 |
 |
|
|
|
Previous literature often describes a relation between runaway behaviour in adolescence and a history of maltreatment as a child (1–6). Estimates of exposure to child physical abuse among teenage runaways range from 6% (4) to 86% (1). For child sexual abuse, the figures range from 1% (3) to 73% (4). Adolescents who run away from home are at risk for poor psychological adjustment, depression, anxiety, substance use, suicidal behaviour, and behaviour problems (4,7–12). Ryan and others emphasize that runaways with a history of physical or sexual abuse have greater needs for mental health services, compared with their nonabused counterparts (5). Female adolescent runaways are particularly vulnerable to adjustment difficulties, compared with male adolescent runaways (1,3,4,8,9,11). In addition, a study conducted by Janus and others observed that female runaways have a tendency to run away at an earlier age than male runaways (1). It is therefore crucial to gain a better understanding of women who run away in adolescence. Previous research has been limited by the selection of samples from youth shelters or drop-in centres. This study investigates the relation between a history of child abuse, lifetime psychiatric disorder, and running away behaviour in a community sample of women. MethodThe 1990 OHS was a province-wide household survey that gathered information about the physical health of those aged 15 years and over. The OHS used a 2-stage stratified design; excluded were foreign service personnel, the homeless, persons residing in institutions, First Nations people living on Reserves, and residents of extremely remote locations. The 1991 OHSUP gathered data on mental health from one randomly selected respondent in each participating OHS household. Boyle and others document the details of the OHSUP methodology elsewhere (13). Measures The question “Before the age of 16, did you ever run away overnight from home (other than from a group or foster home)?” identified runaway behaviour. The CMHSR assessed physical and sexual abuse by an adult while the respondent was “growing up.” To determine physical abuse, respondents were asked how many times they experienced the following 6 incidents: “pushed, grabbed, or shoved you”; “threw something at you”; “kicked, bit, or punched you”; “hit you with something”; “choked, burned, or scalded you”; and “physically attacked you in some other way.” Response options were often, sometimes, rarely, or never. The minimum frequency for qualifying on the item varied with the severity of the item; for example, “choking, burning, or scalding” required a response of rarely, whereas “pushing, grabbing, and shoving” had to occur at least sometimes. Meeting the required frequency on one or more items was deemed physical abuse. Questions on sexual abuse asked whether any adult ever “exposed themselves to you more than once”; “threatened to have sex with you”; “touched the sex parts of your body”; and “tried to have sex with you or sexually attacked you.” Answering yes to one or more items was deemed sexual abuse. Investigation of the CMHSR with a clinical sample of 34 adolescents found test–retest reliability kappas of 0.75 for physical abuse, 0.78 for severe physical abuse, 1.0 for sexual abuse, and 1.0 for severe sexual abuse (MacMillan HL, Fleming JE, unpublished data). MacMillan and others describe the instrument in detail in a separate publication (14). We constructed 4 groups of women: those with no abuse, those with physical abuse only, those with sexual abuse only, and those with physical plus sexual abuse (combined abuse). Also included in the CMHSR was a question regarding the perpetrator(s) of these acts, asked once after the physical abuse items and once following the sexual abuse items: “Who did this to you? Please indicate all the people who did any of these things to you.” Response options were natural father, stepfather, natural mother, stepmother, older brother or sister, other relative, and some other person. To determine whether the respondent had run away from abuse within the home, we dichotomized according to the likelihood that the perpetrator resided in the home; the first 5 response options were tested against the last 2. Women who reported more than one perpetrator, with one (or more) in the home and one (or more) outside the home, were deemed to be in the “abuse within the home” group. To determine lifetime prevalence of psychiatric disorder, trained interviewers administered the UM-CIDI, a revised version of the CIDI (15) based on the DSM-III-R. Offord and others have found the CIDI to have good test–retest reliability, interrater reliability, and validity for the psychiatric disorders included in the study (16). Diagnoses included major depressive disorder, anxiety disorder, alcohol abuse or dependence, illicit drug abuse or dependence, and antisocial behaviours. “Any lifetime psychiatric disorder” was defined as the presence of at least 1 of these 5 disorders. With regard to parents, “any lifetime parental psychiatric disorder” meant that the respondent reported at least 1 of the following 4 disorders in his or her parent: major depressive disorder, mania, schizophrenia, or antisocial disorder. The level of parental education (high school completed or not) measured childhood SES, and the respondent’s current family income (low income, or not, derived from income level, household size, and urban or rural residence) measured current SES. The OHSUP interview was conducted face-to-face, except for the CMHSR, which respondents completed in private and returned to the interviewer in a sealed envelope. Data Analysis Data were weighted to adjust for nonresponse and to account for age and sex distributions in the province. For the current study, we restricted analysis to women under age 65 years. Older adults were administered a briefer interview to reduce respondent burden; thus, psychiatric diagnoses were unavailable for them. Respondents missing data for any measure used in analysis were excluded from all analyses. For analysis, we used SUDAAN (17), a software package that adjusts for survey design. Proportions, means, and their CIs were calculated. Bivariate ORs and CIs were generated by logistic regressions. A full logistic model tested the association of abuse and lifetime psychiatric disorder with runaway behaviour. The neither abuse group was the reference group in the comparison of abuse types in the regression. Age, family income, parent education, and parent psychopathology were control variables. ResultsSample Of 14 758 eligible OHS households, 13 002 (88.1%) households participated. A comparison of respondents with nonrespondents indicated that the latter group tended to be male, were older, and lived in urban settings. The 2 groups were similar in health status, employment, income level, and marital status. Of the participating 13 002 OHS households, 9953 (76.5%) respondents participated in the OHSUP. Of the 4285 OHSUP female respondents under age 65 years, 3760 (87.7%) had complete data. Correlates of Running Away Table 1 presents the characteristics of the sample and the bivariate associations (ORs) of each characteristic with running away. About 6% of women reported running away from home before age 16 years. Bivariate analysis showed that respondents who were younger at the time of the OHSUP interview were significantly more likely to report running away. Other significant correlates were childhood physical, sexual, and combined abuse; psychiatric disorder; parental psychiatric disorder; and current low income, which increased the likelihood of running away by twofold to more than sevenfold. Only parental education, the proxy for childhood SES, showed no significant association with running away.
With regard to sexual abuse, 21.7% of women who reported sexual abuse within the home ran away, compared with 15.0% of women reporting sexual abuse perpetrated by someone outside the home. This difference was not statistically significant (OR 1.57; 95%CI, 0.78 to 3.18). With regard to physical abuse, the cell representing those who experienced physical abuse from persons outside the home and who ran away contained fewer than 30 women (unweighted) and therefore is not reportable. Running Away, Child Maltreatment, and Psychiatric Disorder In the full regression model, physical abuse only, sexual abuse only, and combined abuse significantly increased the risk of running away, relative to the group reporting no abuse, although the sexual abuse only OR was just marginally significant. After psychiatric disorder, parental disorder, parental education, and low income were controlled for, respondents who reported a history of physical abuse were more than twice as likely to report running away, compared with those reporting no abuse. Those reporting a history of sexual abuse were more than 2.5 times more likely to report runaway behaviour, and those who reported experiencing both types of abuse were almost 4 times as likely to run away, compared with their nonabused counterparts. Parental psychiatric disorder was an important correlate of running away; the association was somewhat weaker for respondent psychiatric disorder. Table 2 shows ORs, 95%CIs, and betas from the logistic model.
DiscussionThe findings from this study suggest that the relation between child maltreatment and running away from home in adolescence is not exclusive to runaways involved with youth shelters (1,3–6). We also found a significant association between runaway behaviour and lifetime prevalence of psychiatric disorder in this community sample of women. Recent literature shows adolescent runaways to be at high risk for mental health difficulties (5). Some research suggests that running away puts adolescents at risk for developing disorder (8), whereas other research shows that adolescents with psychiatric disorders are at higher risk for running away (12). Significant relations were also obtained between running away and parental psychiatric disorder. Sullivan and Knutson identify “mental/emotional problems present in a family member” as more common in families of abused runaways, compared with nonabused runaways (12). It is possible that a parental psychiatric disorder is a confounding factor in the link between running away and psychiatric impairment. We were unable to demonstrate a significantly different effect for abuse within the home, compared with outside the home, likely because of small sample sizes. Respondents’ current low income was also related to runaway behaviour. The direction of the relation is unclear: Was income a risk factor for running away, or was it a result of runaway behaviour? The nonsignificant association of childhood SES, as measured by parental education, suggests that SES was not a risk factor; however, the measure may be too crude to draw firm conclusions. Notably, although many of the variables in the analysis were related to each other (for example, respondent lifetime psychiatric disorder and parental psychiatric disorder), all variables except parental education made a significant independent contribution to the full model. The fact that the survey was administered to a large, probability-based community sample is an important strength of this study. Previous work has focused on smaller samples from youth shelters or drop-in centres, with no comparison group of nonrunaways. Abuse was determined from several items (not, as is often the case in large surveys, from one item), a questionnaire with good preliminary reliability was used, and abuse was rigorously defined. We constructed abuse groupings so that the effect of different types of abuse, as well as the cumulative effect of abuse, could be evaluated. Of note, 54% of women in this sample who reported sexual abuse also reported physical abuse. This group was larger than the group reporting sexual abuse only. The small size of the group reporting sexual abuse only may account for its marginal significance level in the full regression. Regarding limitations, the weaknesses inherent in all survey research apply. The cross-sectional design does not allow for causal inferences: we cannot determine the temporal links between exposure to child maltreatment, psychiatric disorder, and running away episodes. Further, self-reported data can be problematic in that child maltreatment and parental psychiatric disorder may be underreported because of stigma or for various other reasons. Recall may also be a factor. It is possible that recall bias accounts for the association of running away and age. These findings have important implications from a clinical and policy perspective. Clearly, female youths with a history of exposure to physical or sexual abuse, and especially combined abuse, have a significantly increased risk for running away. Much of the literature to date has focused on the risks associated with child sexual abuse, yet child physical abuse and other factors such as parental psychiatric disorder and respondent psychiatric disorder are also critical. This type of information can help clinicians when they consider the risk indicators associated with running away among adolescent girls. The association between abuse and running away requires further study; qualitative techniques could have helped us understand the pathways that guided these youth to eventual residence in households (to be included in the OHSUP, respondents had to be living in a household). In quantitative research, other possibly influential variables, such as family characteristics and educational attainment, could be gathered. The fact that these results arose from a community sample underscores the importance of assessing the mental health needs of all adolescent runaways, not just those who use shelters or live on the street. Funding and SupportThis research was supported by the Wyeth CIHR Clinical Research Chair in Women’s Mental Health and the CIHR Institutes of Gender and Health; Aging; Human Development, Child and Youth Health; Neurosciences, Mental Health and Addiction; and Population and Public Health. Dr MacMillan was supported by a William T Grant Faculty Scholar Award. AcknowledgementsSpecial thanks to Stephanie Wong for her assistance in the careful editing of this manuscript. References1. Janus MD, Archambault FX, Brown SW, Welsh LA. Physical abuse in Canadian runaway adolescents. Child Abuse Negl 1995;19:433–47. 2. Kaufman JG, Widom CS. Childhood victimization, running away, and delinquency. Journal of Research in Crime and Delinquency 1999;36:347–70. 3. Kufeldt K, Durieux M, Nimmo M, McDonald M. Providing shelter for street youth: are we reaching those in need? Child Abuse Negl 1992;16:187–99. 4. McCormack A, Janus MD, Burgess AW. Runaway youths and sexual victimization: gender differences in an adolescent runaway population. Child Abuse Negl 1986;10:387–95. 5. Ryan KD, Kilmer RP, Cauce AM, Watanabe H, Hoyt DR. Psychological consequences of child maltreatment in homeless adolescents: untangling the unique effects of maltreatment and family environment. Child Abuse Negl 2000;24:333–52. 6. Whitbeck LB, Hoyt DR, Ackley RA. Families of homeless and runaway adolescents: a comparison of parent/caretaker and adolescent perspectives on parenting, family violence, and adolescent conduct. Child Abuse Negl 1997;21:517–28. 7. Janus MD, Burgess AW, McCormack A. Histories of sexual abuse in adolescent male runaways. Adolescence 1987;22:405–17. 8. Molnar BE, Shade SB, Kral AH, Booth RE, Watters JK. Suicidal behavior and sexual/physical abuse among street youth. Child Abuse Negl 1998;22:213–22. 9. Powers JL, Eckenrode J, Jaklitsch B. Maltreatment among runaway and homeless youth. Child Abuse Negl 1990;14:87–98. 10. Rotheram-Borus MJ, Mahler KA, Koopman C, Langabeer K. Sexual abuse history and associated multiple risk behavior in adolescent runaways. Am J Orthopsychiatry 1996;66;390–400. 11. Stiffman AR. Physical and sexual abuse in runaway youths. Child Abuse Negl 1989;13:417–26. 12. Sullivan PM, Knutson JF. The prevalence of disabilities and maltreatment among runaway children. Child Abuse Negl 2000;24:1275–88. 13. Boyle MH, Offord DR, Campbell D, Catlin G, Goering P, Lin E, and others. Mental health supplement to the Ontario Health Survey: methodology. Can J Psychiatry 1996;41:549–58. 14. MacMillan HL, Fleming JE, Trocmé N, Boyle MH, Wong M, Racine YA, and others. Prevalence of child physical and sexual abuse in the community. Results from the Ontario Health Supplement. JAMA 1997;278:131–5. 15. World Health Organization. Composite International Diagnostic Interview (CIDI): a) CIDI interview (version 1.0), b) CIDI-user manual, c) CIDI training manual, d) CIDI-computer programs. Geneva: World Health Organization; 1990. 16. Offord DR, Boyle MH, Campbell D, Goering P, Lin E, Wong M, and others. One-year prevalence of psychiatric disorder in Ontarians 15 to 64 years of age. Can J Psychiatry 1996;41:559–63. 17. Research Triangle Institute. SUDAAN for Windows; release 7.5.3. Research Triangle Park (NC): Research Triangle Institute; 1999. Author(s)Manuscript received July 2004, revised, and accepted February 2005. 1. Psychologist, formerly, Postdoctoral Fellow, Offord Centre for Child Studies, Department of Psychiatry and Behavioural Neurosciences, Faculty of Health Sciences, McMaster University, Hamilton, Ontario. 2. Research Associate, Offord Centre for Child Studies, Department of Psychiatry and Behavioural Neurosciences, Faculty of Health Sciences, McMaster University, Hamilton, Ontario. 3. Professor, Departments of Psychiatry and Behavioural Neurosciences and Pediatrics, Faculty of Health Sciences, McMaster University, Hamilton, Ontario. Address for correspondence: Dr HL MacMillan, Offord Centre for Child Studies, Patterson Building, Chedoke Division, 1200 Main Street West, Hamilton, ON L8N 3Z5 e-mail: macmilnh@mcmaster.ca
1 | 2
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||