Objective: To determine the prevalence of psychological maladjustment in clinic-based treatment-seeking obese children and adolescents (BMI ≥ 95th percentile) and the degree to which maternal, demographic, and youth factors correlate to the youths’ psychological adjustment.
Research Methods and Procedures: Anthropometrics, demographics (race, sex, insurance status), measures of youth psychological adjustment (self- and mother-report; Behavior Assessment System for Children), and maternal self-report of psychological distress (Symptom Checklist 90-Revised) were collected from 121 obese children and adolescents (55% white, 45% black) and their mothers.
Results: Approximately one-third of youths self-reported some psychological maladjustment, but two-thirds of youth were described by their mothers as experiencing some degree of psychological maladjustment. Adjustment difficulties were specific to social functioning, low self-esteem, and internalizing symptoms. Forty-one percent of mothers of child participants and 56% of mothers of adolescent participants reported clinically significant psychological distress. Youth self-report and mother-report of youths’ psychological difficulties were often most strongly associated with mothers’ level of psychological distress and/or family socioeconomic status rather than to youth characteristics (e.g., percent overweight, race).
Discussion: Psychological maladjustment levels among obese youth and their mothers were higher in this clinic-based sample than in treatment research-based samples. Present correlate findings extended to obese adolescents and reaffirm a potent association between youth self-report of their own psychological adjustment and their mother's level of psychological distress. Demographic characteristics and youth weight status were not consistent correlates of youths’ psychological functioning. Findings have implications for the translation of empirically supported pediatric obesity interventions to clinic-based treatment samples.
See Commentary by Tershakovec on pages 1537–1538
Increasingly prevalent childhood obesity is a growing public health concern (1) associated with both physical (2) and psychological (3) health consequences. However, poor psychological adjustment is not universal for obese children (4). Evidence has suggested that social and internalizing problems (e.g., depression, body image concerns) may be more common than externalizing (e.g., aggression) problems among these children. For instance, parents of obese children seeking weight treatment for their children have reported that their children exhibit greater social withdrawal and somatic complaints and lower social competence, with average scores often equivalent to values of children being seen in mental health clinics (5,6). Braet et al. (7) have found that obese children seeking treatment had higher rates of overall internalizing and externalizing difficulties, specifically with lower social competence than both non—treatment-seeking obese children and nonobese children. In that study, obese children seeking treatment were more likely to be at or above the threshold level of clinically significant impaired psychological functioning (20%) compared with non—treatment-seeking obese (4%) or nonobese children (3%).
There has been considerable focus on whether the level of children's overweight is related to children's psychological health status (7). However, evidence generally has suggested that parent psychopathology accounts for significant variance in their child's psychological functioning, with no evidence that child's level of overweight contributes independently to the prediction of child psychological functioning (8,9).
Limitations to prior research have included that, with few exceptions, these studies have been conducted among samples of children and families who are racially, economically, and psychologically homogeneous, because past samples were often enrolled in treatment research trials for pediatric obesity. Efficacy treatment trials rely on volunteerism for participation and often require children and parents to meet strict exclusion criteria including, for example, that no family member is undergoing psychiatric treatment (10). Furthermore, such samples may not be demographically representative of the population of overweight children or the children presenting for clinic-based weight control treatment. Only a few researchers have included an adequate sampling of ethnic minority populations (5,11), with the majority of data generated from middle to upper middle class white families (12). In addition, prior research examining psychological functioning among overweight children has been limited to children <12 years old, preventing exploration of developmental trends in overweight children's psychological functioning. Obese adolescents may present with unique psychological issues that dictate a different model of treatment than for younger children. The level of overweight in adolescence may be related to adolescents’ psychological functioning above and beyond parental psychopathology because of increasing stigma and weight criticism with child age and greater emphasis on body image and ideals in adolescence (13). Prior research has also relied primarily on mother-report of child psychological functioning. Reliance on mother-report may be problematic given that the data documenting maternal distress may be an important correlate of child adjustment difficulties for obese youth (8,9), and that the data suggest, for youth in general, that mother-report on child adjustment may be biased by maternal distress (14,15).
The purpose of this study was to determine the prevalence of psychological maladjustment, assessed by both child- and parent-report, in obese children and adolescents seeking treatment at an outpatient pediatric weight control clinic. It was hypothesized that rates of clinically significant social and internalizing problems would be higher than in prior studies examining obese children enrolled in treatment research trials and would differ between children and adolescents. The degree to which a mother's level of psychological distress was an additional correlate of child/adolescent psychological adjustment was also examined. It was hypothesized that mother's psychological distress would be related to child and adolescent psychological adjustment, but that child, and particularly adolescent, degree of overweight would be an independent predictor of children's psychological adjustment. Finally, the contribution of socioeconomic status (SES)1 to youth outcomes was examined. The hypothesis that SES would account for variance of child or adolescent psychological adjustment beyond that of maternal distress was presented.
Research Methods and Procedures
Participants and Procedures
Participants were children and adolescents and their mothers enrolling in a pediatric interdisciplinary weight management clinic. Clinic eligibility criteria included physician referral and a BMI (kilograms per meter squared) above the 95th percentile for age- and sex-based national norms from the Centers for Disease Control (16). Based on a retrospective chart review, of the 175 patients consecutively enrolled in the clinic from April 1999 to March 2001 who met age inclusion criteria (age, 8 to 17 years), complete data (i.e., all three psychological measures) were obtained from 79% (N = 138). Given the specific interest in mother's perspective of her child's, as well as her own, psychological health, patients were excluded from analysis if the primary caregiver who completed the clinic assessments was either a father, other relative, or state-appointed legal guardian (n = 17). Data were collected as part of the family's comprehensive intake evaluation that included physiological, anthropometric, nutritional, and psychological assessment. Approval for analysis of these clinical data was obtained from the Institutional Review Board of Cincinnati Children's Hospital Medical Center.
The final sample included 121 children (46 boys and 75 girls; mean age, 12.3 years) of whom 54.5% were white and 45.4% were black. Given the small percentage (2.7%) of other racial/ethnic groups, these children/adolescents were not included. Forty-six percent of families had private insurance or were self-paying, with the remainder being Medicaid recipients. This final sample (n = 121) did not differ by race, sex, age, insurance status, or degree of overweight from those participants who were excluded from analyses. The final sample was grouped by age into children (ages 8 to 11 years; N = 62) and adolescents (ages 12 to 17 years; N = 59) to match age groupings of the normative samples of the psychological measures.
Demographic information including race and type of insurance was obtained from the participants’ medical records.
Youth participants’ body weights were recorded using a digital scale, and heights were measured using a calibrated wall-mounted stadiometer. Children were weighed and measured in light clothing and without shoes. BMI values were calculated (kilograms per meter squared) and standardized (zBMI) using an age- (to the nearest month) and sex-specific median, SD, and power of the Box-Cox transformation (LMS method) based on national norms from the Centers for Disease Control (16). Details regarding the calculation of zBMI using the LMS method are available at www.cdc.govnchsaboutmajornhanesgrowthchartsdatafiles.htm. Mother's BMI was calculated based on the self-reported body weight and height recorded on the weight management program intake form. Epidemiological data (National Health and Nutrition Examination Study III) have shown that self-reported heights and weights are highly correlated with measured heights and weights (r = 0.89 to 0.97) in a nationally representative sample of adults 20 to 60 years of age (17).
Behavior Assessment System for Children
The Behavior Assessment System for Children (BASC) (18) was used to measure both positive (adaptive) and negative (clinical) dimensions of child and adolescent psychological adjustment. The BASC includes a self-report measure (SRP) and parent-rating scale (PRS). The SRP and PRS each have forms at two age levels, child (ages 8 to 11 years; SRP-C and PRS-C) and adolescent (ages 12 to 18 years; SRP-A and PRS-A), with substantial overlap but minor variation in item content across the two age levels. The SRP and PRS were interpreted with reference to national age norms, with standardization samples consistently matching U.S. population percentages of racial/ethnic groups. T scores (50 ± 10) indicate the distance of composite and scale scores from a norm-group mean. T scores in the “at-risk” range (clinical composites/scales: 60 ≤ T score ≤ 69; adaptive composites/scales: 31 ≤ T score ≤ 40) indicate significant problems that require close monitoring and/or intervention. Scores in the clinically significant range (clinical composites/scales: T score ≥ 70; adaptive composites/scales: T score ≤ 30) indicate a pattern of significantly maladaptive behavior. T scores ranging from 41 to 59 are considered average. For the purposes of this study, the “at-risk” and “clinically significant” categories were combined to denote the prevalence of problematic behaviors or self-perceptions outside of the average range.
The SRP includes 186 “True” or “False” statements at both age levels that render three composite scores, School Maladjustment, Clinical Maladjustment, and Personal Adjustment, based on 12 scales of the SRP-C and 2 additional scales of the SRP-A. The SRP shows adequate internal consistency (α, 0.86 to 0.96), satisfactory test-retest reliability, and convergent validity with other self-report measures of child (19) and adolescent (20) psychological adjustment. In this study, multivariate analyses focused on the six specific scale scores that assessed self-perceptions of social functioning (i.e., interpersonal relations, social stress, relations with parents), internalizing symptoms (i.e., depression and anxiety), and self-esteem.
The PRS contains 126 descriptors of adaptive and problem behaviors that parents rate on a four-point scale from “Never” to “Almost Always.” The PRS-C and PRS-A have three identical composite scores, Internalizing Problems, Externalizing Problems, and Adaptive Skills, based on the 15 scales of the PRS-C and 14 scales of the PRS-A. The PRS shows adequate internal consistency (α, 0.85 to 0.93), satisfactory test-retest reliability, and convergent validity with the commonly used parent-report measure—the Child Behavior Checklist (20). In this study, PRS multivariate analyses focused on the three composite scores that assessed internalizing problems (i.e., anxiety, depression, somatization), externalizing problems (i.e., aggression, hyperactivity, conduct), and adaptive skills (i.e., social skills, leadership, adaptability).
Symptom Checklist 90-Revised (SCL-90-R)
The SCL-90-R (21) is a 90-item instrument that measures adults’ current psychological symptom status. Symptoms are classified into nine dimensions of symptomatology and three global indices. The SCL-90-R demonstrates adequate internal consistency (α's range from .77 to .90) and satisfactory test-retest reliability (21). The Global Severity Index (GSI) combines information on the number of symptoms reported and their intensity, and is considered the best single summary measure of symptomatology. In the present study, the GSI T score (mean = 50, SD = 10) indicates the distance from the national age norm-group mean. In addition, for the purposes of comparison with previous research (8, 9), meeting clinical criteria for “caseness,” which is defined as a T score ≥63 on the GSI or on any two of the nine dimension scores of the measure, was also calculated for each participant.
Mean T scores and SD for all BASC PRS and SRP scales were calculated, as were the rates that children and adolescents scored in the at-risk/clinical range. Normal approximation to the binomial tests compared the observed frequencies of at-risk/clinical range scale scores to the population base-rate, based on instrument norms, with α set at 0.0001 to account for multiple comparisons. Zero-order correlations were calculated between all variables within each age group. Hierarchical regression analyses assessed the impact of child/adolescent race (Step 1), child/adolescent degree of overweight (zBMI) (Step 2), mother's level of psychological distress (GSI) (Step 3), and insurance status (Step 4) on the T scores of six specific child/adolescent scales that assessed self-perceptions of social functioning (i.e., Interpersonal Relations, Social Stress, Relations With Parents), internalizing symptoms (i.e., Depression and Anxiety), and Self-Esteem and the three BASC PRS composites (i.e., Internalizing, Externalizing, and Adaptive Skills). Data analyses were completed using SPSS, Version 11.5, SPSS, Inc., Chicago, IL.
The psychological adjustment of obese children and adolescents seeking clinical weight management treatment was not characterized by universally high levels of maladaptive behavior or problematic psychological functioning compared with a general normative sample, according to the report of mothers or youths themselves. However, many mothers of obese children and adolescents described their offspring as being “at risk” for internalizing symptomatology, such as greater depression, anxiety, and somatization. Adolescents were also described by their mothers as being more socially withdrawn and isolated, on average, than the age normative sample. Furthermore, the observed base rates of mother-reported at-risk/clinical range scores for internalizing symptomatology, withdrawal, and impaired social skills were significantly higher than corresponding rates in the general population based on published norms. Obese children and adolescents did not self-report experiencing difficulties across all aspects of psychological functioning, but a significant percentage of adolescents reported greater somatic complaints and lower self-esteem than observed in the general population. Present findings are consistent with prior research among mothers of obese children in both clinical (7) and research settings (9), but present rates of youths with at-risk/clinically significant psychological adjustment difficulties were generally higher than in prior studies, perhaps because of sample (research vs. clinical; less demographically homogeneous) or assessment method differences. The higher rates of psychological maladjustment found within this clinical sample warrant attention from pediatric obesity treatment practitioners.
Mother-reported child and adolescent psychological difficulties were often most strongly associated with the mothers’ own level of psychological distress, and for children specifically, with their insurance status (i.e., socioeconomic status), rather than with other youth characteristics (e.g., percent overweight, race). This finding replicated prior work by Epstein et al. (8,9) among samples of obese children (age, 8 to 12 years) and extended these findings to obese adolescents. These data also expanded the empirical literature by documenting that child and adolescent self-report of their own psychological adjustment was also most strongly associated with their mother's level of psychological distress rather than with youth or demographic factors. It has been argued that maternal depressive mood may distort a mother's perceptions of child behavioral problems (14,15,22), although this assertion has not gone unchallenged (23). By obtaining youth self-report, this study improves on prior research characterized by single-source (e.g., mother) designs (6,9) and studies for which sole reliance on mother-report is a potential confound (7,8,9).
In this study, lower SES, as measured by Medicaid status, was related to mother-report of poorer child, but not adolescent, psychological adjustment. Compared with mothers with higher SES (i.e., insured/self-paying), mothers with low SES described their obese children as at risk for greater internalizing and externalizing difficulties, as well as having fewer adaptive skills. Follow-up analyses revealed that mothers with low SES reported higher psychological distress for themselves than mothers with higher SES (Medicaid GSI = 60.6 ± 10.1; insured/self-paying GSI = 54.7 ± 8.5; t = 2.42; p < 0.05). Post hoc analyses showed that maternal distress (Step 2) mediated the relationship between insurance status (Step 1) and mother-report of child internalizing problems (Step 1: insurance status β = −0.28, p < 0.05; Step 2: insurance status β = −0.18, not significant, GSI β = 0.34, p < 0.01). These data suggest that low SES is associated with mother-report of internalizing problems for obese children only if the mother is distressed. In contrast, maternal distress did not mediate the relationship between insurance status and externalizing problems or adaptive skills, suggesting SES and maternal distress are independent predictors of child outcomes. Interestingly, SES was not related to mother-report of adolescent psychological adjustment, nor was SES a significant predictor of any child or adolescent self-report of psychological adjustment. Present and prior findings (8) lend support for careful analysis of the interaction of SES and maternal distress when examining the psychological correlates of pediatric obesity.
Strong associations between greater maternal distress and child or adolescent psychological adjustment problems are not unique to obese youth (24,25). Furthermore, parental psychopathology is known to disrupt parenting and the family environment. Maternal depression, in particular, has been linked to negativity and indifference in the parent-child relationship, inconsistent parenting, and poorer physical and psychological health outcomes in children (26,27). The consistent association between maternal and both child and adolescent psychological adjustment is particularly problematic, given the higher prevalence of maternal distress reported in this sample. The prevalence of clinically significant maternal distress was markedly higher in this sample of mothers seeking weight treatment for their children (∼50%) than in prior studies (8,9). Interestingly, despite the high rate of obesity in mothers, it was unrelated to her level of distress.
These findings have potential implications for family-based weight control treatment, particularly among clinic-based samples. For example, maternal distress may cause disruptions in parenting practices that reduce the effectiveness of parenting skills needed to support child weight control (28). However, no studies to date have specifically examined the impact of parent distress on parenting practices in families with an obese child or adolescent. Future research should also address whether high maternal distress is a unique phenomenon for treatment-seeking mothers or whether high psychological distress is also present among non—treatment-seeking mothers of obese youth. Greater maternal distress and mothers’ perceptions of their child/adolescent's poorer psychological and social functioning could be leading mothers to pursue weight treatment for their obese child/adolescent. Alternately, mothers without their own or perceived child's psychological adjustment problems may not be seeking weight control treatment for their obese children. This is consistent with evidence that mothers do not necessarily appreciate the physical health risk associated with obesity in childhood (29). Indeed, a significant percentage of mothers do not perceive their obese children as obese, and some indicate only becoming concerned about their child's weight if the child was being teased by peers (30,31). Whether maternal distress becomes a barrier to treatment success remains an unexplored empirical question requiring further examination. It is conceivable that incorporating interventions that specifically target maternal distress into behavioral intervention models might improve both the psychological status of the child and the mother, and obesity treatment outcomes.
The hypothesis that child and adolescent degree of overweight would be an independent predictor of child/adolescent psychological adjustment after controlling for maternal distress and demographic factors was not supported. It is possible that the relationship between child/adolescent psychological adjustment and degree of overweight may not be consistently linear, or a child/adolescent's weight may need to hit a particular threshold for psychological difficulties to develop. The restricted range of degree of overweight (all had BMIs ≥95th percentile) in this study limited a full exploration of these questions, which would require comparing children at different degrees of overweight rather than smaller gradations. To examine this further, we conducted post hoc analyses to determine whether there were differences in the measures of psychological adjustment when comparing those whose zBMI measured within either extreme quartile (1st to 25th vs. 75th to 99th) and found no significant group differences for children or adolescents. Thus, these data documented that, within an exclusively obese sample, there is no support for the hypothesis that those who are more severely obese are experiencing more psychological adjustment issues than those who are less severely obese.
Race was generally not a significant predictor of child and adolescent psychological adjustment. However, white adolescents self-reported lower self-esteem than black adolescents. Thus, these data provide additional empirical evidence that race may moderate the relation between obesity and self-concept for adolescents (32). The more positive self-esteem of black adolescents should be seen as a strength to be emphasized in intervention efforts. However, these data may also reflect a less negative cultural context for obesity in black culture, which, while promoting better self-esteem, may also may present unique barriers to successful weight management for black youth (33,34,35,36).
This study makes a number of important contributions to the empirical evidence regarding obese youths’ psychological adjustment, but it is not without limitations. The lack of a control group (e.g., nonobese, non—treatment-seeking obese) necessitated reliance of comparisons to instrument norms. The use of insurance status as a proxy variable for SES limits the ability to fully capture the social and economic circumstances of the participating households that may prove important to understanding how economic context is associated with pediatric obesity and its psychological correlates. This study was designed to assess traditionally defined dimensions of psychological adjustment, but this and prior research have not directly assessed more contextual information of the daily lives of obese children and adolescents, such as peer group and family factors that may be correlates of pediatric obesity, treatment efficacy, and/or psychological functioning within this increasingly prevalent population. For instance, recent data suggest that obese children have significantly compromised perceived health quality of life, with levels similar to children diagnosed with cancer (37). Recent evidence has suggested that overweight adolescents are more socially marginalized than their non-overweight peers, being much less likely to be identified as friends and having friends who are less popular themselves (38). Some of the obese children and adolescents in this study self-reported significant interpersonal difficulties, and many were described by mothers as being socially withdrawn. While parent- and self-report of peer relations may provide some information about how obese youth fare in the peer environment, both peers and teachers have been established as favored, as well as reliable and valid, sources for assessment of child and adolescent social functioning (39). Future research should use peer- and other-report methodologies, because this may improve our understanding of the impact of obesity on children's concurrent and future adjustment and help to identify social behaviors that could impact successful weight management. Given the important role of maternal distress in this population, further examination of additional family correlates such as paternal psychological health, family functioning (i.e., cohesion, conflict), child-rearing practices, and the parent-child relationship is long overdue. Such exploration would potentially identify variables that could be targeted to maintain positive lifestyle changes necessary for successful weight management among obese children and adolescents.