Objective. To determine whether different factors predicted parental recognition of mental health problems in children and use of services for those problems and to determine whether mental health problems elicit service use in the same way as physical health or developmental problems.
Methods. Survey of the parents of an age- and sex-stratified random sample of children who were born in 1 Connecticut hospital selected from state birth records.
Results. Logistic regression results demonstrate that a number of child and family factors are related to problem recognition, defined as thinking about seeking services and/or talking to a professional about a problem. However, factors related to the third outcome of interest—service use—are more limited and include only physical health and developmental problems, parental worry about language, and family conflict. Furthermore, mental health problems were not discussed with professionals, and they were not associated with receipt of services at the same rates as physical health or developmental problems.
Conclusions. The correlates of problem recognition (defined as thinking about seeking services or actually speaking to a professional about a problem) differ from those of service use, and, whereas mental health, physical health, and developmental problems all affect problem recognition, mental health problems do not result in service use. Furthermore, rates of service use vary dramatically across the 3 categories. These results suggest that needs alone do not drive service use. Rather, other factors may influence which needs are met.
Estimates of behavioral health problems for children indicate that 20% of youths experience such problems, but fewer than 25% of those with problems receive any services.1–4 Many explanations have been put forth for this lack of service, including that children do not seek services for themselves but rather must rely on significant others in their environments to recognize these problems and seek services.5,6 The literature on caregiver behavioral problem recognition and help seeking, particularly among preschool children, is limited. However, parental attitudes and beliefs, perceived need in the child, and family stressors predict problem recognition.7–9 Similar characteristics are cited in the extant literature on predictors of mental health services use among school-age children, including sociodemographic variables such as education, income, insurance coverage, sex, and race/ethnicity; psychosocial factors such as stressful life events, family burden, and social support; and parental and child psychopathology.1,3,5,6,10–18
Most models of help seeking postulate perceived need and susceptibility to illness as important predictors of service use.19–21 However, in the few studies that have examined the impact of perceived need on help seeking for mental health issues, there does not seem to be a direct relationship between perception of need and service use. A study by Horwitz et al,22 which examined correlates of parents’ attitudes toward discussing psychosocial problems versus actual discussions of such problems, suggested that there are considerable discrepancies between expressed parental attitudes and beliefs about psychosocial problems and actual actions about such issues. For example, although most parents believe that it is hypothetically appropriate to discuss behavioral/emotional issues with their pediatrician, many do not actually do so when problems exist.1,22,23 Similarly, data from the Greater Smoky Mountain Study suggested that need must be coupled with parental perceived burden associated with the child’s behavioral and emotional problems to result in service use.5
Given that seeking care for children’s behavioral health issues involves both recognizing the problem and overcoming barriers to seeking service and that information on these issues is sparse, we questioned whether factors related to these 2 distinct aspects of the care-seeking process differed when measured within the same respondents. We also questioned whether mental health problems elicit service use in the same way that physical health or developmental problems might, because mental health problems may not be as readily recognized, particularly in early childhood, when parents may be reluctant to acknowledge such problems or may interpret such problems as age typical. Furthermore, mental health services for infants and toddlers are scarce and often reside in the specialty sector, whereas physical health services are routinely available and developmental services are covered under Part H of the Individuals With Disabilities Education Act Amendments of 1997 (P.L. 105-17).20,24 We investigated these questions by comparing variables associated with parental reports of problem recognition with those associated with reported receipt of services in a community-based sample of children between 11 and 39 months of age.
An age- and sex-stratified random sample (N = 8404) of children who were born at Yale-New Haven Hospital between July 1995 and September 1997 and lived in the 15 towns/cities that composed the New Haven Meriden Standard Metropolitan Statistical Area from the 1990 Census were selected from birth records at the State of Connecticut Department of Public Health. The original sample was designed to refine a new measure of social-emotional adjustment in a developmentally healthy sample of children. For this reason, excluded from the sample were 675 children who were born prematurely (<36 weeks) or determined likely to have significant developmental delays on the basis of birth record data (ie, birth weight <2200 g; 1- and 5-minute Apgar scores <5; chromosomal anomalies [eg, Down syndrome]; and severe anoxia or need for resuscitation at birth). Children who were deceased before sampling (n = 14) or adopted (n = 4) were excluded. One child per mother was randomly selected, resulting in 277 sibling exclusions. An investigator’s child was also excluded. A total of 1788 families were randomly selected from the 7433 eligible subjects that remained after applying the exclusion criteria. These families did not differ significantly from original subjects (N = 8404) in terms of maternal education or child race. Not surprising, differences were observed for infant birth weight and gestational age (t = 6.60 and 6.49, P < .01, respectively), as well as maternal age (meanselected = 29.6, standard deviation = 6.1; meanoriginal = 28.8, standard deviation = 6.2; t = 5.02, P < .01).
Among those sampled, subjects were excluded when 1) neither parent spoke English well enough to participate in a self-report or interview format (n = 50), 2) parents had lost custody (n = 17), and 3) families had moved out of state (n = 116). Excluded families (n = 183) did not differ from remaining families (N = 1605) in terms of maternal education, child race, 1-minute Apgar scores, gestational age, birth weight, or years at the birth address. Parental age and 5-minute Apgar scores were slightly higher in the eligible sample (t = 2.04–4.03, P < .05).7
Of the 1605 eligible subjects, 1278 (79.8%) participated in the baseline data collection. Nonparticipants tended to be younger, less educated, and minority and to have lower birth weight infants. Minority participants were representative of minority nonparticipants on all birth status and sociodemographic variables. Furthermore, birth weight and maternal age were not associated with probability of participation once minority status and education level were taken into account.
From June to September 1998, 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 and/or concerns and encourage participation. Failure to participate after 1 month resulted in a second mailing followed by in-person contacts to offer assistance needed to facilitate participation (eg, interviews, babysitting). All procedures and the informed consent process were approved by the Yale School of Medicine’s Human Investigation Committee. Parents received $25 for participating.
Birth status variables were obtained from birth records provided by the State of Connecticut Department of Public Health. Birth status variables included infant birth weight, gestational age, 1- and 5-minute Apgar scores, parental age, maternal education, and years at the birth address. Additional variables reflecting birth status were used as exclusion criteria in sample selection.
Participants answered several questions about their family sociodemographic status, including the target child’s sex, age, ethnicity, and birth order; maternal age; parental education; marital status; before-tax household income; and use of child care. Parents rated their child’s current physical health on a 5-point scale from “poor” health to “excellent” health. Parents reported whether the child had ever received a diagnosis of a condition or disease, such as lead poisoning or leukemia. In a separate question, parents indicated whether the child had ever received a diagnosis of a developmental delay or problem. Parents (respondents) also rated their own current physical health on a 5-point scale from “poor” health to “excellent” health.
Parents completed the Infant Toddler Social and Emotional Assessment (ITSEA) a new adult-report instrument designed to assess the social-emotional functioning of infants and toddlers. The ITSEA has 4 broad domains—Internalizing, Externalizing, Dysregulation, and Competence—and a separate Maladaptive scale. Items are rated on a 3-point scale. The α coefficients for the broad categories of the ITSEA were very good to excellent (Cronbach’s α: 0.78–0.90).25,26 Parents rated three 5-point questions concerning worry about their child’s social development, emotional development, behavior, and language from 1 = not at all worried to 5 = extremely worried.
Parents completed questions from the Center for Epidemiologic Studies Depression Inventory (CES-D).27 The CES-D is a 20-item self-report scale that assesses depressive symptoms in adults. The CES-D has high internal consistency (coefficient α: 0.84–0.90) and modest test-retest reliability for 2- to 4-week intervals (r = 0.51–0.67).27 The parents also responded to items from the Beck Anxiety Inventory (BAI).28 The BAI is a self-report measure that consists of descriptive statements about common symptoms of anxiety. A parent indicates how much he or she has been bothered by each symptom on a 4-point scale from “not at all” to “severely bothered.” The BAI measure has adequate psychometric properties.
Parents rated expressiveness and conflict in the family using scales from the Family Environment Scale.29 The expressiveness and conflict scales of the Family Environment Scale are composed of a total of 18 items. These scales have demonstrated adequate reliability and validity and have been shown to discriminate between distress and nondistressed families.
Parents completed an adaptation of the Life Events Inventory (LEI).30 For this study, 40 items with the highest severity weights and greatest applicability to parents of young children were selected from the 55-item LEI. The LEI measure was derived in part from the Schedule of Recent Life Experiences.31 Social support was measured using the Medical Outcomes Study parent survey.32 More specific, 2 scales of this self-report measure, Tangible Support and Emotional Informational Support, were used. The scales are composed of 12 items and have demonstrated adequate psychometric properties, including a modest 1-year stability.
Parents first were asked whether they or anyone else had ever been so concerned with the study child’s development or behavior/emotions that they thought about talking to a professional. Parents next were asked whether they ever had discussed the study child’s development or behavior/emotions with a professional and with whom they had discussed the problem. Finally, parents were asked whether the study child had ever received any services for a developmental or behavioral/emotional problem.
Descriptive statistics were generated for all variables. Bivariate analysis was used to examine unadjusted differences in sociodemographic, health, and family characteristics of the study respondents and children for each of the following 3 outcomes: 1) thinking about seeking services (think), 2) talking to someone about services (talk), and 3) receiving services (use).
Because bivariate analyses do not adjust for multiple possible covariates, we developed logistic regression models for each outcome initially using the variables related to the outcomes of interest in the bivariate analysis. Independent variables that did not contribute significantly to the models were deleted individually (ie, the parameter estimates did not change by >10%, and the deletion did not change the G2 statistic significantly). Individual deletion was used to make sure that we noted any potential negative confounders. Additional multivariate logistic regression models were developed to assess the interaction effects between all variables in the main effects models and other variables suggested in the literature. None of the interaction terms was statistically significant. All analyses were conducted using the SAS 8.0 statistical software.
Children ranged in age from 11 to 39 months, were predominantly white, and rarely experienced chronic health or developmental problems (Table 1). However, 25.6% of these children scored above the 90th percentile on at least 1 of the Internalizing, Externalizing, and/or Dysregulation scales of the ITSEA. Respondents (Table 2) were usually biological mothers, had at least a high school education, and had incomes >185% of the federal poverty line. Respondents reported very good or excellent health, with 5% of respondents receiving scores on the BAI above the clinical cut point and 16.3% scoring above the clinical cut point on the CES-D.
The multivariate logistic regression models predicting the 3 outcomes of interest show that covariates differ depending on the stage of help seeking (Table 3). Variables were statistically significantly related to the outcome when the 95% confidence interval and odds ratio are displayed. Factors associated with thinking about seeking help (think) include the child’s age, race/ethnicity, developmental problems, and high ITSEA score, as well as the respondent’s educational level, anxiety score, low tangible support, parental worry about a child’s behavior and language, and use of child care. Child sex was retained in the model because deleting it changed the estimates by >10%. Speaking with a professional about a developmental or emotional problem (talk) was related to a child’s age, the presence of a developmental problem, and child stressful life events. Respondent and family characteristics associated with speaking with a professional included nonparticipation in public assistance programs, high anxiety scores on the BAI, and parental worry about language development. Again, child’s sex was retained in the model because of its effects on the other covariates. In contrast to the first 2 models, which show a range of child, respondent, and family factors related to problem identification and initial help seeking, the model describing the features related to actual receipt of any service for a developmental or emotional/behavioral problem showed only 3 statistically significant relationships. Whether a child received a service for a developmental or emotional/behavioral problem (use) was related to the presence of a developmental problem, family conflict, and parental worry about language development. A low rating of general child health was retained in the model because of its effect on the remaining 3 covariates.
To determine whether reported behavioral and emotional problems elicited services use in the same way that physical health or developmental problems did, we examined the percentage of parents who reported talking to a professional and subsequent service use by the type of problem reported (Table 4). Of the 46 parents who reported a child physical health problem, 30.4% reported speaking with a professional about that problem, and 57.1% of those who spoke to someone reported receiving a service. Thus, 17.4% of those who reported a physical health problem received a service for that problem. Of the 39 parents who reported a developmental problem, 82% spoke with a professional, and 90.6% of those parents reported a service. Seventy-four percent of children with developmental problems received a service. Of the 305 parents who reported a behavioral or emotional problem, only 20% spoke to a professional, and of those who spoke to someone, only 38.3% reported receiving a service. Overall, only 7.5% of parents who reported a behavioral or emotional problem also reported receiving services for that problem. Ten times as many children with developmental problems received services compared with children with mental health problems who received services.
Consistent with a recent study of adults, this examination of the factors related to parental problem recognition and receipt of services for a developmental, behavioral, or emotional problem suggests that different features may be related to the distinct aspects of the care-seeking process.33 Furthermore, although many parents think about speaking with a professional about their children’s problems (18%), fewer actually speak with someone (13%), and fewer still receive any services (5.6%). Parents who think about speaking with a professional about a child’s developmental, behavioral, or emotional problem have children who are older and white. These parents report considerable problem behaviors on the ITSEA. Furthermore, they report that their children have received a diagnosis of a developmental problem. These parents tend to be better educated, be high in anxiety symptoms, have little social support available to them, worry about their children’s behavior and language development, and use child care for the index child. Features related to actually speaking with a professional, most often (85% of the time) the child’s pediatrician, are very similar to the features related to thinking about speaking to a professional with a few exceptions. Stressful life events as well as other family factors seem to trigger speaking with a professional. The nonuse of public assistance programs, parental anxiety, and worry about a child’s language development prompt discussions with a professional. Although these 2 models differ somewhat in their composition, they point to a constellation of well-recognized child factors such as age and ethnicity, as well as the importance of parental education, recognition of need, parental anxiety, parental worry, and family stressors, as precursors to problem recognition and early attempts to evaluate the seriousness of the problem.
Whether a child received services for a problem, however, depended on far fewer features; specifically, parent-reported developmental problems, family conflict, and parental worry about language development with physical health status retained because of its influence on the other covariates. The co-occurrence of physical health and developmental problems in the model predicting service use suggests that the services that are available to treat young children may shape which needs are met. Although behavioral issues within pediatric primary care have received increased attention as a result of the development of a sanctioned educational curriculum and certification examinations, studies of practicing pediatricians document that, although practitioners identify significant percentages of their patients as having emotional and behavioral problems, clinicians consistently underidentify and undertreat children with emotional and behavioral problems.34–37 A recent survey of a random sample of American Academy of Pediatric members38 demonstrated that although pediatricians believed that they should be identifying a common mental health problem, childhood depression (90%), a significant percentage of these physicians believed that they are unprepared to make that diagnosis, and even fewer believed that they had the expertise to treat depression. The poor performance of pediatricians in the area of emotional and behavioral disorders is well documented; thus, the lack of attention to mental health problems is not unexpected.
When we examined whether physical, developmental, and behavioral/emotional problems prompt speaking to professionals or receipt of services once parents have spoken to a professional, we continued to see vast differences by type of problem reported. Most parents of children with developmental problems spoke to a professional (80%), and virtually all of these parents reported receiving services (90.6%), resulting in a high percentage of parents’ reporting services for developmental problems (74.4%). This is in contrast to behavioral and emotional problems, for which few parents reported speaking to professionals (20%) or receiving services even when they did speak with someone (38.3%). In contrast to developmental problems, for which almost everyone is served (74.4%), <10% of children with behavioral and emotional problems received any care.
The importance of developmental but not mental health problems as predictors of service use may reflect the availability of services for very young children in Connecticut. Children’s mental health services in Connecticut, because they are housed within the agency responsible for child welfare issues, receive less attention and funding than child protection activities. Although a unified child services agency has the potential for developing a seamless continuum of services to meet children’s multiple needs, in actuality, such an agency must set priorities, and those priorities begin and end with child protection. Furthermore, for the sample in this study, the chief services for very young children, the Connecticut Birth to Three System, is administered through the Department of Mental Retardation, and the services emphasize cognitive, speech and language, motor, and adaptive skills. Indeed, although some states provide services for children with mental health problems under Section H of the Individuals With Disabilities Education Act Amendments of 1997, the Connecticut Birth to Three System has chosen not to categorize children as eligible for services on the basis of only mental health problems. Until the principal systems available for children become expert in identifying, diagnosing, and treating mental health problems, mental health problems will receive no attention. Furthermore, until third-party reimbursement for mental health services is equal to that for physical health and developmental services, access to mental health services will be severely limited.39 Finally, until professionals, such as pediatricians, who routinely interact with young children gain skills in identifying and treating or referring such problems, children with mental health problems will continue to be underserved.34,37,38
In addition to provider and systems explanations for these findings, there are at least 2 explanations that are more family focused. Although most parents have some notion of age-appropriate physical and developmental milestones, parents may have less well-developed norms for behavioral health issues.20,40–42 Similarly, because of the stigma associated with behavioral health issues, parents and providers alike may be reluctant to identify such problems in young children.2 Consequently, family barriers to service use for behavioral and emotional problems may be considerable.
As with all research, the data reported here have certain limitations. First, these data are cross-sectional. Because variables were measured at 1 point in time, no assumptions can be made about causality. Second, the data all are generated from parent surveys. No attempt was made to verify any of the responses through the use of a second source of information (eg, medical records). Third, because this research was originally designed to test the reliability and validity of the infant mental health measure, the ITSEA, and to assess the prevalence of early emerging socioemotional problems and competencies, certain variables that may be important explanatory factors (eg, insurance benefits for mental health problems, stigma) were measured inadequately. Finally, because the original sample excluded children who were likely to have developmental delays, the rates of such delays are lower than would be expected in a community sample.
The data reported here demonstrate that covariates of problem recognition differ from those of service use. Furthermore, given evidence that the presence of mental health, physical health, and developmental problems affects problem recognition but that only physical health and developmental problems are related to service use and that rates of use of services vary dramatically across these 3 types of problems, it seems reasonable to conclude that the needs that are met may be dependent not on identification of problems but rather on either the availability of services or a family’s ability or willingness to access available services. In a system with available medical and developmental services and commonly accepted norms for these issues, these are the problems that receive attention. Until services for the full range of physical and mental health needs are available and accessible and professionals who are in contact with young children can identify and treat or refer mental health problems, these problems are likely to remain underserved.
This work was supported by National Institute of Mental Health Grant Problems and Competencies Under Age 5 (Carter PI) and the National Institute of Mental Health–funded Training Program in Mental Health Services Research (S.M.H.).
- ↵US Department of Health and Human Services (USDHHS). Mental Health: A Report of the Surgeon General. Washington, DC: US Government Printing Office; 2000
- ↵Burns BJ, Costello EJ, Angold A, et al. Data watch: Children’s mental health services use across service sectors. Health Aff.1995;14 :147– 159
- Garralda ME, Bailey D. Child and family factors associated with referral to child psychiatrists. Br J Psychiatry.1988;153 :81– 89
- ↵Individuals with Disabilities Education Act Amendments of 1997 (IDEA). P.L. 1997;105–117
- ↵Carter AS, Briggs-Gowan M. Infant-Toddler Social and Emotional Assessment (ITSEA) Manual. New Haven, CT: Yale University; 1999
- ↵Carter AS, Briggs-Gowan MJ, Jones SM, Little TD. The Infant Toddler Social and Emotional Assessment (ITSEA): factor structure, reliability and validity. J Abnorm Child Psychol. In press
- ↵Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. App Psych Meas.1977;1 :385– 401
- ↵Moos RH, Moos BM. Family Environment Scale: Manual. Palto Alto, CA: Consulting Psychologists Press; 1981
- ↵Horwitz SM, Leaf PJ, Leventhal JM, Forsyth B, Speechley KN. Identification and management of psychosocial and developmental problems in community-based, primary care pediatric practices. Pediatrics.1992;89 :480– 485
- Kelleher KJ, Childs GE, Wasserman RC, McInerny TK, Nutting PA, Gardner P. Insurance status and recognition of psychosocial problems: a report from the Pediatric Research in Office Settings and the Ambulatory Sentinel Practice Networks. Arch Pediatr Adolesc Med.1997;151 :1109– 1115
- Costello EJ, Edelbrock C, Costello AJ, Dulcan MK, Burns BJ, Brent D. Psychopathology in pediatric primary care: the new hidden morbidity. Pediatrics.1988;82 :415– 424
- ↵Stroul BA. Children’s Mental Health: Creating Systems of Care in a Changing Society. Baltimore, MD: Paul H. Brooks; 1996
- McMiller WP, Weisz, JR. Help seeking preceding mental health clinic intake among African-American, Latino and Caucasian youths. J Am Acad Child Adolesc Psychiatry.1996;5 :1086– 1094
- Copyright © 2003 by the American Academy of Pediatrics