BACKGROUND. Dr Van Cleave’s current address is Department of Pediatrics, Center for Child and Adolescent Health Policy, MassGeneral Hospital for Children, 50 Staniford St, Room 901, Boston, MA 02114. The association between bullying, being bullied, or being a bully/victim and having a special health care need has not been well described in a national sample of children with a broad variety of special needs.
OBJECTIVE. We aimed to determine the prevalence of bullying, being bullied, or being a bully/victim in children with special health care needs and associations of behaviors with particular types of special needs.
DESIGN. We performed a secondary data analysis using the National Survey of Children's Health, a nationally representative telephone survey conducted by the National Center for Health Statistics of >102000 US households.
METHODS. We measured associations between having a special health care need and being a victim of bullying, bullying other children, and being a bully/victim in children and adolescents aged 6 to 17 years. Multiple logistic-regression models were used to examine the association of children with special health care needs overall, and of particular special needs, with the bullying measures.
RESULTS. Overall, children with special health care needs were 21% of the population. In multivariate models adjusting for sociodemographic factors, being a child with special health care needs was associated with being bullied but not with bullying or being a bully/victim. Having a chronic behavioral, emotional, or developmental problem was associated with bullying others and with being a bully/victim.
CONCLUSIONS. Having a special health care need generally is associated with being bullied, and having a behavioral, emotional, or developmental problem is associated with bullying others and being a bully/victim. These findings may help pediatricians, mental health providers, and schools use targeted screening and interventions to address bullying for children with special health care needs.
Improving the quality of life and well-being of children with special health care needs (CSHCN) is a priority for those who care for them. CSHCN, as defined by the Maternal and Child Health Bureau, “have or are at increased risk for a chronic physical, developmental, behavioral, or emotional condition and who also require health and related services of a type or amount beyond that required by children generally.”1 Several recent articles have shed light on psychological challenges encountered by CSHCN,2–4 suggesting that CSHCN may be at particular risk for poor psychosocial functioning.
One area of psychosocial function that has not been studied extensively among CSHCN is bullying, including bullying other children, being bullied (also called peer victimization), or both. CSHCN may be at risk for involvement in bullying for a variety of reasons. First, CSHCN have a high rate of conduct problems,2 which may be associated with bullying other children. Also, many CSHCN may be singled out by their peers because their conditions cause them to appear or act differently from other children.
Previous studies have found that children with certain conditions have social relationships with peers that are similar to those of children without these conditions.5,6 However, other studies of bullying and peer victimization in children with chronic conditions indicate a high prevalence of being bullied among those with certain conditions (cancer,7 speech/language impairment,8 vision problems,9 and diabetes10) and of bullying others among children with psychological impairment.11 Studies that focus on children with 1 specific chronic condition are certainly informative; however, results may not be generalizable to other conditions, and broader population estimates of prevalence cannot be performed. We found only 1 study12 that used a population-based sample. This study was performed in Scandinavia, which limits its generalizability, especially given that the prevalence of bullying behaviors varies across countries.13
Being a victim of bullying is associated with anxiety, depression, and low self-esteem.14,15 Bullying other children (ie, as a perpetrator) is associated with social aggression and may be a manifestation of more complex mental health conditions.11 Bully/victims, children who bully others and are themselves also victimized by their peers, may represent a group that is at particularly high risk for poor social adjustment.16
The objective of this study was to test the hypothesis that having a special health care need is associated with bullying, being bullied, and being a bully/victim. Using a nationally representative data set of US households, we examined the association between having a special health care need and (1) being a victim of bullying, (2) bullying other children, and (3) participating in both behaviors. We investigated these associations among children aged 6 to 17 years while controlling for other factors known to be associated with bullying in general (age, race, gender, and family income)14,17–19 and with psychosocial adjustment and conduct problems for CSHCN (maternal and child health status).2,4
We used data from the National Survey of Children's Health (NSCH),20 a telephone survey of US households conducted by the National Center for Health Statistics (NCHS) from January 2003 to July 2004. This survey (available at www.cdc.gov/nchs/about/major/slaits/nsch.htm) asked about prevalence of health conditions as well as the child's family and community. In this survey, data were gathered on 102353 children aged 0 to 17 years from a member of each household that was most knowledgeable about the child's health. Fifty-one percent of the nonweighted sample were male, 69% were white, non-Hispanic, 10% were black, and 12% were Hispanic. The weighted response rate was 55.3%, and sampling weights are included with the data set to permit estimates that are generalizable to the noninstitutionalized population of children in the United States. Interviews were conducted in either English or Spanish. The survey was translated from English to Spanish by an experienced translator and was reviewed by several Spanish-speaking telephone interviewers. Bilingual (English/Spanish) interviewers conducted the interviews in Spanish.
The NCHS assured confidentiality of the families interviewed. For the public-use data set, the NCHS edited certain variables that represented information that could potentially identify certain families.
The NSCH contains the Children With Special Health Care Needs Screener, a previously validated instrument21 used to identify children with a special health care need on the basis of the Maternal and Child Health Bureau definition. The screener contains 5 stem questions that assess a child's ongoing (1) need for prescription medication, (2) need for more medical or mental health care or educational services than is usual for other children the child's age, (3) limitations in doing things that are normal for the child's age, (4) need for physical, occupational, or speech therapy, and (5) emotional, developmental, or behavioral problem for which treatment or counseling is needed. We used any affirmative responses to this screener to create a dichotomous variable that represents presence or absence of a special health care need. Because we hypothesized that bullying, being bullied, or both might vary for children who have different special health care needs (eg, need for prescription medication versus need for speech therapy), we also created separate variables that were based on affirmative answers to each of the 5 screening questions.
Bullying and Being Bullied
The NSCH contained 2 questions related to bullying that were asked of respondents with children aged 6 to 17 years. Respondents were asked how often the following item was true for the child in the past month: “He/she bullies, or is cruel or mean to others.” This question was derived from the Behavior Problems Index20 and has been used previously to assess bullying.14 Respondents chose from 4 possible answers: always, usually, sometimes, or never. We designated the sample child as a bully if the respondent chose always, usually, or sometimes. The respondents were also asked, “Are you concerned a lot, a little, or not at all about [the child] being ‘bullied’ by a classmate?” We used responses to this question to identify those children who were bullied; if the respondent acknowledged any concern (a lot or a little), we designated the sample child as being bullied. This question was developed by Child Trends (www.childtrends.org) for the NSCH. Information about validation of this question is not available, but similar questions have been used previously to assess, by parental report, peer victimization in children.12 As had previous investigators,19,22,23 we combined these variables to identify children who were both bullies and victims of bullying (bully/victims).
Several sociodemographic variables were used in the analysis because they represented potential confounders in the association of bullying, being bullied, and being a bully/victim with CSHCN status. Age in years was treated as a continuous variable. We used 5 categories for race and ethnicity: non-Hispanic white only, non-Hispanic black only, Hispanic only, multiracial, and other. In creating a variable to describe family structure, we categorized the families as 1-parent, 2-parent, or other. The NSCH contained a question regarding the highest level of education achieved by anyone in the household. We used 3 categories for this household member's education: less than high school, high school diploma or equivalent, and at least some college. We also examined the role of family income, using percent of the federal poverty level (FPL), which was calculated by the NCHS. Income was categorized as <100% FPL, 100%–200% FPL, or >200% FPL. Health insurance was categorized as private, public (Medicaid or State Children's Health Insurance Program), or not insured at the time of the survey.
Because child and maternal health statuses have been shown in other studies to be related to psychosocial functioning in CSHCN,4 we included these variables in our analysis. The NSCH contained items asking respondents to rate both the child's health and the mother's health. Maternal health was dichotomized by excellent/very good/good and fair/poor. We kept the 5 separate categories for the child health-status variable as excellent, very good, good, fair, and poor.
Subjects with incomplete data or missing data on the variables of interest were not included in the analysis. The most common sources of missing data were failure to report income and failure to report maternal health status (for cases in which the mother was not a member of the household). In comparing subjects with missing data to those with complete data, we found that children with complete data were slightly younger (mean ± SD: 11.5 ± 3.5 vs 11.8 ± 3.4 years; P < .001) and had an adult household member with higher levels of education (for those with complete data: 68% had more than high school, 26% had high school or less, and 6% had less than high school; for incomplete data: 55% had more than high school, 32% had more than high school, and 12% had less than high school; P < .001 for group comparison). The 2 groups did not vary with respect to gender or presence of a special health care need. We judged these variables (age and adult household member education level) not material to our analyses and, therefore, chose not to impute values for these variables.
The institutional review board at the University of Michigan Medical School approved this study.
We hypothesized that CSHCN would be more likely than children without CSHCN to be victims of bullying, bullies, and bully/victims. We also hypothesized that these associations may differ among the various subgroups of CSHCN; specifically, we hypothesized that children with emotional, developmental, or behavioral problems would be more likely than other CSHCN to bully, be bullied, and be bully/victims.
To test these hypotheses, we first conducted bivariate tests (Student's t test and χ2) to determine differences in outcome, sociodemographic, and health-status variables for CSHCN and other children.
We then performed several logistic-regression analyses to determine associations between bullying, being bullied, and being a bully/victim and having a special health care need using the inclusive CSHCN variable (ie, CSHCN overall) and the 5 variables representing each of the different screening questions. To test the effect of having an ongoing emotional, developmental, or behavioral problem on these associations, we repeated the regression models while adjusting for having an emotional, developmental, or behavioral condition (because of the hypothesized strong association of this variable with bullying, being bullied, and being a bully/victim). We then incorporated sociodemographic and health-status factors because they represented potential confounders. We hypothesized that CSHCN who had an emotional, behavioral, or developmental problem and also 1 other category of special health care need might be at particular high risk of bullying, being bullied, or both. We suspected an interaction when the odds ratio (OR) for each of these bullying categories deviated further from the null in the full models than in the unadjusted analyses. When this occurred, we assessed for interactions between having an emotional, behavioral, or developmental problem and the other categories of CSHCN using logistic-regression models that included the sociodemographic and health-status covariates. As a sensitivity analysis, we removed those responding “a little” to the question of whether the respondent was concerned about the child being bullied and those responding “sometimes” to the question of how often the child bullies others from the sample and recalculated the ORs of being bullied, bullying others, and being a bully/victim using the full models.
We used Stata 9 (Stata Corp, College Station, TX) for our analysis. Because the NCHS uses a complex sampling frame, survey weights provided by them were used to generate nationally representative estimates.
The weighted prevalence of having a special health care need was 21% for the 6- to 17-year-old children in our US-based sample. Results of bivariate comparisons of sociodemographic and health-status characteristics of CSHCN and children without special needs (non-CSHCN) are shown in Table 1. CSHCN were more likely to be male (58% vs 50%; P < .001) and had a mean age that was 3 months older than their non-CSHCN peers. CSHCN were more likely to be white (70% vs 63%; P < .001) and less likely to be Hispanic (9% vs 16%; P < .001). CSHCN were also more likely to be from single-adult households (30% vs 24%; P < .001), more likely to have a household member with greater than a high school education (71% vs 67%; P < .001), and less likely to report excellent health status (37% vs 67%; P < .001). CSHCN were more likely to be publicly insured (29% vs 21%; P < .001) but less likely to be uninsured (5% vs 10%; P < .001) than other children. Mothers with CSHCN endorsed worse health status than mothers of non-CSHCN (16% vs 9% with fair/poor health status; P < .001). The 2 groups did not vary significantly in family income.
Of note, there were specific diagnoses associated with parent-reported emotional, developmental, or behavioral problems. Parents reported that 57% of these children had attention-deficit/hyperactivity disorder, 55% had a behavioral or conduct problem, 46% had anxiety or depression, 26% had developmental delay or a physical impairment, and 7% had autism. Nine percent reported having none of these specific conditions. The categories are not mutually exclusive. In addition, we were able to quantify the proportion of CSHCN with respiratory allergies (39%), asthma (37%), eczema (15%), headaches (14%), bone, joint, or muscle problems (11%), food allergies (8%), diabetes (2%), and speech problems (1%). However, these problems may not represent the actual chronic problem that qualifies the child as having a special health care need. Regarding the various subcategories of CSHCN, 53% of CSHCN were in only 1 group, 21% of CSHCN were in 2 groups, and 26% were in >2 groups. Those who use more health services, have functional limitations, or receive special therapy were most likely to have overlap with another group.
Bullying, Being Bullied, and Being a Bully/Victim
Prevalence rates of bullying, being a victim of bullying, and both for CSHCN versus non-CSHCN are presented in Table 2. The overall prevalence of being bullied was 34.4%, for bullying others was 23.5%, and for being a bully/victim was 10.2%. In these unadjusted analyses, CSHCN were significantly more likely than non-CSHCN to be bullied (victims) and significantly more likely to be bullies (perpetrators). Children who were bully/victims were fewer in number than those reporting a single bullying behavior; however, CSHCN were bully/victims at about twice the rate of non-CSHCN.
Unadjusted and adjusted ORs for being bullied among CSHCN are presented in Table 3. Before and after adjustment for sociodemographic characteristics and health status, CSHCN were significantly more likely to be bullied than were non-CSHCN (adjusted OR: 1.16). Across all subcategories of CSHCN, the odds of being bullied was elevated in adjusted models.
The association with CSHCN was different for bullying (Table 4). In unadjusted analyses, CSHCN and all subcategories were significantly associated with bullying. However, once we controlled for having an emotional, developmental, or behavioral problem, this association remained significant only for children having functional limitations. Moreover, as shown in the full adjusted models in Table 4, having an emotional, developmental, or behavioral problem was the only CSHCN category that was significantly associated with bullying other children.
Being a bully/victim was also associated with having a special health care need. This was primarily a result of having an emotional, developmental, or behavioral problem or a functional limitation (Table 5), similar to the pattern observed for bullying only.
We assessed the interactions between having an emotional, developmental, or behavioral condition and other categories of CSHCN. The only interaction that reached statistical significance was the interaction between having an emotional, developmental, or behavioral condition and using prescription medication. This interaction seemed to be protective against being a bully (compared with having an emotional, developmental, or behavioral condition alone), but this term did not remain statistically significant when it was adjusted for sociodemographic covariates.
As a sensitivity analysis, the data were analyzed by using the full models with those responding “a little” to the question assessing concern about the child being bullied and those responding “sometimes” to the question of how often the child bullies others removed from the sample. These results using this subsample were similar to results using the complete sample with 2 exceptions. First, for the odds of being bullied, the associations with having a special health care need overall, using medication, and using more health services were no longer significant. Second, for the odds of bullying others, the magnitude of the effect of having a behavioral, emotional, or developmental problem was increased (OR: 11.5 [confidence interval (CI): 8.5–15.6] vs 3.13 [CI: 2.76–3.54]).
In our analysis of US children, we found a significant association between having a special health care need and being bullied by other children, bullying other children, and being both a bully and a victim. Being bullied was associated with each of the 5 categories of special health care needs (as defined by the CSHCN screening questionnaire), and this association persisted when adjusting for several sociodemographic variables and health-status variables. In contrast, bullying others was associated only with an emotional, developmental, or behavioral problem requiring treatment. Having an emotional, developmental, or behavioral problem and having a functional limitation was associated with being a bully/victim.
Our study is the first in the US to examine the relationship between bullying, being bullied, and being a bully/victim and having a special health care need on a national level, using a non–disease-specific definition of CSHCN. Our findings are consistent with 1 previous study that examined bullying in Scandinavian children with chronic conditions and with other previous studies in smaller samples that reported an association between certain chronic conditions, such as speech and vision problems, and peer victimization.8,9 In addition, our study echoes previous studies that found an association between having a chronic condition and a high prevalence of conduct disorders and poorer psychosocial functioning in general.2,4,24
CSHCN as Victims of Bullying
CSHCN were significantly more likely than non-CSHCN to be victims of bullying. The effect size is small (OR: 1.16) for CSHCN generally, but those with a functional limitation or a behavioral, emotional, or developmental problem are 1.5 to 2 times more likely to be a victim of bullying. We propose several explanations for this association. Some CSHCN may look or act differently from other children because of their health conditions. They may differ from their peers in physical appearance, or they may have physical limitations that prevent them from doing things that other children can do. Also, these children may have mannerisms or speech patterns that differ from other children, or they may have cognitive delays that cause them to interact differently with their peers. Previous studies offer conflicting evidence as to whether these signs of chronic conditions contribute to peer victimization,9,25,26 and these conflicting results may be a function of who is reporting the bullying (children versus parents versus teachers).
It is also possible that being bullied makes some chronic conditions worse. In previous studies, victims of certain kinds of bullying have more health problems, including depression, headaches, and abdominal pain, and worse health status, although these studies were cross-sectional in nature, and causality could not be established.15,27 In addition, the psychological effects of bullying may influence adherence to treatment. Storch et al10 proposed that some children may not adhere to treatment regimens at school (eg, checking blood sugar if diabetic) for fear of peer victimization. A recent meta-analysis has shown a link between depression and adherence to treatment regimens in both children and adults.28
CSHCN as Bullies
Our study found that bullying other children is associated with having a chronic developmental, emotional, or behavioral problem, but not associated with any of the other sub-categories of CSHCN. CSHCN with a developmental, emotional, or behavioral problem were >3 times more likely to bully as other children. This finding is expected, given that bullying itself can be a behavioral problem and is a diagnostic criterion for conduct disorder.18 However, previous research has been mixed regarding whether bullies tend to have more psychological and emotional problems.17,19 In fact, there is some evidence that bullies may be psychologically “stronger” than the general population.17 Nonetheless, several studies report a link between bullying and having other conduct problems, such as fighting and alcohol use,19 and psychosocial distress.14
CSHCN as Bully/Victims
Bully/victims likely represent a smaller, unique group of children who are at even higher risk for psychosocial problems than children in the previous 2 categories.16 Thus, our finding that CSHCN who have behavioral, developmental, or emotional problems are >3 times as likely to be a bully/victim was expected.
Our study has several limitations. The NSCH is designed to cover many topics broadly; as a result, questions that explore bullying, being bullied, and being a bully/victim more in depth were not available. Bullying was not defined in this survey for the parent respondents, and parents' perceptions of what constitutes bullying others or being bullied may vary. Also, this survey relied on parental report; self-report by children and adolescents, ideally in combination with peer and teacher reporting, is likely to be more accurate.29 Some children were excluded from the analysis because of missing data, a potential source of bias. However, excluded subjects varied little from those with complete data for several demographic and health factors (age, education level of adult household member, gender and CSHCN status). In addition, the CSHCN screener does not differentiate among children with developmental, behavioral, and emotional problems, and the prevalence of bullying or being bullied may be different for children with, for example, developmental problems compared with those with behavioral problems.
In this survey, parents were asked not if their child was a victim of bullying but if the parent was concerned about their child being bullied at school. The nature of this question may overestimate or underestimate victimization. Some parents might interpret the question as asking if they are concerned about the potential for their child to become the victim of bullying, although the child is not currently experiencing this problem. Other parents might feel victimization is a normal part of childhood and may minimize the problem for various reasons or may have discounted victimization occurring outside of the school setting.
Nonetheless, our study has several important implications for the care of CSHCN. First, our study reinforces the need for the primary care provider, responsible for comprehensive care for CSHCN, to screen for involvement in bullying, either as a perpetrator or a victim, at preventive care visits. Because CSHCN tend to use more health care than non-CSHCN and may have stronger connections with their providers, those providers are in the unique position to bring this problem to attention. Second, school-based bullying-prevention programs, which have been shown to be effective in decreasing bullying in schools,30 may benefit from a particular focus on CSHCN. Third, this study helps the parents of CSHCN to be particularly attuned to the risk of being a bully or the victim of bullying, which may prompt them to address this issue with teachers and other adults who are responsible for supervising children's peer-to-peer interactions.
Future investigations in this area will help further elucidate the relationship between having a chronic condition and bullying, being bullied, and being a bully/victim. Of particular value would be studies examining the role of bullying in the psychosocial functioning of CSHCN. Longitudinal studies could examine whether the long-term effects of bullying in the population generally, which include increased criminality and rates of depression, are more pronounced for CSHCN. Longitudinal studies could also examine other consequences of bullying, particularly the psychological effects of being bullied by others, on treatment adherence and clinical outcomes. Studies are also warranted to determine the prevalence of screening by children's health care providers for issues related to bullying among CSHCN and whether interventions to increase screening help to identify and affect outcomes for bullying, being bullied, and being a bully/victim in this population.
This work was supported by the Robert Wood Johnson Clinical Scholars Program (Dr Van Cleave).
- Accepted May 3, 2006.
- Address correspondence to Jeanne Van Cleave, MD, 50 Staniford St, Room 901, Boston, MA 02114. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵McPherson M, Arango P, Fox H, et al. A new definition of children with special health care needs. Pediatrics.1998;102 :137– 140
- Wallander JL, Varni JW. Effects of pediatric chronic physical disorders on child and family adjustment. J Child Psychol Psychiatry.1998;39(1) :29– 46
- ↵Horwood J, Waylen A, Herrick D, Williams C, Wolke D. Common visual defects and peer victimization in children. Invest Ophthalmol Vis Sci.2005;46 :1177– 1181
- ↵Storch EA, Lewin A, Silverstein JH, et al. Peer victimization and psychosocial adjustment in children with type 1 diabetes. Clin Pediatr (Phila).2004;43 :467– 471
- ↵Juvonen J, Graham S, Schuster MA. Bullying among young adolescents: the strong, the weak, and the troubled. Pediatrics.2003;112 :1231– 1237
- ↵Glew G, Rivara F, Feudtner C. Bullying: children hurting children. Pediatr Rev.2000;21 :183– 189; quiz 190
- ↵Blumberg S, Olson L, Frankel M, Osborn L, Srinath K, Giambo P. Design and operation of the National Survey of Children's Health, 2003. Vital Health Stat 1.2005;(43) :1– 114
- ↵Janssen I, Craig WM, Boyce WF, Pickett W. Associations between overweight and obesity with bullying behaviors in school-aged children. Pediatrics.2004;113 :1187– 1194
- ↵Cadman D, Boyle M, Szatmari P, Offord DR. Chronic illness, disability, and mental and social well-being: findings of the Ontario Child Health Study. Pediatrics.1987;79 :805– 813
- ↵Wolke D, Woods S, Bloomfield L, Karstadt L. Bullying involvement in primary school and common health problems. Arch Dis Child.2001;85 :197– 201
- ↵Ladd GW, Kochenderfer-Ladd B. Identifying victims of peer aggression from early to middle childhood: analysis of cross-informant data for concordance, estimation of relational adjustment, prevalence of victimization, and characteristics of identified victims. Psychol Assess.2002;14 :74– 96
- Copyright © 2006 by the American Academy of Pediatrics