G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS Child Abuse & Neglect xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Child Abuse & Neglect
Mental health correlates of victimization classes among homeless youth Kimberly Bender a,∗ , Kristin Ferguson b , Sanna Thompson c , Lisa Langenderfer a a b c
Graduate School of Social Work, University of Denver, 2148 South High Street, Denver, CO 80208, USA Silberman School of Social Work at Hunter College, The City University of New York, 2180 Third Avenue, New York, NY 10035, USA School of Social Work, University of Texas at Austin, 1717 West 6th Street, Suite 295, Austin, TX 78703, USA
a r t i c l e
i n f o
Article history: Received 22 November 2013 Received in revised form 26 February 2014 Accepted 3 March 2014 Available online xxx
Keywords: Victimization Homeless youth Mental health Depression Posttraumatic stress disorder Latent class
a b s t r a c t Literature reports high rates of street victimization among homeless youth and recognizes psychiatric symptoms associated with such victimization. Few studies have investigated the existence of victimization classes that differ in type and frequency of victimization and how youth in such classes differ in psychiatric profiles. We used latent class analysis (LCA) to examine whether classes of homeless youth, based on both type and frequency of victimization experiences, differ in rates of meeting diagnostic criteria for major depressive episodes and posttraumatic stress disorder (PTSD) in a sample of homeless youth (N = 601) from three regions of the United States. Results suggest youth who experience high levels of direct and indirect victimization (high-victimization class) share similarly high rates of depressive episodes and PTSD as youth who experience primarily indirect victimization only (witness class). Rates of meeting criteria for depressive episodes and PTSD were nearly two and three times greater, respectively, among the high victimization and witness classes compared to youth who never or rarely experienced victimization. Findings suggest the need for screening and intervention for homeless youth who report direct and indirect victimization and youth who report indirect victimization only, while prevention efforts may be more relevant for youth who report limited victimization experience. © 2014 Elsevier Ltd. All rights reserved.
Introduction Street victimization is a common occurrence among homeless youth (Stewart et al., 2004; Tyler & Beal, 2010). This victimization is correlated with a host of negative consequences for the youth, notably poorer mental health (Whitbeck, Hoyt, Johnson, & Chen, 2007). Specifically, previous research has found that street victimization is associated with substance abuse (Hoyt, Ryan, & Cauce, 1999; Whitbeck, Hoyt, & Bao, 2000), self-mutilation (Tyler, Whitbeck, Hoyt, & Johnson, 2003), depressive symptoms (Whitbeck et al., 2000), and the development of posttraumatic stress disorder (PTSD; Whitbeck et al., 2007). Despite the pervasiveness of street victimization and its consequences, there is limited research on the existence of subgroups of youth based on the type and frequency of their victimization and how such victimization profiles may lead to differential assessment and treatment protocols for homeless youth. As such, the present study examined psychiatric symptoms of major depressive episode and posttraumatic stress disorder to determine their association with homeless youth victimization profiles.
∗ Corresponding author. http://dx.doi.org/10.1016/j.chiabu.2014.03.001 0145-2134/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
2
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Background Literature Many homeless youth report histories of victimization prior to living on the streets; previous research has shown that 47% of homeless youth experience physical abuse in the home, and 29% experience sexual abuse (Tyler & Cauce, 2002). This maltreatment is often cited as the reason for leaving home (Lindsey, Kurtz, Jarvis, Williams, & Nackerud, 2000; Rosenthal, Mallett, & Myers, 2006) and is associated with greater subsequent victimization once on the streets (Slesnick, Erdem, Collins, Patton, & Buettner, 2010; Thrane, Hoyt, Whitbeck, & Yoder, 2006). Homeless youth frequently experience direct forms of abuse (Stewart et al., 2004), which has been defined as verbal, emotional, physical, and sexual forms of victimization (Ferguson, 2009). On the streets, it has been estimated that 83% of these youth experience direct physical and/or sexual victimization, such as being beaten up, assaulted with a weapon, or raped (Stewart et al., 2004). Youth also endure indirect traumatic events, such as witnessing the victimization of another, experiencing threats of violence, or losing a loved one (Bender, Thompson, Ferguson, & Langenderfer, in press; Bender, Thompson, Ferguson & Pollio, in press; Ferguson, 2009). Approximately 72% of homeless youth report having witnessed a physical attack, 20% having seen someone killed, and 16% having seen someone being sexually assaulted since becoming homeless (Kipke, Simon, Montgomery, Unger, & Iversen, 1997). Several factors have been identified as increasing risk for victimization among homeless youth. Youth who meet criteria for an alcohol use disorder are more likely to experience traumatic events, such as direct and indirect victimization, on the streets (Bender et al., 2010). In addition, transience, or movement from one city to the next, has been shown to exacerbate victimization risk, as it disrupts building and maintaining a network of resources and increases vulnerability in unfamiliar environments (Bender et al., 2010). Youth who spend more time on the streets, run away at earlier ages, and report multiple episodes of running away are more likely to experience physical victimization, while youth who participate in survival behaviors (e.g., prostitution, panhandling) are at increased risk for sexual victimization (Tyler & Beal, 2010). Much of the extant research on homeless youth has focused on singular types of victimization, such as experiencing sexual or physical abuse; however, given the high prevalence of both direct (Stewart et al., 2004) and indirect (Kipke et al., 1997) victimization among homeless youth, the concept of poly-victimization has emerged as a framework for investigating how homeless youth in particular experience victimization (Ferguson, 2009). Poly-victimization has been defined as having experienced more than one type of victimization (e.g., sexual abuse, bullying, witnessing familial violence), emphasizing experiencing multiple types rather than multiple occurrences of victimization (Finkelhor, Turner, Hamby, & Ormrod, 2011). There is no consensus over what constitutes a poly-victim (Finkelhor et al., 2011), but authors of one study considered youth who experienced four or more types of victimization to be poly-victims (Finkelhor, Ormrod, & Turner, 2007a). By this standard, 22% of children in a nationally representative sample were considered poly-victims in the previous year (Finkelhor et al., 2007a) and had higher rates of anxiety, depression, and trauma compared to those never experiencing abuse or had a single experience (Finkelhor et al., 2007a). Moreover, research has shown poly-victims experience more serious and injurious victimization (Finkelhor et al., 2011) and are at increased risk for subsequent victimization (Finkelhor, Ormrod, & Turner, 2007b). Experiencing multiple forms of victimization has not been well studied among homeless youth, though limited existing research shows multiple forms of abuse may negatively affect psychosocial and behavioral outcomes (Ferguson, 2009). Specifically, in her examination of types of abuse experienced (physical abuse, sexual, abuse, emotional abuse, psychological abuse, verbal abuse, and witnessing familial abuse). Ferguson (2009) found that homeless youth who reported experiencing multiple forms of abuse (i.e., both direct and indirect) demonstrated the most psychological symptoms (Ferguson, 2009). Although Ferguson’s (2009) study is limited by its small convenience sample, it provides a foundation for applying the concept of poly-victimization to homeless youth and introduces the notion of direct and indirect victimization as one conceptualization of poly-victimization and its effects. Most research on victimization among homeless youth has not investigated, in detail, the combinations of types and frequencies of victimization experienced. Instead, studies report prevalence estimates that suggest high overall rates (see Stewart et al., 2004). High rates, although concerning, do not differentiate subgroups of homeless youth who may differ in their service needs. Service agencies for homeless youth, without a more nuanced understanding of victimization profiles, may fail to address victimization or, at best, provide trauma-informed services generally. These agencies will likely fail to recognize the differential needs of youth with particularly harmful experiences. Differentiation in the mental health needs of youth with various victimization profiles will, thus, inform under-resourced service agencies regarding how to most efficiently serve victimized youth in their care. For example, agencies may identify methods of screening for victimization experiences that place youth at greatest risk for mental health problems and provide further assessment, referral, treatment, and supportive housing to those youth most in need. Recently, in a multi-site study of homeless youth, latent class analysis was conducted to identify specific categories of victimization among these youth (Bender et al., in press). Latent class analysis (LCA) allowed individuals to naturally group into classes based on the different types and frequencies of victimization they experienced. Rather than the researcher creating artificial cut points or predetermined criteria for group inclusions, the classes revealed in the LCA were naturally occurring groups of youth with similar patterns of victimization. The LCA identified three distinct victimization classes: (a) a low victimization class of youth who never or rarely experienced indirect or direct victimization; (b) a high-victimization class of youth who experienced high rates of direct and indirect victimization, typically experiencing each victimization event multiple times; and (c) a witness class of youth who remained on the periphery of direct violence but reported high rates of threats and witnessed violence, typically experiencing each victimization event only once (Bender et al., in press). This Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
3
previous work not only suggests the existence of distinct victimization profiles among homeless youth, but also finds that these subgroups differ in their substance use patterns such that youth in the low-victimization class reported significantly less substance abuse/dependence compared to witnesses and high-victims, while youth in the high-victimization group reported significantly greater substance dependence, particularly in regard to alcohol, compared to youth in the other two groups (Bender et al., in press). The purpose of the current study is to build on this previous class analysis by investigating whether these classes of victimized youth differ in their mental health needs, specifically rates of meeting diagnostic criteria for major depressive episodes and PTSD. Such investigation may inform differential screening, assessment, treatment, and prevention needs for homeless youth utilizing services at under-resourced shelters and drop in centers. Thus, two broad research questions guided our present investigation: controlling for other relevant correlates, do homeless youth in different victimization classes differ significantly in the degree to which they (a) meet criteria for having had a major depressive episode or (b) meet criteria for PTSD? Methods Sample and Recruitment A total of 601 street youth (ages 18–24) were recruited from homeless youth-serving host agencies in Los Angeles (n = 200), Denver (n = 201), and Austin (n = 200) using purposive sampling. Researchers selected host agencies based on their existing partnerships and the agency’s commitment to host the study. In each study, recruitment sites included drop-in centers offering case management, referral services, and basic subsistence items (food, hygiene supplies); shelters offering short-term (40 days) stay; and transitional housing apartments offering temporary (6 months) housing. Each investigator received human subjects’ approval from her university. Data collection occurred from March 2010 to April 2011. Recruitment procedures were nearly identical across cities with minor variations due to services emphasized in each location (e.g., more crisis-shelter users in Los Angeles, more drop-in service users in Denver and Austin). To participate in the study, youth had to meet three inclusion criteria: (a) be 18–24 years of age, (b) have spent at least 2 weeks away from home in the month before the interview (Whitbeck, 2009), and (c) provide written informed consent. Youth were excluded if they could not understand the consent form because of cognitive limitations (psychotic symptoms or developmental delays) or if they were noticeably intoxicated or high at the time of the interview. In the latter case, youth were asked to return at a later time when they could more competently answer interview questions. Agency case managers made the determination whether a particular individual was eligible for recruitment into the study based on their knowledge of each individual and their current level of intoxication. Data Collection and Measures Agency staff facilitated introductions and referred youth to research staff who explained the study procedures and secured written consent. Researchers administered a 45-minute quantitative retrospective interview containing both standardized self-report instruments and researcher-developed items that together assessed demographic and background information, victimization experiences, and psychiatric symptoms. Interviewers read questions and response options aloud to participants and youth responded verbally. Youth were compensated for their time with a $10.00 gift card to a local food vendor. Demographics. Basic demographics included age, gender (0 = male, 1 = female), and ethnicity (1 = white, 2 = black, 3 = Hispanic, 4 = other). Ethnicity was subsequently dummy coded to include Black (0 = no, 1 = yes), Latino (0 = no, 1 = yes), and other (0 = no, 1 = yes), with White as a reference category. To assess and control for inter-city differences, the city in which data were collected was recorded (1 = Los Angeles, 2 = Denver, 3 = Austin) and then dummy-coded to include Los Angeles (0 = no, 1 = yes) and Austin (0 = no, 1 = yes), with Denver as a reference category. Victimization Class. To capture the many forms of victimization homeless youth experience, the authors built on Ferguson’s (2009) definitions of direct and indirect abuse and operationalized “victimization” as experiencing any form of direct (i.e., physical, sexual, emotional, or verbal) or indirect (i.e., witnessing) abuse, or experiencing a traumatic event (e.g., sudden death of close friend or loved one). To do so, street victimization experiences were collected using a revised version of the Traumatic Life Events Questionnaire (Kubany et al., 2000). This measure assessed the frequency (0 = never, 1 = once, 2 = more than once) of experiencing different types of direct (e.g., robbery involving a weapon, physical assault by an acquaintance or stranger) and indirect victimization (e.g., sudden death of a close friend or loved one, witnessing a physical assault) since leaving home. This instrument was revised to exclude items with less relevance for this study (e.g., natural disasters, warfare or combat). Patterns in responses to these items were used to form latent victimization classes. Due to space limitations, this process is described elsewhere (Bender et al., in press) or details can be provided by contacting the first author. Fig. 1 displays the percentage of youth in each victimization class who experienced each type of victimization never, once, or more than once. Youths’ patterned responses placed them in one of three classes: (a) low victimization class (n = 215; never or rarely experienced indirect and direct victimization); (b) high-victimization class (n = 136); experienced high rates of direct and indirect victimization, typically experiencing each victimization event more than once; and (c) witness class (n = 250) included youth who remained on the periphery of direct violence but reported high rates of threats, witnessing violence, and Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
4
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Fig. 1. Percentage of youth in each class who experienced each victimization event. Adapted from Bender et al., in press.
typically experiencing each victimization event only once. The 3-category victimization class variable (1 = low-victimization class, 2 = witness class, 3 = high-victimization class) was recoded into a series of three dummy coded variables for use in multivariate analyses: low-victimization class (0 = no, 1 = yes), high-victimization class (0 = no, 1 = yes), and witness class (0 = no, 1 = yes). Homelessness Variables. Several variables associated with victimization in the extant homeless youth literature were measured, including primary living arrangement (0 = homeless, 1 = temporarily housed with relative, friend, foster parent, facility); transience or number of inter-city moves since leaving home; and length of time homeless (number of months between interview date and the date the youth last left home). Psychiatric Problems. Major depressive episode and PTSD were assessed by the Mini International Neuropsychiatry Interview (MINI; Sheehan et al., 1998). This is a widely used, brief, structured interview that facilitates screening for Axis I psychiatric disorders as outlined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV-TR; American Psychological Association, 2000). The MINI has demonstrated good reliability (Lecrubier et al., 1997) as well as convergent validity with the Structured Clinical Interview for DSM-IV-TR Axis 1 Disorders (SCID), an established measure of diagnostic criteria (Sheehan et al., 1998). In assessing whether youth met criteria for a major depressive episode, respondents were asked two screening questions regarding feeling down and losing interest during the past 2 weeks. Those who responded affirmatively were then asked a series of questions regarding symptoms, such as changes in appetite, sleep patterns, energy level, concentration, and suicidal thoughts experienced during the past 2 weeks. Participants who reported 5 or more of the 9 symptoms listed were coded positive for meeting criteria for a current major depressive episode (0 = no, 1 = yes). The MINI was also used to assess whether youth met criteria for PTSD. Respondents were first asked three screening questions regarding: (1) whether or not they had ever experienced or witnessed an extremely traumatic event that included actual or threatened death or serious injury to them or someone else, (2) whether they responded with intense fear, helplessness, or horror, and (3) whether, during the past month, they re-experienced the event in a distressing way (dreams, intense recollections, flashbacks, or physical reactions). Those who answered affirmatively to all 3 screeners were then asked 13 additional questions reflecting DSM-IV symptoms of re-experiencing, avoidance, and arousal symptoms during the past month. Those who responded affirmatively to three or more re-experiencing and avoidance symptoms and two or more arousal symptoms, and confirmed that these symptoms had been significantly interfering with their lives within the past month were categorized as meeting criteria for PTSD (0 = no, 1 = yes). Data Analysis Univariate analyses were used to describe the full sample’s demographic and psychiatric characteristics. Bivariate analyses (chi square, ANOVA) examined differences in sample characteristics across data collection sites. Utilizing two binary logistic regression analyses, first PTSD, then depression, were regressed on demographic variables (age, gender, ethnicity, city), homelessness variables (primary living situation, transience, time homeless) and dummy coded victimization Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model
ARTICLE IN PRESS
CHIABU-2746; No. of Pages 8
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
5
class assignment variables (high-victimization class and witness class with low-victimization class as the reference category). These models examined whether victimization class membership was associated with meeting criteria for a major depressive episode and PTSD, controlling for other correlates, and these initial models compared youth in the witness and high-victimization classes to youth in the low-victimization class reference category. These models were then re-analyzed in order to compare the effects of high-victimization class membership to witness class membership on depression and PTSD by changing the dummy coded class assignment variables to include low-victimization class and witness class, leaving high-victimization class as the reference category. Results Sample Characteristics Sample characteristics are described in Table 1, including demographic variables, correlates, and victimization class assignment for the full sample (N = 601) and for each subsample by data collection site (Los Angeles, Denver, and Austin). Because of differences on some variables across sites, city was controlled for in the multivariate analysis. Correlates of Class Membership To address the study’s two research questions, a series of binary logistic regression analyses were conducted. First, depression (model 1) and then PTSD (model 2) were regressed on victimization class assignment variables (high-victimization class and witness class with low-victimization class as the reference category), controlling for demographic (age, gender, ethnicity, city) and homelessness variables (primary living situation, transience, time homeless). Results for these models are displayed in Table 2. Model results indicated that, compared to low-victims (i.e., youth who experienced no or very little indirect or direct victimization), witnesses (i.e., youth experiencing indirect victimization) and high-victims (i.e., youth experiencing indirect and direct victimization on multiple occasions) were more likely to meet criteria for a major depressive episode and PTSD. Specifically, in regards to depression, 2 [12] = 34.37, p < .001, compared to low-victims, witnesses were over 2 times more likely (OR = 2.34, p < .001) and high-victims were over 2 times more likely (OR = 2.12, p < .01) to meet criteria for a major depressive episode. Similarly, in regards to PTSD, 2 [12] = 59.46, p < .001, witnesses were 3 times more likely (OR = 3.16, p < .001) and high-victims were 3 times more likely (OR = 3.39, p < .001) to meet criteria for PTSD compared to low-victims.
Table 1 Full and city-level sample characteristics of homeless young adults. Full sample (N = 601)
Gender Male Female Race/ethnicity White Black Latino Other Primary living situation Streets Temp housing Meet diagnostic criteria Major depressive episode PTSD Co-occurring depression and PTSD Class assignment Low-victimization High-victimization Witness Mean Age Number inter-city moves Months homeless ***
20.5 3.5 32.4
Denver (n = 201)
Los Angeles (n = 200) n
%
n
Austin (n = 200)
n
%
%
n
385 216
64.1 35.9
128 72
64.0 36.0
129 72
64.2 35.8
128 72
64.0 36.0
240 152 107 101
39.9 25.3 17.8 16.8
18 92 52 38
9.0 46.0 26.0 19.0
70 54 37 39
34.8 26.9 18.4 19.4
152 6 18 24
76.0 3.0 9.0 12.0
303 298
50.4 49.6
87 113
43.5 56.5
90 111
44.8 55.2
126 74
63.0 37.0
187 136 79
31.1 22.7 13.2
62 65 32
31 32.5 16
73 39 28
36.3 19.4 13.9
52 32 19
26.4 16.2 9.7
215 136 250
35.8 22.6 41.6
95 24 81
47.5 12.0 40.5
70 42 89
34.8 20.9 44.3
50 70 80
25.0 35.0 40
Chi-square/ ANOVA
% .002
2 = 207.9***
2 = 19.1***
SD 1.6 3.7 31.0
Mean 19.4 2.6 22.9
SD 1.1 3.2 26.0
Mean 19.9 1.9 31.6
SD 1.6 2.8 29.6
Mean 20.8 5.9 42.8
SD 1.7 3.7 33.9
2 = 4.6 2 = 17.0*** 2 = 3.56 2 = 38.5***
ANOVA F = 40.6*** F = 85.3*** F = 22.1***
p < .001.
Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
6
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Table 2 Logistic regression models predicting depression and PTSD. Depression model 1
Age Gender (male) Ethnicity (Black) Ethnicity (Latino) Ethnicity (White) Transience Time homeless Primary living situation (streets) Los Angeles (other cities) Austin (other cities) Witness class (low-victimization reference category) High-victimization class (low-victimization reference category) Witness class (high-victimization reference category)
PTSD model 2
Depression model 3
PTSD model 4
OR
CI
OR
CI
OR
CI
OR
1.05 1.38 73 1.23 92 96 1.00 82 95 60 2.34***
.92–1.20 .95–2.00 .40–1.32 .67–2.24 .52–1.60 .90–1.02 .99–1.01 .57–1.19 .60–1.50 .36–1.02 1.52–3.61
1.00 1.98** .62 .84 .61 .98 1.00 .82 2.53*** .77 3.16***
86–1.17 1.30–3.00 33–1.18 44–1.61 33–1.14 91–1.05 1.00–1.01 54–1.25 1.50–4.25 41–1.45 1.89–5.28
1.05 1.38 73 1.23 92 96 1.00 82 95 60
92–1.20 95–2.00 40–1.32 67–2.24 52–1.60 90–1.02 99–1.01 57–1.19 60–1.50 36–1.02
1.00 1.98** .62 .84 .61 .98 1.00 .82 2.53*** .77
.86–1.17 1.30–3.00 .33–1.18 .44–1.61 .33–1.14 .91–1.05 1.00–1.01 .54–1.25 1.50–4.25 .41–1.45
2.12**
1.24–3.65
3.39***
1.79–6.42
1.10
69–1.77
.93
.55–1.58
2 (12) = 34.37, p < .001
2 (12) = 59.46, p < .001
2 (12) = 34.37, p < .001
CI
2 (12) = 59.46, p < .001
Note. Shaded region displays results comparing witness class to high-victimization class as predictors of depression and PTSD; OR = odds ratio; 95% CI = 95% confidence interval. ** p < .01. *** p < .001.
Among control variables, gender (youth who were female; OR = 1.98, p < .001) and data collection site (youth who participated in the study in LA; OR = 2.53, p < .001) increased odds of meeting criteria for PTSD; no control variables predicted depression. For a final analysis, we included witness class but exchanged the high-victimization class variable for the lowvictimization class variable (making high-victimization class the new reference category) and predicted depression (model 3) and PTSD (model 4), controlling for the same demographic and homelessness control variables. The effects of all control variables in model 3 and 4 did not change, but this allowed us to compare youth in the witness class to youth in the high-victimization class. Compared to youth in the high-victimization class, youth in the witness class were not significantly more or less likely to meet criteria for depression (OR = 1.10, p = .887) or PTSD (OR = .930, p = .789). These results appear in the shaded column of Table 2. Discussion The current study represents the first known investigation of whether homeless youth victimization profiles differ significantly in psychiatric problems, specifically rates of meeting criteria for major depressive episode and PTSD. Victimization profiles were derived from our previously conducted latent class analysis (Bender et al., in press), which identified three distinct victimization classes: (a) youth in a low-victimization class never or rarely experienced either direct or indirect forms of victimization; (b) youth in a high-victimization class reported higher rates of experiencing both direct and indirect victimization, often experiencing each victimization event more than once; and (c) youth in a witness class experienced elevated rates of indirect trauma rather than their own direct victimization, typically experiencing each victimization event only once. The results of the current study suggest that homeless youth in the high-victimization and witness classes were significantly more likely to meet criteria for a major depressive episode and for PTSD than youth in the low-victimization class. Thus, those homeless youth who experience victimization, whether direct and indirect, or indirect only, are likely to be at higher risk for major depressive episodes and PTSD than those youth who experience no or low rates of victimization. Our findings support previous literature indicating high rates of homeless youth victimization (e.g., Tyler & Beal, 2010) and an association between said victimization and mental health problems (e.g., Whitbeck et al., 2000, 2007). However, this study uniquely contributes an understanding that homeless youth do not represent a heterogeneous population with universally high victimization exposure and psychiatric needs, but rather a collection of subgroups of youth with varying levels of exposure and psychiatric symptoms. Recognizing subgroups with differential treatment needs suggests under-resourced service agencies may need to develop screening protocols that differentiate high-risk subgroups in need of referral and more costly treatment from low-risk subgroups who may benefit from less costly efforts to prevent victimization and associated psychiatric outcomes. That more than one-third (n = 215) of this sample was classified as low-victims suggests lower-cost prevention efforts should be a major emphasis in service settings.‘ Of note, results demonstrate that homeless youth exposed to indirect victimization (witness class) are at similarly high risk for psychiatric problems as youth exposed to a combination of direct and indirect victimization (high-victimization class). Previous research has demonstrated similar findings of the deleterious effects of indirect trauma on youths’ mental Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
7
health in a general context. For example, using the National Survey of Adolescents, direct victimization (i.e., physical assault, sexual assault) and indirect victimization (i.e., witnessing violence) were found to significantly predict mental health outcomes, including comorbid diagnosis of PTSD and depression (Kilpatrick et al., 2003). High rates of mental health diagnoses, specifically PTSD, have been found in other adolescent samples that have witnessed violence, particularly children exposed to domestic violence (Kilpatrick & Williams, 1997; Luthra et al., 2009). Specifically, in a small sample of children exposed to domestic violence, only one of 20 children did not meet criteria for a PTSD diagnosis ranging from mild to severe (Kilpatrick & Williams, 1997). Luthra et al.’s (2009) examination of various types of adolescent trauma exposure similarly demonstrated a correlation between witnessing violence and PTSD diagnosis. Our study adds to this literature, indicating that indirect forms of victimization can have serious mental health consequences for homeless youth – a population with elevated exposure to dangerous street situations – and suggests that even those homeless youth who are on the periphery of violence are likely in need of mental health services.
Limitations Study limitations should be considered in interpreting our findings. The cross-sectional design used in this study limits assumptions of causal order; thus, it may be that youth who meet criteria for depressive episodes or PTSD are placed at greater risk for victimization. Longitudinal research, although challenging with this population, is necessary to determine the causal order or the relationship discussed here. Furthermore, our sample of service-seeking youth limits the generalizability of our results to youth disconnected from services – a subpopulation with potentially greater risk for victimization due to disconnection from formal/institutional supports. In addition, social desirability bias may have prevented youth from sharing sensitive information about victimization and psychiatric symptoms. In attempts to reduce this bias, our interviewers received extensive training in building rapport with homeless youth as well as stressing confidentiality and assuring privacy during interviews. Nonetheless, rates of victimization and psychiatric symptoms may have been underreported. Conversely, with few existing measures of psychiatric problems normed for homeless youth, it is possible that our measure of PTSD could have inflated rates slightly given some homeless youths’ protective hyper-vigilance in a street environment. Finally, although a strength of this study was the use of a large multi-site dataset, this also created challenges. For example, data collection sites varied in rates of PTSD. While we controlled for data collection site in all analyses, this site level difference deserves further investigation, including whether mental health service provisions differ significantly across cities.
Implications Screening for victimization experiences and associated psychiatric symptoms should be systematically conducted across homeless youth service agencies. Our study suggests rates for trauma, PTSD, and depressive episodes are high; thus, greater understanding of the degree and type of trauma experienced is critical to providing adequate assessment and treatment. Screening should include questions about youths’ experiences of various forms of direct and indirect victimization. Youth who screen positive for experiencing either (or both) forms of victimization should be engaged in mental health services if amenable. Service providers should take seriously experiences of indirect victimization, as it appears witnessing violence and experiencing threats affect youths’ psychological health. Engaging youth in mental health treatment following positive screening for victimization and associated psychiatric symptoms may prove challenging. Research finds that although 84% of youth have histories of victimization, only 44% report an interest in treatment to address these issues (Keeshin & Campbell, 2011). While some youth simply do not see the need for treatment, believing instead that their struggles make them stronger, others describe past treatment experiences that were ineffective as reasons not to seek further help (Keeshin & Campbell, 2011). Careful consideration should be given to when and in what contexts trauma treatment is provided. Intensive approaches to processing previous trauma, including exposure to trauma stimuli, may not be appropriate for youth living on the streets or in other unstable contexts where they are consistently confronted with victimization risk. Hyper-vigilance associated with previous victimization, for example, may be protective for those currently living in dangerous situations (McManus & Thompson, 2008). For these street-living youth, interventions should be developed and tested that teach self-protection and safety in short-term shelter and drop-in centers. In contrast, trauma treatments should be tested with youth who have found safe/stable living environments off the streets in long-term transitional or permanent housing. These are likely to be the most appropriate venues for intervention and allow a determination of whether such interventions are effective with this highly vulnerable group (Thompson, McManus, & Voss, 2006). Although treatment is recommended for youth with high-victimization and witness profiles, a large subsample of youth reported little victimization and low rates of psychiatric symptoms. For this group, victimization prevention programming is warranted. Skills-based programs aimed at preventing high-risk behaviors (e.g., substance use and sexual risk behaviors) have demonstrated effectiveness with homeless youth (Rotheram-Borus et al., 2003). Similar programs should be developed to integrate safety skills, helping youth to identify dangerous situations and problem-solve assertively to avoid victimization. Such programs may be best provided in short-term homeless youth shelters where staff may address victimization risk factors before youth leave shelters for the streets where their vulnerability is heightened. Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
8
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Homeless youth encounter multiple forms of victimization on the streets. However, their experiences and resulting mental health needs vary considerably. Assessing for youths’ victimization experiences and providing appropriate prevention or treatment programming is critical to helping youth remain safe on the streets and cope with past traumatic experiences. References American Psychological Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed. text rev.). Washington, DC: American Psychiatric Association. Bender, K., Thompson, S., Ferguson, K., & Langenderfer, L. (in press) Substance use predictors of victimization profiles among homeless youth: A latent class analysis. Journal of Adolescence. Bender, K., Ferguson, K., Thompson, S. J., Komlo, C., & Pollio, D. (2010). Factors associated with trauma and posttraumatic stress disorder among homeless youth in three U.S. cities: The importance of transience. Journal of Traumatic Stress, 23, 161–168. http://dx.doi.org/10.1002/jts.20501 Ferguson, K. (2009). Exploring the psychosocial and behavioral outcomes of multi-type abuse among homeless young adults. Social Work Research, 33, 219–230. http://dx.doi.org/10.1093/swr/33.4.219 Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2007a). Poly-victimization: A neglected component in child victimization. Child Abuse & Neglect, 31, 7–26. http://dx.doi.org/10.1016/j.chiabu.2006.06.008 Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2007b). Re-victimization patterns in a national longitudinal sample of children and youth. Child Abuse & Neglect, 31, 479–502. http://dx.doi.org/10.1016/j.chiabu.2006.03.012 Finkelhor, D., Turner, H. A., Hamby, S., & Ormrod, R. (2011). Polyvictmization: Children’s exposure to multiple types of violence, crime, and abuse. Juvenile Justice Bulletin.,. Retrieved from: https://www.ncjrs.gov/pdffiles1/ojjdp/235504.pdf Hoyt, D. R., Ryan, K. D., & Cauce, A. M. (1999). Personal victimization in a high-risk environment: Homeless and runaway adolescents. Journal of Research in Crime and Delinquency, 36, 371–392. http://dx.doi.org/10.1177/0022427899036004002 Keeshin, B. R., & Campbell, K. (2011). Screening homeless youth for histories of abuse: Prevalence, enduring effects, and interest in treatment. Child Abuse & Neglect, 35, 401–407. http://dx.doi.org/10.1016/j.chiabu.2011.01.015 Kilpatrick, D. G., Ruggiero, K. J., Acierno, R., Saunders, B. E., Resnick, H. S., & Best, C. L. (2003). Violence and risk of PTSD, major depression, substance abuse/dependence, and comorbidity: Results from the National Survey of Adolescents. Journal of Consulting and Clinical Psychology, 71, 692–700. http://dx.doi.org/10.1037/0022-006X.71.4.692 Kilpatrick, K. L., & Williams, L. M. (1997). Post-traumatic stress disorder in child witnesses to domestic violence. American Journal of Orthopsychiatry, 67, 639–644. http://dx.doi.org/10.1037/h0080261 Kipke, M. D., Simon, T. R., Montgomery, S. B., Unger, J. B., & Iversen, E. F. (1997). Homeless youth and their exposure to and involvement in violence while living on the streets. Journal of Adolescent Health, 20, 360–367. http://dx.doi.org/10.1016/S1054-139X(97)00037-2 Kubany, E. S., Leisen, M. B., Kaplan, A. S., Watson, S. B., Haynes, S. N., Owens, J. A., & Burns, K. (2000). Development and preliminary validation of a brief-broad-spectrum measure of trauma exposure: The Traumatic Life Events Questionnaire. Psychological Assessment, 12, 210–224. http://dx.doi.org/10.1037/1040-3590.12.2.210 Lecrubier, Y., Sheehan, D., Weiller, E., Amorim, P., Bonora, I., Sheehan, K., Sheehan, K. H., & Dunbar, G. (1997). The MINI International Neuropsychiatric Interview. A short diagnostic structured interview: Reliability and validity according to the CIDI. European Psychiatry, 12, 224–231. http://dx.doi.org/10.1016/S0924-9338(97)83296-8 Lindsey, E. W., Kurtz, P. D., Jarvis, S., Williams, N. R., & Nackerud, L. (2000). How runaway and homeless youth navigate troubled waters: Personal strengths and resources. Child and Adolescent Social Work Journal, 17, 115–141. Luthra, R., Abramovitz, R., Greenberg, R., Schoor, A., Newcorn, J., Schmeidler, J., Levine, P., & Chemtob, C. M. (2009). Relationship between type of trauma exposure and posttraumatic stress disorder among urban children and adolescents. Journal of Interpersonal Violence, 24, 1919–1927. http://dx.doi.org/10.1177/0886260508325494 McManus, H. H., & Thompson, S. J. (2008). Trauma among unaccompanied homeless youth: The integration of street culture into a model of intervention. Journal of Aggression, Maltreatment & Trauma, 16, 92–109. http://dx.doi.org/10.1080/10926770801920818 Rosenthal, D., Mallett, S., & Myers, P. (2006). Why do homeless young people leave home? Australian and New Zealand Journal of Public Health, 30, 281–285. http://dx.doi.org/10.1111/j.1467-842X.2006.tb00872.x Rotheram-Borus, M., Song, J., Gwadz, M., Lee, M., Van Rossem, R., & Koopman, C. (2003). Reductions in HIV risk among runaway youth. Prevention Science, 4, 173–187. http://dx.doi.org/10.1023/A:1024697706033 Sheehan, D., Lecrubier, Y., Sheehan, K., Amorim, P., Janavs, J., Weiller, E., & Dunbar, G. (1998). The Mini-International Neuropsychiatric Interview (MINI): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59, 22–33. Slesnick, N., Erdem, G., Collins, J., Patton, R., & Buettner, C. (2010). Prevalence of intimate partner violence reported by homeless youth in Columbus, Ohio. Journal of Interpersonal Violence, 25, 1579–1593. http://dx.doi.org/10.1177/0886260509354590 Stewart, A. J., Steiman, M., Cauce, A. M., Cochran, B. N., Whitbeck, L. B., & Hoyt, D. R. (2004). Victimization and posttraumatic stress disorder among homeless adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 325–331. http://dx.doi.org/10.1097/00004583-200403000-00015 Thompson, S. J., McManus, H. H., & Voss, T. (2006). PTSD and substance abuse among youth who are homeless: Treatment issues and implications. Brief Treatment and Crisis Intervention, 6, 1–12. Thrane, L. E., Hoyt, D. R., Whitbeck, L. B., & Yoder, K. A. (2006). Impact of family abuse on running away, deviance, and street victimization of homeless rural and urban youth. Child Abuse & Neglect, 30, 1117–1128. http://dx.doi.org/10.1016/j.chiabu.2006.03.008 Tyler, K., & Beal, M. R. (2010). The high-risk environment of homeless young adults: Consequences for physical and sexual victimization. Violence and Victims, 25, 101–115. http://dx.doi.org/10.1891/0886-6708.25.1.101 Tyler, K. A., & Cauce, A. M. (2002). Perpetrators of early physical and sexual abuse among homeless and runaway adolescents. Child Abuse & Neglect, 26, 1261–1274. http://dx.doi.org/10.1016/S0145-2134(02)00413-1 Tyler, K. A., Whitbeck, L. B., Hoyt, D. R., & Johnson, K. D. (2003). Self-mutilation and homeless youth: The role of family abuse, street experiences, and mental disorders. Journal of Research on Adolescence, 13, 457–474. http://dx.doi.org/10.1046/j.1532-7795.2003.01304003.x Whitbeck, L. B. (2009). Mental health and emerging adulthood among homeless young people. New York, NY: Taylor & Francis Group. Whitbeck, L. B., Hoyt, D. R., & Bao, W. (2000). Depressive symptoms and co-occurring depressive symptoms, substance abuse, and conduct problems among runaway and homeless adolescents. Child Development, 71, 721–732. http://dx.doi.org/10.1111/1467-8624.00181 Whitbeck, L. B., Hoyt, D. R., Johnson, K. D., & Chen, X. (2007). Victimization and posttraumatic stress disorder among runaway and homeless adolescents. Violence and Victims, 22, 721–734. http://dx.doi.org/10.1891/088667007782793165
Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
ARTICLE IN PRESS Child Abuse & Neglect xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Child Abuse & Neglect
Mental health correlates of victimization classes among homeless youth Kimberly Bender a,∗ , Kristin Ferguson b , Sanna Thompson c , Lisa Langenderfer a a b c
Graduate School of Social Work, University of Denver, 2148 South High Street, Denver, CO 80208, USA Silberman School of Social Work at Hunter College, The City University of New York, 2180 Third Avenue, New York, NY 10035, USA School of Social Work, University of Texas at Austin, 1717 West 6th Street, Suite 295, Austin, TX 78703, USA
a r t i c l e
i n f o
Article history: Received 22 November 2013 Received in revised form 26 February 2014 Accepted 3 March 2014 Available online xxx
Keywords: Victimization Homeless youth Mental health Depression Posttraumatic stress disorder Latent class
a b s t r a c t Literature reports high rates of street victimization among homeless youth and recognizes psychiatric symptoms associated with such victimization. Few studies have investigated the existence of victimization classes that differ in type and frequency of victimization and how youth in such classes differ in psychiatric profiles. We used latent class analysis (LCA) to examine whether classes of homeless youth, based on both type and frequency of victimization experiences, differ in rates of meeting diagnostic criteria for major depressive episodes and posttraumatic stress disorder (PTSD) in a sample of homeless youth (N = 601) from three regions of the United States. Results suggest youth who experience high levels of direct and indirect victimization (high-victimization class) share similarly high rates of depressive episodes and PTSD as youth who experience primarily indirect victimization only (witness class). Rates of meeting criteria for depressive episodes and PTSD were nearly two and three times greater, respectively, among the high victimization and witness classes compared to youth who never or rarely experienced victimization. Findings suggest the need for screening and intervention for homeless youth who report direct and indirect victimization and youth who report indirect victimization only, while prevention efforts may be more relevant for youth who report limited victimization experience. © 2014 Elsevier Ltd. All rights reserved.
Introduction Street victimization is a common occurrence among homeless youth (Stewart et al., 2004; Tyler & Beal, 2010). This victimization is correlated with a host of negative consequences for the youth, notably poorer mental health (Whitbeck, Hoyt, Johnson, & Chen, 2007). Specifically, previous research has found that street victimization is associated with substance abuse (Hoyt, Ryan, & Cauce, 1999; Whitbeck, Hoyt, & Bao, 2000), self-mutilation (Tyler, Whitbeck, Hoyt, & Johnson, 2003), depressive symptoms (Whitbeck et al., 2000), and the development of posttraumatic stress disorder (PTSD; Whitbeck et al., 2007). Despite the pervasiveness of street victimization and its consequences, there is limited research on the existence of subgroups of youth based on the type and frequency of their victimization and how such victimization profiles may lead to differential assessment and treatment protocols for homeless youth. As such, the present study examined psychiatric symptoms of major depressive episode and posttraumatic stress disorder to determine their association with homeless youth victimization profiles.
∗ Corresponding author. http://dx.doi.org/10.1016/j.chiabu.2014.03.001 0145-2134/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
2
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Background Literature Many homeless youth report histories of victimization prior to living on the streets; previous research has shown that 47% of homeless youth experience physical abuse in the home, and 29% experience sexual abuse (Tyler & Cauce, 2002). This maltreatment is often cited as the reason for leaving home (Lindsey, Kurtz, Jarvis, Williams, & Nackerud, 2000; Rosenthal, Mallett, & Myers, 2006) and is associated with greater subsequent victimization once on the streets (Slesnick, Erdem, Collins, Patton, & Buettner, 2010; Thrane, Hoyt, Whitbeck, & Yoder, 2006). Homeless youth frequently experience direct forms of abuse (Stewart et al., 2004), which has been defined as verbal, emotional, physical, and sexual forms of victimization (Ferguson, 2009). On the streets, it has been estimated that 83% of these youth experience direct physical and/or sexual victimization, such as being beaten up, assaulted with a weapon, or raped (Stewart et al., 2004). Youth also endure indirect traumatic events, such as witnessing the victimization of another, experiencing threats of violence, or losing a loved one (Bender, Thompson, Ferguson, & Langenderfer, in press; Bender, Thompson, Ferguson & Pollio, in press; Ferguson, 2009). Approximately 72% of homeless youth report having witnessed a physical attack, 20% having seen someone killed, and 16% having seen someone being sexually assaulted since becoming homeless (Kipke, Simon, Montgomery, Unger, & Iversen, 1997). Several factors have been identified as increasing risk for victimization among homeless youth. Youth who meet criteria for an alcohol use disorder are more likely to experience traumatic events, such as direct and indirect victimization, on the streets (Bender et al., 2010). In addition, transience, or movement from one city to the next, has been shown to exacerbate victimization risk, as it disrupts building and maintaining a network of resources and increases vulnerability in unfamiliar environments (Bender et al., 2010). Youth who spend more time on the streets, run away at earlier ages, and report multiple episodes of running away are more likely to experience physical victimization, while youth who participate in survival behaviors (e.g., prostitution, panhandling) are at increased risk for sexual victimization (Tyler & Beal, 2010). Much of the extant research on homeless youth has focused on singular types of victimization, such as experiencing sexual or physical abuse; however, given the high prevalence of both direct (Stewart et al., 2004) and indirect (Kipke et al., 1997) victimization among homeless youth, the concept of poly-victimization has emerged as a framework for investigating how homeless youth in particular experience victimization (Ferguson, 2009). Poly-victimization has been defined as having experienced more than one type of victimization (e.g., sexual abuse, bullying, witnessing familial violence), emphasizing experiencing multiple types rather than multiple occurrences of victimization (Finkelhor, Turner, Hamby, & Ormrod, 2011). There is no consensus over what constitutes a poly-victim (Finkelhor et al., 2011), but authors of one study considered youth who experienced four or more types of victimization to be poly-victims (Finkelhor, Ormrod, & Turner, 2007a). By this standard, 22% of children in a nationally representative sample were considered poly-victims in the previous year (Finkelhor et al., 2007a) and had higher rates of anxiety, depression, and trauma compared to those never experiencing abuse or had a single experience (Finkelhor et al., 2007a). Moreover, research has shown poly-victims experience more serious and injurious victimization (Finkelhor et al., 2011) and are at increased risk for subsequent victimization (Finkelhor, Ormrod, & Turner, 2007b). Experiencing multiple forms of victimization has not been well studied among homeless youth, though limited existing research shows multiple forms of abuse may negatively affect psychosocial and behavioral outcomes (Ferguson, 2009). Specifically, in her examination of types of abuse experienced (physical abuse, sexual, abuse, emotional abuse, psychological abuse, verbal abuse, and witnessing familial abuse). Ferguson (2009) found that homeless youth who reported experiencing multiple forms of abuse (i.e., both direct and indirect) demonstrated the most psychological symptoms (Ferguson, 2009). Although Ferguson’s (2009) study is limited by its small convenience sample, it provides a foundation for applying the concept of poly-victimization to homeless youth and introduces the notion of direct and indirect victimization as one conceptualization of poly-victimization and its effects. Most research on victimization among homeless youth has not investigated, in detail, the combinations of types and frequencies of victimization experienced. Instead, studies report prevalence estimates that suggest high overall rates (see Stewart et al., 2004). High rates, although concerning, do not differentiate subgroups of homeless youth who may differ in their service needs. Service agencies for homeless youth, without a more nuanced understanding of victimization profiles, may fail to address victimization or, at best, provide trauma-informed services generally. These agencies will likely fail to recognize the differential needs of youth with particularly harmful experiences. Differentiation in the mental health needs of youth with various victimization profiles will, thus, inform under-resourced service agencies regarding how to most efficiently serve victimized youth in their care. For example, agencies may identify methods of screening for victimization experiences that place youth at greatest risk for mental health problems and provide further assessment, referral, treatment, and supportive housing to those youth most in need. Recently, in a multi-site study of homeless youth, latent class analysis was conducted to identify specific categories of victimization among these youth (Bender et al., in press). Latent class analysis (LCA) allowed individuals to naturally group into classes based on the different types and frequencies of victimization they experienced. Rather than the researcher creating artificial cut points or predetermined criteria for group inclusions, the classes revealed in the LCA were naturally occurring groups of youth with similar patterns of victimization. The LCA identified three distinct victimization classes: (a) a low victimization class of youth who never or rarely experienced indirect or direct victimization; (b) a high-victimization class of youth who experienced high rates of direct and indirect victimization, typically experiencing each victimization event multiple times; and (c) a witness class of youth who remained on the periphery of direct violence but reported high rates of threats and witnessed violence, typically experiencing each victimization event only once (Bender et al., in press). This Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
3
previous work not only suggests the existence of distinct victimization profiles among homeless youth, but also finds that these subgroups differ in their substance use patterns such that youth in the low-victimization class reported significantly less substance abuse/dependence compared to witnesses and high-victims, while youth in the high-victimization group reported significantly greater substance dependence, particularly in regard to alcohol, compared to youth in the other two groups (Bender et al., in press). The purpose of the current study is to build on this previous class analysis by investigating whether these classes of victimized youth differ in their mental health needs, specifically rates of meeting diagnostic criteria for major depressive episodes and PTSD. Such investigation may inform differential screening, assessment, treatment, and prevention needs for homeless youth utilizing services at under-resourced shelters and drop in centers. Thus, two broad research questions guided our present investigation: controlling for other relevant correlates, do homeless youth in different victimization classes differ significantly in the degree to which they (a) meet criteria for having had a major depressive episode or (b) meet criteria for PTSD? Methods Sample and Recruitment A total of 601 street youth (ages 18–24) were recruited from homeless youth-serving host agencies in Los Angeles (n = 200), Denver (n = 201), and Austin (n = 200) using purposive sampling. Researchers selected host agencies based on their existing partnerships and the agency’s commitment to host the study. In each study, recruitment sites included drop-in centers offering case management, referral services, and basic subsistence items (food, hygiene supplies); shelters offering short-term (40 days) stay; and transitional housing apartments offering temporary (6 months) housing. Each investigator received human subjects’ approval from her university. Data collection occurred from March 2010 to April 2011. Recruitment procedures were nearly identical across cities with minor variations due to services emphasized in each location (e.g., more crisis-shelter users in Los Angeles, more drop-in service users in Denver and Austin). To participate in the study, youth had to meet three inclusion criteria: (a) be 18–24 years of age, (b) have spent at least 2 weeks away from home in the month before the interview (Whitbeck, 2009), and (c) provide written informed consent. Youth were excluded if they could not understand the consent form because of cognitive limitations (psychotic symptoms or developmental delays) or if they were noticeably intoxicated or high at the time of the interview. In the latter case, youth were asked to return at a later time when they could more competently answer interview questions. Agency case managers made the determination whether a particular individual was eligible for recruitment into the study based on their knowledge of each individual and their current level of intoxication. Data Collection and Measures Agency staff facilitated introductions and referred youth to research staff who explained the study procedures and secured written consent. Researchers administered a 45-minute quantitative retrospective interview containing both standardized self-report instruments and researcher-developed items that together assessed demographic and background information, victimization experiences, and psychiatric symptoms. Interviewers read questions and response options aloud to participants and youth responded verbally. Youth were compensated for their time with a $10.00 gift card to a local food vendor. Demographics. Basic demographics included age, gender (0 = male, 1 = female), and ethnicity (1 = white, 2 = black, 3 = Hispanic, 4 = other). Ethnicity was subsequently dummy coded to include Black (0 = no, 1 = yes), Latino (0 = no, 1 = yes), and other (0 = no, 1 = yes), with White as a reference category. To assess and control for inter-city differences, the city in which data were collected was recorded (1 = Los Angeles, 2 = Denver, 3 = Austin) and then dummy-coded to include Los Angeles (0 = no, 1 = yes) and Austin (0 = no, 1 = yes), with Denver as a reference category. Victimization Class. To capture the many forms of victimization homeless youth experience, the authors built on Ferguson’s (2009) definitions of direct and indirect abuse and operationalized “victimization” as experiencing any form of direct (i.e., physical, sexual, emotional, or verbal) or indirect (i.e., witnessing) abuse, or experiencing a traumatic event (e.g., sudden death of close friend or loved one). To do so, street victimization experiences were collected using a revised version of the Traumatic Life Events Questionnaire (Kubany et al., 2000). This measure assessed the frequency (0 = never, 1 = once, 2 = more than once) of experiencing different types of direct (e.g., robbery involving a weapon, physical assault by an acquaintance or stranger) and indirect victimization (e.g., sudden death of a close friend or loved one, witnessing a physical assault) since leaving home. This instrument was revised to exclude items with less relevance for this study (e.g., natural disasters, warfare or combat). Patterns in responses to these items were used to form latent victimization classes. Due to space limitations, this process is described elsewhere (Bender et al., in press) or details can be provided by contacting the first author. Fig. 1 displays the percentage of youth in each victimization class who experienced each type of victimization never, once, or more than once. Youths’ patterned responses placed them in one of three classes: (a) low victimization class (n = 215; never or rarely experienced indirect and direct victimization); (b) high-victimization class (n = 136); experienced high rates of direct and indirect victimization, typically experiencing each victimization event more than once; and (c) witness class (n = 250) included youth who remained on the periphery of direct violence but reported high rates of threats, witnessing violence, and Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
4
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Fig. 1. Percentage of youth in each class who experienced each victimization event. Adapted from Bender et al., in press.
typically experiencing each victimization event only once. The 3-category victimization class variable (1 = low-victimization class, 2 = witness class, 3 = high-victimization class) was recoded into a series of three dummy coded variables for use in multivariate analyses: low-victimization class (0 = no, 1 = yes), high-victimization class (0 = no, 1 = yes), and witness class (0 = no, 1 = yes). Homelessness Variables. Several variables associated with victimization in the extant homeless youth literature were measured, including primary living arrangement (0 = homeless, 1 = temporarily housed with relative, friend, foster parent, facility); transience or number of inter-city moves since leaving home; and length of time homeless (number of months between interview date and the date the youth last left home). Psychiatric Problems. Major depressive episode and PTSD were assessed by the Mini International Neuropsychiatry Interview (MINI; Sheehan et al., 1998). This is a widely used, brief, structured interview that facilitates screening for Axis I psychiatric disorders as outlined by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV-TR; American Psychological Association, 2000). The MINI has demonstrated good reliability (Lecrubier et al., 1997) as well as convergent validity with the Structured Clinical Interview for DSM-IV-TR Axis 1 Disorders (SCID), an established measure of diagnostic criteria (Sheehan et al., 1998). In assessing whether youth met criteria for a major depressive episode, respondents were asked two screening questions regarding feeling down and losing interest during the past 2 weeks. Those who responded affirmatively were then asked a series of questions regarding symptoms, such as changes in appetite, sleep patterns, energy level, concentration, and suicidal thoughts experienced during the past 2 weeks. Participants who reported 5 or more of the 9 symptoms listed were coded positive for meeting criteria for a current major depressive episode (0 = no, 1 = yes). The MINI was also used to assess whether youth met criteria for PTSD. Respondents were first asked three screening questions regarding: (1) whether or not they had ever experienced or witnessed an extremely traumatic event that included actual or threatened death or serious injury to them or someone else, (2) whether they responded with intense fear, helplessness, or horror, and (3) whether, during the past month, they re-experienced the event in a distressing way (dreams, intense recollections, flashbacks, or physical reactions). Those who answered affirmatively to all 3 screeners were then asked 13 additional questions reflecting DSM-IV symptoms of re-experiencing, avoidance, and arousal symptoms during the past month. Those who responded affirmatively to three or more re-experiencing and avoidance symptoms and two or more arousal symptoms, and confirmed that these symptoms had been significantly interfering with their lives within the past month were categorized as meeting criteria for PTSD (0 = no, 1 = yes). Data Analysis Univariate analyses were used to describe the full sample’s demographic and psychiatric characteristics. Bivariate analyses (chi square, ANOVA) examined differences in sample characteristics across data collection sites. Utilizing two binary logistic regression analyses, first PTSD, then depression, were regressed on demographic variables (age, gender, ethnicity, city), homelessness variables (primary living situation, transience, time homeless) and dummy coded victimization Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model
ARTICLE IN PRESS
CHIABU-2746; No. of Pages 8
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
5
class assignment variables (high-victimization class and witness class with low-victimization class as the reference category). These models examined whether victimization class membership was associated with meeting criteria for a major depressive episode and PTSD, controlling for other correlates, and these initial models compared youth in the witness and high-victimization classes to youth in the low-victimization class reference category. These models were then re-analyzed in order to compare the effects of high-victimization class membership to witness class membership on depression and PTSD by changing the dummy coded class assignment variables to include low-victimization class and witness class, leaving high-victimization class as the reference category. Results Sample Characteristics Sample characteristics are described in Table 1, including demographic variables, correlates, and victimization class assignment for the full sample (N = 601) and for each subsample by data collection site (Los Angeles, Denver, and Austin). Because of differences on some variables across sites, city was controlled for in the multivariate analysis. Correlates of Class Membership To address the study’s two research questions, a series of binary logistic regression analyses were conducted. First, depression (model 1) and then PTSD (model 2) were regressed on victimization class assignment variables (high-victimization class and witness class with low-victimization class as the reference category), controlling for demographic (age, gender, ethnicity, city) and homelessness variables (primary living situation, transience, time homeless). Results for these models are displayed in Table 2. Model results indicated that, compared to low-victims (i.e., youth who experienced no or very little indirect or direct victimization), witnesses (i.e., youth experiencing indirect victimization) and high-victims (i.e., youth experiencing indirect and direct victimization on multiple occasions) were more likely to meet criteria for a major depressive episode and PTSD. Specifically, in regards to depression, 2 [12] = 34.37, p < .001, compared to low-victims, witnesses were over 2 times more likely (OR = 2.34, p < .001) and high-victims were over 2 times more likely (OR = 2.12, p < .01) to meet criteria for a major depressive episode. Similarly, in regards to PTSD, 2 [12] = 59.46, p < .001, witnesses were 3 times more likely (OR = 3.16, p < .001) and high-victims were 3 times more likely (OR = 3.39, p < .001) to meet criteria for PTSD compared to low-victims.
Table 1 Full and city-level sample characteristics of homeless young adults. Full sample (N = 601)
Gender Male Female Race/ethnicity White Black Latino Other Primary living situation Streets Temp housing Meet diagnostic criteria Major depressive episode PTSD Co-occurring depression and PTSD Class assignment Low-victimization High-victimization Witness Mean Age Number inter-city moves Months homeless ***
20.5 3.5 32.4
Denver (n = 201)
Los Angeles (n = 200) n
%
n
Austin (n = 200)
n
%
%
n
385 216
64.1 35.9
128 72
64.0 36.0
129 72
64.2 35.8
128 72
64.0 36.0
240 152 107 101
39.9 25.3 17.8 16.8
18 92 52 38
9.0 46.0 26.0 19.0
70 54 37 39
34.8 26.9 18.4 19.4
152 6 18 24
76.0 3.0 9.0 12.0
303 298
50.4 49.6
87 113
43.5 56.5
90 111
44.8 55.2
126 74
63.0 37.0
187 136 79
31.1 22.7 13.2
62 65 32
31 32.5 16
73 39 28
36.3 19.4 13.9
52 32 19
26.4 16.2 9.7
215 136 250
35.8 22.6 41.6
95 24 81
47.5 12.0 40.5
70 42 89
34.8 20.9 44.3
50 70 80
25.0 35.0 40
Chi-square/ ANOVA
% .002
2 = 207.9***
2 = 19.1***
SD 1.6 3.7 31.0
Mean 19.4 2.6 22.9
SD 1.1 3.2 26.0
Mean 19.9 1.9 31.6
SD 1.6 2.8 29.6
Mean 20.8 5.9 42.8
SD 1.7 3.7 33.9
2 = 4.6 2 = 17.0*** 2 = 3.56 2 = 38.5***
ANOVA F = 40.6*** F = 85.3*** F = 22.1***
p < .001.
Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
6
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Table 2 Logistic regression models predicting depression and PTSD. Depression model 1
Age Gender (male) Ethnicity (Black) Ethnicity (Latino) Ethnicity (White) Transience Time homeless Primary living situation (streets) Los Angeles (other cities) Austin (other cities) Witness class (low-victimization reference category) High-victimization class (low-victimization reference category) Witness class (high-victimization reference category)
PTSD model 2
Depression model 3
PTSD model 4
OR
CI
OR
CI
OR
CI
OR
1.05 1.38 73 1.23 92 96 1.00 82 95 60 2.34***
.92–1.20 .95–2.00 .40–1.32 .67–2.24 .52–1.60 .90–1.02 .99–1.01 .57–1.19 .60–1.50 .36–1.02 1.52–3.61
1.00 1.98** .62 .84 .61 .98 1.00 .82 2.53*** .77 3.16***
86–1.17 1.30–3.00 33–1.18 44–1.61 33–1.14 91–1.05 1.00–1.01 54–1.25 1.50–4.25 41–1.45 1.89–5.28
1.05 1.38 73 1.23 92 96 1.00 82 95 60
92–1.20 95–2.00 40–1.32 67–2.24 52–1.60 90–1.02 99–1.01 57–1.19 60–1.50 36–1.02
1.00 1.98** .62 .84 .61 .98 1.00 .82 2.53*** .77
.86–1.17 1.30–3.00 .33–1.18 .44–1.61 .33–1.14 .91–1.05 1.00–1.01 .54–1.25 1.50–4.25 .41–1.45
2.12**
1.24–3.65
3.39***
1.79–6.42
1.10
69–1.77
.93
.55–1.58
2 (12) = 34.37, p < .001
2 (12) = 59.46, p < .001
2 (12) = 34.37, p < .001
CI
2 (12) = 59.46, p < .001
Note. Shaded region displays results comparing witness class to high-victimization class as predictors of depression and PTSD; OR = odds ratio; 95% CI = 95% confidence interval. ** p < .01. *** p < .001.
Among control variables, gender (youth who were female; OR = 1.98, p < .001) and data collection site (youth who participated in the study in LA; OR = 2.53, p < .001) increased odds of meeting criteria for PTSD; no control variables predicted depression. For a final analysis, we included witness class but exchanged the high-victimization class variable for the lowvictimization class variable (making high-victimization class the new reference category) and predicted depression (model 3) and PTSD (model 4), controlling for the same demographic and homelessness control variables. The effects of all control variables in model 3 and 4 did not change, but this allowed us to compare youth in the witness class to youth in the high-victimization class. Compared to youth in the high-victimization class, youth in the witness class were not significantly more or less likely to meet criteria for depression (OR = 1.10, p = .887) or PTSD (OR = .930, p = .789). These results appear in the shaded column of Table 2. Discussion The current study represents the first known investigation of whether homeless youth victimization profiles differ significantly in psychiatric problems, specifically rates of meeting criteria for major depressive episode and PTSD. Victimization profiles were derived from our previously conducted latent class analysis (Bender et al., in press), which identified three distinct victimization classes: (a) youth in a low-victimization class never or rarely experienced either direct or indirect forms of victimization; (b) youth in a high-victimization class reported higher rates of experiencing both direct and indirect victimization, often experiencing each victimization event more than once; and (c) youth in a witness class experienced elevated rates of indirect trauma rather than their own direct victimization, typically experiencing each victimization event only once. The results of the current study suggest that homeless youth in the high-victimization and witness classes were significantly more likely to meet criteria for a major depressive episode and for PTSD than youth in the low-victimization class. Thus, those homeless youth who experience victimization, whether direct and indirect, or indirect only, are likely to be at higher risk for major depressive episodes and PTSD than those youth who experience no or low rates of victimization. Our findings support previous literature indicating high rates of homeless youth victimization (e.g., Tyler & Beal, 2010) and an association between said victimization and mental health problems (e.g., Whitbeck et al., 2000, 2007). However, this study uniquely contributes an understanding that homeless youth do not represent a heterogeneous population with universally high victimization exposure and psychiatric needs, but rather a collection of subgroups of youth with varying levels of exposure and psychiatric symptoms. Recognizing subgroups with differential treatment needs suggests under-resourced service agencies may need to develop screening protocols that differentiate high-risk subgroups in need of referral and more costly treatment from low-risk subgroups who may benefit from less costly efforts to prevent victimization and associated psychiatric outcomes. That more than one-third (n = 215) of this sample was classified as low-victims suggests lower-cost prevention efforts should be a major emphasis in service settings.‘ Of note, results demonstrate that homeless youth exposed to indirect victimization (witness class) are at similarly high risk for psychiatric problems as youth exposed to a combination of direct and indirect victimization (high-victimization class). Previous research has demonstrated similar findings of the deleterious effects of indirect trauma on youths’ mental Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
7
health in a general context. For example, using the National Survey of Adolescents, direct victimization (i.e., physical assault, sexual assault) and indirect victimization (i.e., witnessing violence) were found to significantly predict mental health outcomes, including comorbid diagnosis of PTSD and depression (Kilpatrick et al., 2003). High rates of mental health diagnoses, specifically PTSD, have been found in other adolescent samples that have witnessed violence, particularly children exposed to domestic violence (Kilpatrick & Williams, 1997; Luthra et al., 2009). Specifically, in a small sample of children exposed to domestic violence, only one of 20 children did not meet criteria for a PTSD diagnosis ranging from mild to severe (Kilpatrick & Williams, 1997). Luthra et al.’s (2009) examination of various types of adolescent trauma exposure similarly demonstrated a correlation between witnessing violence and PTSD diagnosis. Our study adds to this literature, indicating that indirect forms of victimization can have serious mental health consequences for homeless youth – a population with elevated exposure to dangerous street situations – and suggests that even those homeless youth who are on the periphery of violence are likely in need of mental health services.
Limitations Study limitations should be considered in interpreting our findings. The cross-sectional design used in this study limits assumptions of causal order; thus, it may be that youth who meet criteria for depressive episodes or PTSD are placed at greater risk for victimization. Longitudinal research, although challenging with this population, is necessary to determine the causal order or the relationship discussed here. Furthermore, our sample of service-seeking youth limits the generalizability of our results to youth disconnected from services – a subpopulation with potentially greater risk for victimization due to disconnection from formal/institutional supports. In addition, social desirability bias may have prevented youth from sharing sensitive information about victimization and psychiatric symptoms. In attempts to reduce this bias, our interviewers received extensive training in building rapport with homeless youth as well as stressing confidentiality and assuring privacy during interviews. Nonetheless, rates of victimization and psychiatric symptoms may have been underreported. Conversely, with few existing measures of psychiatric problems normed for homeless youth, it is possible that our measure of PTSD could have inflated rates slightly given some homeless youths’ protective hyper-vigilance in a street environment. Finally, although a strength of this study was the use of a large multi-site dataset, this also created challenges. For example, data collection sites varied in rates of PTSD. While we controlled for data collection site in all analyses, this site level difference deserves further investigation, including whether mental health service provisions differ significantly across cities.
Implications Screening for victimization experiences and associated psychiatric symptoms should be systematically conducted across homeless youth service agencies. Our study suggests rates for trauma, PTSD, and depressive episodes are high; thus, greater understanding of the degree and type of trauma experienced is critical to providing adequate assessment and treatment. Screening should include questions about youths’ experiences of various forms of direct and indirect victimization. Youth who screen positive for experiencing either (or both) forms of victimization should be engaged in mental health services if amenable. Service providers should take seriously experiences of indirect victimization, as it appears witnessing violence and experiencing threats affect youths’ psychological health. Engaging youth in mental health treatment following positive screening for victimization and associated psychiatric symptoms may prove challenging. Research finds that although 84% of youth have histories of victimization, only 44% report an interest in treatment to address these issues (Keeshin & Campbell, 2011). While some youth simply do not see the need for treatment, believing instead that their struggles make them stronger, others describe past treatment experiences that were ineffective as reasons not to seek further help (Keeshin & Campbell, 2011). Careful consideration should be given to when and in what contexts trauma treatment is provided. Intensive approaches to processing previous trauma, including exposure to trauma stimuli, may not be appropriate for youth living on the streets or in other unstable contexts where they are consistently confronted with victimization risk. Hyper-vigilance associated with previous victimization, for example, may be protective for those currently living in dangerous situations (McManus & Thompson, 2008). For these street-living youth, interventions should be developed and tested that teach self-protection and safety in short-term shelter and drop-in centers. In contrast, trauma treatments should be tested with youth who have found safe/stable living environments off the streets in long-term transitional or permanent housing. These are likely to be the most appropriate venues for intervention and allow a determination of whether such interventions are effective with this highly vulnerable group (Thompson, McManus, & Voss, 2006). Although treatment is recommended for youth with high-victimization and witness profiles, a large subsample of youth reported little victimization and low rates of psychiatric symptoms. For this group, victimization prevention programming is warranted. Skills-based programs aimed at preventing high-risk behaviors (e.g., substance use and sexual risk behaviors) have demonstrated effectiveness with homeless youth (Rotheram-Borus et al., 2003). Similar programs should be developed to integrate safety skills, helping youth to identify dangerous situations and problem-solve assertively to avoid victimization. Such programs may be best provided in short-term homeless youth shelters where staff may address victimization risk factors before youth leave shelters for the streets where their vulnerability is heightened. Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
G Model CHIABU-2746; No. of Pages 8
ARTICLE IN PRESS
8
K. Bender et al. / Child Abuse & Neglect xxx (2014) xxx–xxx
Homeless youth encounter multiple forms of victimization on the streets. However, their experiences and resulting mental health needs vary considerably. Assessing for youths’ victimization experiences and providing appropriate prevention or treatment programming is critical to helping youth remain safe on the streets and cope with past traumatic experiences. References American Psychological Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed. text rev.). Washington, DC: American Psychiatric Association. Bender, K., Thompson, S., Ferguson, K., & Langenderfer, L. (in press) Substance use predictors of victimization profiles among homeless youth: A latent class analysis. Journal of Adolescence. Bender, K., Ferguson, K., Thompson, S. J., Komlo, C., & Pollio, D. (2010). Factors associated with trauma and posttraumatic stress disorder among homeless youth in three U.S. cities: The importance of transience. Journal of Traumatic Stress, 23, 161–168. http://dx.doi.org/10.1002/jts.20501 Ferguson, K. (2009). Exploring the psychosocial and behavioral outcomes of multi-type abuse among homeless young adults. Social Work Research, 33, 219–230. http://dx.doi.org/10.1093/swr/33.4.219 Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2007a). Poly-victimization: A neglected component in child victimization. Child Abuse & Neglect, 31, 7–26. http://dx.doi.org/10.1016/j.chiabu.2006.06.008 Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2007b). Re-victimization patterns in a national longitudinal sample of children and youth. Child Abuse & Neglect, 31, 479–502. http://dx.doi.org/10.1016/j.chiabu.2006.03.012 Finkelhor, D., Turner, H. A., Hamby, S., & Ormrod, R. (2011). Polyvictmization: Children’s exposure to multiple types of violence, crime, and abuse. Juvenile Justice Bulletin.,. Retrieved from: https://www.ncjrs.gov/pdffiles1/ojjdp/235504.pdf Hoyt, D. R., Ryan, K. D., & Cauce, A. M. (1999). Personal victimization in a high-risk environment: Homeless and runaway adolescents. Journal of Research in Crime and Delinquency, 36, 371–392. http://dx.doi.org/10.1177/0022427899036004002 Keeshin, B. R., & Campbell, K. (2011). Screening homeless youth for histories of abuse: Prevalence, enduring effects, and interest in treatment. Child Abuse & Neglect, 35, 401–407. http://dx.doi.org/10.1016/j.chiabu.2011.01.015 Kilpatrick, D. G., Ruggiero, K. J., Acierno, R., Saunders, B. E., Resnick, H. S., & Best, C. L. (2003). Violence and risk of PTSD, major depression, substance abuse/dependence, and comorbidity: Results from the National Survey of Adolescents. Journal of Consulting and Clinical Psychology, 71, 692–700. http://dx.doi.org/10.1037/0022-006X.71.4.692 Kilpatrick, K. L., & Williams, L. M. (1997). Post-traumatic stress disorder in child witnesses to domestic violence. American Journal of Orthopsychiatry, 67, 639–644. http://dx.doi.org/10.1037/h0080261 Kipke, M. D., Simon, T. R., Montgomery, S. B., Unger, J. B., & Iversen, E. F. (1997). Homeless youth and their exposure to and involvement in violence while living on the streets. Journal of Adolescent Health, 20, 360–367. http://dx.doi.org/10.1016/S1054-139X(97)00037-2 Kubany, E. S., Leisen, M. B., Kaplan, A. S., Watson, S. B., Haynes, S. N., Owens, J. A., & Burns, K. (2000). Development and preliminary validation of a brief-broad-spectrum measure of trauma exposure: The Traumatic Life Events Questionnaire. Psychological Assessment, 12, 210–224. http://dx.doi.org/10.1037/1040-3590.12.2.210 Lecrubier, Y., Sheehan, D., Weiller, E., Amorim, P., Bonora, I., Sheehan, K., Sheehan, K. H., & Dunbar, G. (1997). The MINI International Neuropsychiatric Interview. A short diagnostic structured interview: Reliability and validity according to the CIDI. European Psychiatry, 12, 224–231. http://dx.doi.org/10.1016/S0924-9338(97)83296-8 Lindsey, E. W., Kurtz, P. D., Jarvis, S., Williams, N. R., & Nackerud, L. (2000). How runaway and homeless youth navigate troubled waters: Personal strengths and resources. Child and Adolescent Social Work Journal, 17, 115–141. Luthra, R., Abramovitz, R., Greenberg, R., Schoor, A., Newcorn, J., Schmeidler, J., Levine, P., & Chemtob, C. M. (2009). Relationship between type of trauma exposure and posttraumatic stress disorder among urban children and adolescents. Journal of Interpersonal Violence, 24, 1919–1927. http://dx.doi.org/10.1177/0886260508325494 McManus, H. H., & Thompson, S. J. (2008). Trauma among unaccompanied homeless youth: The integration of street culture into a model of intervention. Journal of Aggression, Maltreatment & Trauma, 16, 92–109. http://dx.doi.org/10.1080/10926770801920818 Rosenthal, D., Mallett, S., & Myers, P. (2006). Why do homeless young people leave home? Australian and New Zealand Journal of Public Health, 30, 281–285. http://dx.doi.org/10.1111/j.1467-842X.2006.tb00872.x Rotheram-Borus, M., Song, J., Gwadz, M., Lee, M., Van Rossem, R., & Koopman, C. (2003). Reductions in HIV risk among runaway youth. Prevention Science, 4, 173–187. http://dx.doi.org/10.1023/A:1024697706033 Sheehan, D., Lecrubier, Y., Sheehan, K., Amorim, P., Janavs, J., Weiller, E., & Dunbar, G. (1998). The Mini-International Neuropsychiatric Interview (MINI): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59, 22–33. Slesnick, N., Erdem, G., Collins, J., Patton, R., & Buettner, C. (2010). Prevalence of intimate partner violence reported by homeless youth in Columbus, Ohio. Journal of Interpersonal Violence, 25, 1579–1593. http://dx.doi.org/10.1177/0886260509354590 Stewart, A. J., Steiman, M., Cauce, A. M., Cochran, B. N., Whitbeck, L. B., & Hoyt, D. R. (2004). Victimization and posttraumatic stress disorder among homeless adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 325–331. http://dx.doi.org/10.1097/00004583-200403000-00015 Thompson, S. J., McManus, H. H., & Voss, T. (2006). PTSD and substance abuse among youth who are homeless: Treatment issues and implications. Brief Treatment and Crisis Intervention, 6, 1–12. Thrane, L. E., Hoyt, D. R., Whitbeck, L. B., & Yoder, K. A. (2006). Impact of family abuse on running away, deviance, and street victimization of homeless rural and urban youth. Child Abuse & Neglect, 30, 1117–1128. http://dx.doi.org/10.1016/j.chiabu.2006.03.008 Tyler, K., & Beal, M. R. (2010). The high-risk environment of homeless young adults: Consequences for physical and sexual victimization. Violence and Victims, 25, 101–115. http://dx.doi.org/10.1891/0886-6708.25.1.101 Tyler, K. A., & Cauce, A. M. (2002). Perpetrators of early physical and sexual abuse among homeless and runaway adolescents. Child Abuse & Neglect, 26, 1261–1274. http://dx.doi.org/10.1016/S0145-2134(02)00413-1 Tyler, K. A., Whitbeck, L. B., Hoyt, D. R., & Johnson, K. D. (2003). Self-mutilation and homeless youth: The role of family abuse, street experiences, and mental disorders. Journal of Research on Adolescence, 13, 457–474. http://dx.doi.org/10.1046/j.1532-7795.2003.01304003.x Whitbeck, L. B. (2009). Mental health and emerging adulthood among homeless young people. New York, NY: Taylor & Francis Group. Whitbeck, L. B., Hoyt, D. R., & Bao, W. (2000). Depressive symptoms and co-occurring depressive symptoms, substance abuse, and conduct problems among runaway and homeless adolescents. Child Development, 71, 721–732. http://dx.doi.org/10.1111/1467-8624.00181 Whitbeck, L. B., Hoyt, D. R., Johnson, K. D., & Chen, X. (2007). Victimization and posttraumatic stress disorder among runaway and homeless adolescents. Violence and Victims, 22, 721–734. http://dx.doi.org/10.1891/088667007782793165
Please cite this article in press as: Bender, K., et al. Mental health correlates of victimization classes among homeless youth. Child Abuse & Neglect (2014), http://dx.doi.org/10.1016/j.chiabu.2014.03.001
