Risk behaviors in maltreated youth placed in foster care: a longitudinal study of protective and vulnerability factors☆
Introduction
The number of children in out-of-home care increased by 95.3% (280,000–547,000) from 1986 to 1999 (Petit & Curtis, 1997; US Department of Health and Human Services, 2000). Despite the increase, systematic research, especially longitudinal studies on maltreated foster care populations, has been lacking. What is known about maltreated children entering foster care is that they manifest a host of emotional, behavioral, social and developmental problems, and are in need of many specialized services (Clausen, Landsverk, Ganger, Chadwick, & Litrownik, 1998; Halfon, Mendonca, & Berkowitz, 1995; Hochstadt, Jaudes, Zimo, & Schachter, 1987; Landsverk & Garland, 1999; Reams, 1999).
Unfortunately, for many maltreated children, the experience of trauma does not cease when they are placed in out-of-home care; victims are at an increased risk of experiencing additional negative events throughout their lives (cf McDonald, Allen, Westerfelt, & Piliavin, 1996). In addition to suffering the consequences of child maltreatment, children placed in foster care are likely to experience additional trauma by being removed and often isolated from their homes, schools, friends and family. Furthermore, these stressors may be exacerbated by frequent placement changes, not uncommon for older foster children (Taussig, 1998). Given their already increased risk of emotional and behavioral problems, youth in foster care may lack appropriate coping resources to handle the multiplicity of stressors associated with multiple life transitions.
Maltreated children in foster care are overrepresented among multiple service systems including juvenile justice and mental health (Blumberg, Landsverk, Ellis-MacLeod, Ganger, & Culver, 1996; Halfon, Berkowitz, & Klee, 1992; Widom, 1991b). The long-term research on foster children grown up suggests that they are at risk of experiencing continued difficulties in adulthood. A recent study found that 12–18 months after leaving foster care (due to emancipation), 27% of the males and 10% of the females had been incarcerated, 37% had not finished high school, and 50% were unemployed (Courtney & Piliavin, 1998).
One of the explanations posited for why maltreated and foster care youth may be more likely to experience subsequent negative life events is that they are engaging in behaviors that place them at risk. Although the types of risk behaviors in which maltreated and foster care youth engage do not differ from those of general adolescents, research shows that maltreated and foster youth tend to engage in these behaviors at an earlier age, at a greater frequency, and with more intensity (McDonald et al., 1996; Springs & Friedrich, 1992).
There are four major risk behavior domains that have been examined in maltreated and foster care adolescents: Sexual, Delinquent/Violent, Substance Use, and Suicidal/Self-Destructive behaviors (Widom, 1994). These behaviors all have the potential for severe short- and long-term negative consequences. The current study examined these four domains of risk behavior in a longitudinal study of children who had been placed in foster care due to abuse and/or neglect. Data collected from these children and their caregivers shortly after their placement in foster care were used to predict engagement in risk behaviors over 5 years later.
Studies have found that maltreated individuals and youth placed in foster care are more likely to engage in a host of risky sexual behaviors than are their non-abused peers. Such behaviors include initiating sexual intercourse at an earlier age, having a greater number of sexual partners, infrequently or inconsistently using contraception, and exchanging sex for money, drugs, or other things (Boyer & Fine, 1992; Polit, White, & Morton, 1990; Risley-Curtiss, 1997; Rotheram-Borus, Mahler, Koopman, & Langabeer, 1996; Silbert & Pines, 1981).
Maltreated youth are also more likely than their non-maltreated peers to engage in delinquent and violent behaviors (Paperny & Deisher, 1983; Sandberg, 1989). This finding appears to hold whether maltreatment rates are examined within incarcerated samples (e.g., Stein & Lewis, 1992) or rates of arrests or incarceration are examined within maltreated samples (Burgess, Hartman, & McCormack, 1987; Widom, 1989, Widom, 1991b). Elevated rates of maltreatment (relative to the population at large) have also been consistently reported among teenage runaway samples (Janus, Archambault, Brown, & Welsh, 1995; Stiffman, 1989, Widom, 1994). Adults with a history of placement in foster care appear to have increased criminal activity and arrests although there are methodological problems with many of the studies that demonstrate this finding (McDonald et al., 1996).
Research has also demonstrated that adolescents with a history of maltreatment have experimented more with alcohol and drugs, have done so at an earlier age, and are more likely to report current use of drugs and alcohol (Riggs, Alario, & McHorney, 1990; Rotheram-Borus et al., 1996; Silverman, Reinherz, & Giaconia, 1996). Additionally, drug use has been found to be strongly associated with violent and delinquent behaviors among adolescents (Dembo, Dertke, Borders, Washburn, & Schmeidler, 1988; Dembo et al., 1989). There is some limited evidence for increased drug and alcohol use in foster care populations (McDonald et al., 1996).
Although substance use is conceptualized by some researchers as a self-destructive behavior (cf Lindberg & Distad, 1985), other researchers have focused on the specific associations between child maltreatment and suicidal/self-mutilation behaviors. Studies have found that those adolescents who report past suicide attempts are more likely to have experienced maltreatment than are youth who have not attempted suicide (Molnar, Shade, Kral, Booth, & Waters, 1998; Silverman et al., 1996, Widom, 1994; Wilde, Kienhorst, Diekstra, & Wolters, 1992). In addition, adolescent psychiatric inpatients with a history of maltreatment have more severe suicidal ideation and a greater number of suicide attempts than patients without a history of maltreatment (Sansonnet-Hayden, Haley, Marriage, & Fine, 1987; Shaunesey, Cohen, Plummer, & Berman, 1993). In terms of self-mutilation, studies have consistently found a higher prevalence of self-injurious behaviors among child and adult survivors of maltreatment (Boudewyn & Liem, 1995; Green, 1978; Rosenthal & Rosenthal, 1984; van der Kolk, Perry, & Herman, 1991).
Although maltreated youth and youth in foster care are elevated in their levels of engagement in various risk behaviors, it is critical to remember that not all, and not even the majority of these youth engage in such behaviors (Widom, 1994). This study examined some of the factors that might predict why certain youth who were maltreated and placed in foster care engaged in risky and destructive behaviors, whereas others were able to avoid these negative behavioral sequelae. Because risk behaviors can result in severe negative consequences for both the youth (e.g., suicide) as well as others (e.g., assault), the ultimate goal is to design prevention and intervention programs that will mitigate the display of such behaviors, thereby reducing negative consequences. This, however, must rely on a solid understanding of the life-course trajectories of maltreated youth in foster care in order to understand better those factors that increase the likelihood of risky and destructive behaviors and those that protect children from the development of these behavioral patterns.
From studies of specific behavioral and emotional sequelae of maltreatment, certain vulnerability and protective factors have been identified. No longitudinal studies of psychosocial risk and protective factors in a foster care sample, however, were identified. The current study sought to examine whether psychosocial predictive factors (i.e., measures of social support and self-perception), that have been identified as protective factors in maltreated populations were also predictive in a longitudinal study of maltreated youth placed in foster care.
Studies of maltreated youth suggest that they may have a difficult time forming secure attachments with their caregivers (cf Cicchetti, Barnett, Rabideau, & Toth, 1991; Emery, 1989). Removal from one’s home and placement with strangers may exacerbate attachment difficulties. The long-term effects of insecure attachments include interpersonal difficulties, disturbances of self, impaired affect regulation, and internalizing and externalizing behavior problems (Alexander, 1992; Armsden, McCauley, Greenberg, Burke, & Mitchell, 1990; Greenberg, Speltz, & DeKlyen, 1993). Alternatively, studies of maltreated youth have found that an early, consistent relationship with a supportive adult has a protective effect (Rutter, 1990). Perceived levels of social support, both family support and support outside the family, have been found to moderate the relationship between early maltreatment and later emotional and behavioral problems (Burgess et al., 1987; Conte & Schuerman, 1987; Folkman, Chesney, Pollack, & Phillips, 1992; Romans, Martin, Anderson, O’Shea, & Mullen, 1995; Testa, Miller, Downs, & Panek, 1992; Widom, 1994; Wyatt & Mickey, 1989).
Peer support and acceptance also have been demonstrated to serve as protective mechanisms, although for at-risk populations, high social acceptance may actually be a risk factor (Cauce, Felner, & Primavera, 1982; Farrington, Gallagher, Morley, St. Ledger, & West, 1988; Luthar, 1993, Seidman et al., 1999). One study of maltreated youth demonstrated that a high level of social support from peers and perceived social acceptance were positively correlated with engagement in various risk behaviors, although social support from deviant peers mediated this effect (Taussig, Brown, & Litrownik, 1996).
Low self-esteem in maltreated individuals is a consistent finding, and researchers have hypothesized that this results from the internalization of negative messages resulting from abusive and/or neglectful experiences (Briere, 1992; Butler & Burton, 1990; Finkelhor & Browne, 1986; Green, 1978, Rotheram-Borus et al., 1996, Widom, 1994). These negative messages may become self-fulfilling prophesies, as engaging in risky and destructive acts may be a way to reinforce both one’s own and others’ negative perceptions (Jehu, 1992). Additionally, maltreated youth may have difficulty understanding the causal relationships between their cognitions, behaviors, and the outcomes, which can then lead to learned helplessness, feelings of powerlessness, and decreased self-efficacy (Finkelhor & Browne, 1986; Thompson, Potter, Sanderson, & Maibach, 1997). Such a cognitive style may account for maltreated children engaging in risky and destructive behaviors if: (1) they do not view these behaviors as risky, and (2) they do not think their actions will have any effect on outcomes. Researchers have found that low self-esteem, self-derogation, perceived lack of control, and hopelessness among maltreated youth are related to engagement in risk behaviors (Becker-Lausen, Saunders, & Chinsky, 1995; Dembo et al., 1989; Jessor, Van Den Bos, Vanderryn, Costa, & Turbin, 1995; Kaliski, Rubinson, Lawrance, & Levy, 1990; Lanz, 1995, Reinherz et al., 1995, Stiffman, 1989, Widom, 1994).
While several studies have demonstrated relationships between these psychosocial predictors and risk behaviors in maltreated individuals, methodological problems plague this area of research, as non-representative or convenience samples are often used, self-report and retrospective reports of maltreatment are relied upon without confirmation, and definitions of maltreatment vary greatly (Widom, 1994; Zingraff, Leiter, Myers, & Johnsen, 1993). The temporal relationship between maltreatment and engagement in risk behaviors is rarely examined, making it difficult to examine causal relationships. Furthermore, behaviors are often inferred from outcomes; for example, delinquent behaviors are inferred from arrest records, and risky sexual behavior is inferred from pregnancy. While such inferences are probably correct, sole reliance on official records may not capture a complete picture of risk behaviors, as many youth who engage in such behaviors either avoid negative consequences or escape detection (Johnson, Wish, Schmeidler, & Huizinga, 1991; Zingraff et al., 1993).
The current investigation was designed to reduce these problems. First, the study was longitudinal, thereby avoiding problems associated with retrospective reports of maltreatment. In this study, the occurrence of maltreatment for each subject was court-substantiated, and in fact, severe enough to warrant removing the child from the home. Another strength of the current study was that it examined a cohort of foster children at two distinct timepoints, rather than using cross-sectional methods. Third, the determination of involvement in risk behaviors did not solely rely on official records or specific outcomes. Finally, this study was able to examine whether some early psychosocial factors, measured shortly after the youth were placed in foster care, predicted differential engagement in risk behaviors in adolescence after controlling for demographic factors, type of maltreatment, Time 1 behavior problems and cognitive and adaptive behavioral functioning.
This study was guided by a working model in which: (1) risk behaviors (the dependent variables) were conceptualized and examined both as four distinct domains (i.e., Delinquent, Self-destructive, Substance Use, and Sexual Risk behaviors) and as a unitary construct (i.e., a Total Risk Behavior Scale), (2) the control variables consisted of demographic factors (i.e., age, gender, ethnicity), maltreatment type(s), Time 1 behavior problems, and cognitive and adaptive behavioral functioning, and (3) the psychosocial predictor variables included distinct domains of social support and self-perception, as described below.
The psychosocial predictor variables of social support and self-perception have theoretical relevance to the risk behavior construct, yet these variables have not been examined in a longitudinal study of maltreated youth in foster care. In addition, because risk behavior, as a unitary construct, has not been studied before in maltreated youth in foster care, there is little precedence for building predictor models. For this reason, the current study was somewhat exploratory, initially examining the bivariate relationships between the psychosocial variables and the risk behaviors and, secondly, exploring whether the psychosocial variables, as a group, accounted for a significant amount of variance over and above the control variables. The goal of the current study was to identify whether there were any factors, measured shortly after placement in foster care, that would be predictive of risk behaviors in adolescence. The psychosocial variables were conceptualized as being salient for this investigation (relative to the control variables), because intervention efforts may maximally impact these areas. For the most part, the control variables in this study were variables that are fixed or generally stable (i.e., age, gender, ethnicity, maltreatment type, cognitive and adaptive behavioral functioning). Because early behavior problems are predictive of later behavior problems (Koot & Verhulst, 1992; Lanz, 1995, Loeber, 1990; Runyan & Gould, 1985; White, 1992, Widom, 1991a, Widom, 1991c), Time 1 behavior problems were also controlled for this study in order to examine the specific influence of the psychosocial predictor variables.
The two central questions of the current study were:
- 1.
Which Time 1 control and psychosocial predictor variables are related to engagement in Time 2 risk behaviors?
- 2.
Does the group of psychosocial predictor variables from Time 1 predict engagement in adolescent risk behaviors, over and above the control variables?
Section snippets
Participants
The flowchart in Figure 1 delineates the recruitment process at both timepoints as well as the sample for the current study.
Time 1
Participants for the current study were recruited from a cohort of youth who previously participated in the Screening Impact on Services and Costs for Foster Children (SISC) study (Landsverk, Litrownik, Newton, Ganger, & Remmer, 1996). Children were recruited for the SISC (Time 1) study after they were removed from their homes and placed in a Southern California county’s
Correlations between Time 1 control and predictor variables and Time 2 risk behaviors
As shown in Table 1, age was positively correlated with Substance Use, Sexual behaviors, and Total Risk behaviors. Gender was not significantly related to any of the dependent variables, although there was a trend (p=.06), for males to engage in more Delinquent behaviors. In examining the two ethnicity variables, African American youth were less likely (relative to Caucasian and Hispanic youth) to engage in Self-Destructive behaviors, and Caucasian youth were less likely to engage in Delinquent
Discussion
Literature on resiliency has brought to the forefront the importance of examining vulnerability and protective factors in an effort to improve intervention and prevention efforts based on empirical findings. This longitudinal study contributes to the existing literature by examining factors associated with adolescent risk behaviors within a sample of maltreated youth placed in foster care. As a group, the Time 1 control and psychosocial predictor variables were significant predictors of Time 2
Acknowledgements
The author wishes to express her appreciation to the project interviewers, especially Stephanie Brown, members of her dissertation committee (Alan Litrownik, John Landsverk, Joseph Price, and Ann Garland), and to David Olds and Sara Culhane for their comments on earlier drafts of the manuscript. A special thank you to all of the youth who made this work possible.
References (104)
- et al.
Factors associated with an increased impact of child sexual abuse
Child Abuse & Neglect
(1987) - et al.
Type of maltreatment as a predictor of mental health service use for children in foster care
Child Abuse & Neglect
(1996) - et al.
The medical and psychosocial needs of children entering foster care
Child Abuse & Neglect
(1987) - et al.
Physical abuse in Canadian runaway adolescents
Child Abuse & Neglect
(1995) - et al.
Survival responses to incest: Adolescents in crisis
Child Abuse & Neglect
(1985) Development and risk factors of juvenile antisocial behavior and delinquency
Clinical Psychology Review
(1990)- et al.
Suicidal behavior and sexual/physical abuse among street youth
Child Abuse & Neglect
(1998) - et al.
Child sexual abuse and premarital intercourse among high-risk adolescents
Journal of Adolescent Health Care
(1990) - et al.
Multicultural perspectives on suicide
Aggression and Violent Behavior
(1999) - et al.
Early psychosocial risks for adolescent suicidal ideation and attempts
Journal of the American Academy of Child and Adolescent Psychiatry
(1995)
Health risk behaviors and attempted suicide in adolescents who report prior maltreatment
The Journal of Pediatrics
Sexual abuse and psychopathology in hospitalized adolescents
Journal of the American Academy of Child and Adolescent Psychiatry
Sexual child abuse as an antecedent to prostitution
Child Abuse & Neglect
The long-term sequelae of child and adolescent abuse: A longitudinal community study
Child Abuse & Neglect
Health risk behaviors and medical sequelae of childhood sexual abuse
Mayo Clinic Proceedings
Discovering physical abuse: Insights from a follow-up study of delinquents
Child Abuse & Neglect
Physical and sexual abuse in runaway youths
Child Abuse & Neglect
The relationship of sexual abuse and HIV risk behaviors among heterosexual adult female STD patients
Child Abuse & Neglect
Application of attachment theory to the study of sexual abuse
Journal of Consulting and Clinical Psychology
The development of problem behavior syndromes in at-risk adolescents
Development and Psychopathology
Parent and peer attachment in early adolescent depression
Journal of Abnormal Child Psychology
Involvement in multiple problem behaviors of young urban adolescents
Journal of Primary Prevention
Mediation of abusive childhood experiences: Depression, dissociation, and negative life outcomes
American Journal of Orthopsychiatry
Use of public mental health system by children in foster care: Client characteristics and service use patterns
Journal of Mental Health Administration
Childhood sexual abuse as a precursor to depression and self-destructive behavior in adulthood
Journal of Traumatic Stress
Sexual abuse as a factor in adolescent pregnancy and child abuse
Family Planning Perspectives
Abused to abuser: Antecedents of socially deviant behaviors
American Journal of Psychiatry
Rethinking teenage childbearing: Is sexual abuse a missing link
Family Relations
Social support in high-risk adolescents: Structural components and adaptive impact
American Journal of Community Psychology
Mental health problems of children in foster care
Journal of Child and Family Studies
The relationship between physical and sexual abuse and tobacco, alcohol, and illicit drug use among youths in a juvenile detention center
The International Journal of the Addictions
Physical abuse, sexual victimization, and illicit drug use: Replication of a structural analysis among a new sample of high-risk youths
Violence and Victims
Juvenile crime and drug abuse: A prospective study of high risk youth
Journal of Addictive Diseases
Biological and genetic contributors to violence—Widom’s untold tale
Psychological Bulletin
Family violence
American Psychologist
Are there any successful men from criminogenic backgrounds
Psychiatry
Stress, coping, and high-risk sexual behavior
Health Psychology
Self-destructive behavior in battered children
American Journal of Psychiatry
The role of attachment in the early development of disruptive behavior problem
Development and Psychopathology
Cited by (0)
- ☆
This research was conducted at the Child and Adolescent Services Research Center, San Diego, CA, and the San Diego State University/University of California, San Diego Joint Doctoral Program in Clinical Psychology. The research reported herein was based on a portion of the author’s doctoral dissertation. It was supported by grants from NIMH (1 R03 MH56781 and 1 K01 MH01972, H. Taussig, PI) as well as grants 1 R01 MH46078, NIMH, and 90-CA-1458, NCCAN (J. Landsverk, PI).