibition to novelty. Dysregulation measures problems in the areas of sleep, eating, and sensory sensitivities. Finally, ITSEA social competence assesses attention skills, compliance, empathy, imitation, and peer interactions. The ITSEA has acceptable internal consistency, test-retest reliability, and interrater reliability. Associations between the ITSEA and independent evaluator ratings and parental ratings of child behavior problems, temperament, and parental distress support its validity.19

Image not available

Generation of study sample (n = 269) from original sample (n = 1278).

Of the study sample of all children exceeding the 90th percentile cutpoint on at least 1 ITSEA problem domain (n = 269), 129 children (48.0%) exceeded the cutpoint on the internalizing domain, 127 (47.4%) on the externalizing domain, and 125 (46.5%) on the dysregulation domain. Eighty-three children (approximately 31.0%) in the study sample exceeded the 90th percentile on 2 ITSEA problem domains, and 29 (10.8%) had high scores on all 3 domains. Additionally, 43 children (16.0%) in the study sample scored below the 10th percentile for social-emotional competence on the ITSEA. The study sample included 129 boys (48.0%), 135 children (50.2%) who were ethnic minorities, 44 children (16.2%) living below the poverty line, 84 children (31.3%) living in poverty or borderline poverty (up to 185% of the poverty line), and 49 children (18.1%) who lived in households with no working parent. Respondents included 252 mothers (93.7%) and 16 fathers (6.0%) (hence, respondents are referred to in this article as "parents"), 80 single parents (29.7%), and 99 (37.1%) with a high school education or less.

The study sample (n = 269) was compared with the remainder of the original sample (those without a high problem score on any ITSEA problem domain, n = 1006). Bivariate analyses reveal that children in the study sample were significantly more likely than the rest of the original sample to be ethnic minorities, live in families with no working parent, live in poverty or borderline poverty, and have a parent with lower education (Table 1). Study sample children also were more likely to have scores below the 10th percentile on ITSEA social-emotional competence and a score below the 10th percentile in age-related vocabulary skills (Table 1). Study sample parents were significantly less likely to have education exceeding high school and more likely to score high on the depression/anxiety indices of the Center for Epidemiological Studies Depression Scale,22 report difficulty with parenting on the Parenting Stress Index,23 and report high levels of family burden on the Family Life Impairment Scale.2 Finally, study sample parents were more likely to worry about their child's behavior. There were no detectable differences between the original and selected study sample with regard to child's age, respondent's age, or severe child physical health problems.

Table 1. 
Image not available
Description of Study Sample With Elevated Symptoms in at Least 1 Infant-Toddler Social Emotional Assessment (ITSEA) Domain vs the Rest of the Original Sample*

Parents were mailed a letter describing the study, followed 1 week later by a questionnaire and children's book. Staff members subsequently telephoned parents to address questions or concerns and to encourage participation in the study. When appropriate, parents were remailed questionnaires and/or visited in person to offer assistance. Informed consent procedures were followed, and parents received $25.


The ITSEA was used to define the sample and was described earlier. Other instruments and measures used in this analysis, including measures of child characteristics, parental factors, parent-child stress, and outcome variables, are summarized in Table 2.

Table 2. 
Image not available
Study Measures

Bivariate analyses were used followed by multivariate logistic regression analyses (using stepwise backward elimination models) to control for variance shared among variables and to test for interactions. Owing to the small sample (n = 269) and low baseline prevalence of outcome variables, multivariate models only included those variables exhibiting significant bivariate associations with the outcome of interest. Variables were then individually removed from the full model. Variables were retained if their removal resulted in a significant deviance from the full (or previously reduced) model. Additionally, t tests were conducted to test for differences in raw ITSEA scores among children of parents who spoke to a provider (vs those who did not) as well as for parents who worried (vs those who did not). To run these t tests, we created a composite ITSEA score by taking the mean of raw ITSEA scores on all 3 domains. All statistical analyses were conducted using SAS software version 8.0 (SAS Institute Inc, Cary, NC).


Only 47 parents (17.7%) in our sample with young children exhibiting elevated problematic behavior reported having spoken to a provider about their child's behavioral health. Of parents who had spoken to a provider, 41 (87.2%) spoke with a pediatrician.

Bivariate analyses (Table 3) identified 4 factors as significantly associated with speaking to a provider about a child's behavior: parental worry, low social competence, elevated parental depression/anxiety, and disruption of family routines because of the child's behavior. Notably, sociodemographic factors, such as ethnic minority status, living in poverty/borderline poverty, parental unemployment, and parental education, were not significantly associated with the likelihood of speaking to a provider. Several characteristics specific to the child, including the child's sex, age, wide-ranging behavioral problems (scores above the 90th percentile in all ITSEA domains), and low vocabulary skills, also showed no association with speaking to a provider.

Table 3. 
Image not available
Bivariate Analyses Within a Study Sample of 269 for Speaking to a Provider and Parental Worry Outcomes*

The multivariate logistic regression analysis (Table 4) used to predict the likelihood of a parent speaking to a provider regarding early behavioral problems showed that only parental worry (odds ratio [OR], 3.47 [95% confidence interval (CI), 1.74-6.92]) and low social-emotional competence (OR, 2.68 [95% CI, 1.23-5.84]) remained significant predictors of speaking to a provider in our final model. Interruption of family life because of the child's behavior was eliminated from the model via backward stepwise elimination. However, elevated parental depression/anxiety remained in the final model because removing it changed the fit of the model significantly. There were no statistically significant interactions. Finally, a t test conducted among parents who spoke to a provider (vs those parents who did not) revealed that there was no difference in raw ITSEA scores between the 2 groups (P = .48), indicating no difference in overall symptoms between the 2 groups.

Table 4. 
Image not available
Adjusted Multivariate Logistic Regression Analysis Modeling the Probability of Speaking to a Pediatric Care Provider About a Child's Behavior (n = 269)*

Given the importance of parental worry as a predictor of accessing services, characteristics associated with worry were investigated. Parents who reported being worried about their child's behavior (n = 101; 37.8%) were compared with parents who did not worry in a series of analyses. Bivariate analyses (Table 3) indicated that parental worry was associated with child sex (male), living in poverty/borderline poverty, low parent-reported social competence, and difficult parent-child relations. Further, parents worried more about children's wide-ranging behavior problems (high ITSEA scores in all domains) and about children who contributed to high levels of disruption in family routines on the Family Life Impairment Scale.

All significant predictors of parental worry about a child's behavior from bivariate analyses were included in multivariate regression analyses to adjust for shared variance (Table 5). Using stepwise backward elimination, poverty/borderline poverty and wide-ranging behavioral problems (high ITSEA scores in all domains) were eliminated from the final model. All remaining variables were significant, including low social competence (OR, 2.18 [95% CI, 1.07-4.22]), family life impairment due to the child's behavioral problem (OR, 2.38 [95% CI, 1.31-4.33]), male child (OR, 1.87 [95% CI, 1.10-3.19]), and elevated parental depression/anxiety (OR, 1.81 [95% CI, 1.03-3.17]).

Table 5. 
Image not available
Multivariate Logistic Regression Analyses Modeling Parental Worry About a Child's Behavior*

Finally, a t test conducted among parents who worried about their child's behavioral health (vs those who did not) revealed that there was a significant difference in raw ITSEA scores between the 2 groups (high worry, mean ± SD, 0.831 ± 0.182; low worry, mean ± SD, 0.747 ± 0.162; P = .001). Thus, among our sample of children with elevated symptom reporting, parents who worried about their children were more likely to have children with significantly higher ITSEA scores than those parents who did not worry.


Our results strongly suggest that parental worry is a primary trigger in the initiation of contact with a provider for children with behavior problems, even after controlling for low age-related social competence and elevated parental depression/anxiety. Our findings confirm assertions that parental worry is an indication that a child is failing to meet developmental or behavioral expectations within the family context and may be a key determinant of whether parents access any services for behavioral health problems.

Additionally, we found that having elevated scores on multiple domains of the ITSEA did not significantly predict either worry or speaking to a provider. However, when we ran t tests on the composite raw ITSEA scores by the outcomes of worry and speaking to a provider, we found that parents who worried tended to have children with significantly higher overall ITSEA scores. However, parents who spoke to providers about behavior problems did not have significantly higher raw ITSEA scores than those parents who did not. This underscores our conclusion that parental worry is an important correlate of behavioral problems in young children and suggests that there are parents who report high levels of problems and worry yet do not speak to a provider regarding these problems. There may be several reasons why parents choose not to speak to a health care provider about their child's behavioral problems. Parents may not consider problematic behavior atypical or may be reluctant to acknowledge their concerns because of the stigma of mental health problems, fear of being blamed for their child's behavior, or skepticism about the value and willingness of providers to treat such problems effectively.25

When parental characteristics associated with worry were examined, we found that disruption of family routines and low social competence were the most robust predictors of worry in adjusted models. However, only social competence also predicted seeking services, suggesting that if parents had better norms for judging age-appropriate social skills, they might be more likely to initiate discussions about problematic behaviors with their pediatricians. Thus, social-emotional competence may be one avenue through which to begin a dialogue about behavioral health concerns. In addition, this finding emphasizes the importance of educating parents so that they become better able to distinguish normative infant-toddler behavior from atypical behavior, a task that is particularly challenging in early childhood because many symptomatic behaviors have normative correlates. Atypical behaviors are often distinguishable from normal behaviors because of heightened frequency or intensity.26 Continued research efforts to clarify the boundaries between normal and atypical behaviors in early childhood and parent education about development and psychopathologic features are central to the goal of promoting service use for mental health problems in young children.

Despite the fact that the sample was drawn from a representative birth cohort, the size of the sample of children with high levels of behavioral difficulties limited our ability to evaluate a large number of variables and to detect interactions. Additionally, the cross-sectional nature of the data does not permit the establishment of causality. Continuing this line of research with longitudinal data is an important future contribution.


Given the discrepancy that currently exists between the need for and use of services for children's behavioral health problems, our findings suggest that systematically soliciting, through interviewing or screening, parental concerns about their children's behavioral health problems may be a promising strategy to increase access to services. This study underscores the important role of parental worry in help seeking for behavioral services and the continuing importance of future research exploring parental attitudes as key determinants of service use. Educating pediatricians and other providers who come into contact with young children to elicit parental concerns may be an efficient way to identify early behavioral health needs. Additionally, encouraging parents to act on concerns about their children's social-emotional and behavioral health has the potential to promote appropriate service use for young children with early behavioral problems.

However, given the complex interplay between child difficulties and parental concern in the help-seeking process, it may be beneficial to also use screening measures that rely on parental reports. For young children, screening tools such as the Ages & Stages Questionnaires: Social-Emotional27 and the Brief Infant-Toddler Social and Emotional Assessment,28 both of which address behavioral health problems and social-emotional competencies, can be administered in a health care setting. Such screening strategies have proved to be accurate and cost-effective for several reasons.29 First, these screens eliminate the need for obtaining children's cooperation, provide a more thorough and extensive sampling of behaviors than is typically obtained with direct observation of the child, and usually have flexible administration methods (in-person interviews, telephone interviews, take-home questionnaires, completed in preparation for a second appointment, self-administered in waiting rooms, etc). The number of screens available for early childhood has grown in recent years, and these should, if possible, be used routinely in pediatric practices.

What This Study Adds

Behavioral health problems in young children may persist into later childhood, adolescence, and adulthood, exacting a great toll on individuals and society. Despite the existence of effective early interventions and expressed concern from both the medical community and policy makers, recognition and treatment of young children with behavioral problems continue to occur at very low rates. This study attempts to better understand the predictors of parental discussions with pediatric health care providers regarding early emerging behavioral problems, and in particular the influence of parental worry. Understanding the role of parental worry in help seeking has the potential to guide providers in their systematic inquiry of parents about behavioral concerns.

Back to top
Article Information

Correspondence: Katherine D. Ellingson, BS, Yale University School of Epidemiology and Public Health, 60 College St, PO Box 208034, New Haven, CT 06520-8034 (

Accepted for publication April 7, 2004.

Glascoe  FPMacLean  WEStone  WL The importance of parents' concerns about their child's behavior. Clin Pediatr (Phila). 1991;308- 11
Briggs-Gowan  MCarter  ASkuban  EHorwitz  S Prevalence of social-emotional and behavior problems in a community sample of 1- and 2-year-old children. J Am Acad Child Adolesc Psychiatry. 2001;40811- 819
Jenkins  SBax  MHart  H Behaviour problems in pre-school children. J Child Psychol Psychiatry. 1980;215- 17
Cohen  PHesselbar  C The role of caregiver strain in determining children's mental health service utilization patterns. Diss Abstr Int. 2001;615794
Cause  AMDomenech-Rodriguez  MParadise  M  et al.  Cultural and contextual influences in mental health help seeking: a focus on ethnic minority youth. J Consult Clin Psychol. 2002;7044- 55
The National Advisory Mental Health Council Workgroup on Child and Adolescent Mental Health Intervention Development and Deployment, Blueprint for Change: Research on Child and Adolescent Mental Health.  Washington, DC National Institute of Mental Health2001;
Satcher  D Executive summary: a report of the Surgeon General on mental health. Public Health Rep. 2000;11589- 101
Costello  EJaniszewski  S How can epidemiology improve mental health services for children and adolescents? J Am Acad Child Adolesc Psychiatry. 1993;321106- 1113discussion 1114-1117
Angold  AMesser  SCStangl  DFarmer  EMCostello  EJBurns  BJ Perceived parental/family burden and service use for child and adolescent psychiatric disorders. Am J Public Health. 1998;8875- 80
Burns  BJCostello  EJTweed  DStangl  DFarmer  EM Data watch: children's mental health services across service sectors. Health Aff. 1995;14147- 159
Andersen  RNewman  JF Societal and individual determinants of medical care utilization in the United States. Milbank Mem Fund Q Health Soc. 1973;5195- 124
Rosenstock  IM Why people use health services. Milbank Mem Fund Q. 1966;4494- 106
Pescosolido  BABrooks Gardner  CLubell  KM How people get into mental health services: stories of choice, coercion and "muddling through" from "first-timers." Soc Sci Med. 1998;46275- 286
Weisz  JRSuwanlert  SChaiyasit  WWeiss  BJackson  EW Adult attitudes toward over- and undercontrolled child problems: urban and rural parents and teachers from Thailand and the United States. J Child Psychol Psychiatry. 1991;32645- 654
Lin  KInui  TSKleinman  AMWomack  WM Sociocultural determinants of the help-seeking behavior of patients with mental illness. J Nerv Ment Dis. 1982;17078- 85
Brennon  AMHelfinger  CA Distinguishing caregiver strain from psychological distress: modeling the relationships among child, family, and caregiver variables. J Child Fam Stud. 2001;10405- 418Article
Verhulst  FCvan der Ende  J Factors associated with child mental health service use in the community. J Am Acad Child Adolesc Psychiatry. 1997;36901- 909
Horwitz  SMGary  LCBriggs-Gowan  MJCarter  AS Do needs drive services use in young children? Pediatrics. 2003;112 (pt 6) 1373- 1378
Carter  ASBriggs-Gowan  MJones  SMLittle  T The Infant-Toddler Social and Emotional Assessment (ITSEA): factor structure, reliability, and validity. J Abnorm Child Psychol. 2003;31495- 514
Fenson  LPethick  SRenda  CCox  JLDale  PSReznick  JS Short-form versions of the MacArthur Communicative Development Inventories. Appl Psycholinguist. 2000;2195- 116Article
Radloff  L The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1997;1385- 401Article
Beck  AEpstein  NBrown  GSteer  R An inventory for measuring clinical anxiety: psychometric properties. J Consult Clin Psychol. 1988;56893- 897
Abidin  RR Parenting Stress Index Short Form-Test Manual.  Charlottesville, Va Pediatric Psychology Press1990;
Fenson  LDale  PReznick  J  et al.  The MacArthur Communicative Development Inventories: User's Guide and Technical Manual.  San Diego, Calif Singular Publishing Group1993;
Glascoe  FPDworkin  PH Obstacles to effective developmental surveillance: errors in clinical reasoning. J Dev Behav Pediatr. 1993;14344- 349
Eggers  H The reliability of the preschool age psychiatric assessment (PAPA).  Paper presented at: 50th Annual Meeting of the American Academy of Child and Adolescent Psychiatry October 15, 2003 Miami, Fla
Squires  JBricker  DTwombly  E Ages & Stages Questionnaires: Social-Emotional: A Parent-Completed, Child-Monitoring System for Social-Emotional Behaviors.  Baltimore, Md Paul H. Brookes Publishing Co Inc2002;
Briggs-Gowan  MJCarter  ASIrwin  JRWachtel  KCicchetti  DV The Brief Infant-Toddler Social and Emotional Assessment: screening for social-emotional problems and delays in competence. J Pediatr Psychol. 2004;29143- 155
Glascoe  FPFoster  EMWolraich  ML Parental concern has proven to be key in early identification of developmental problems in children. Pediatrics. 1997;99830- 837