|Home | About | Journals | Submit | Contact Us | Français|
Cross-national comparisons of homeless youth in Melbourne, Australia, and Los Angeles, CA, United States were conducted. Newly (n = 427) and experienced (n = 864) homeless youth were recruited from each site. Compared to Australia, homeless youth in the United States were younger, more likely to be in school or jail, demonstrated fewer sexual and substance use risk acts, fewer suicidal acts, and reported less need for social services. Across sites, experienced homeless youth were more likely to be older, male, engage in sexual and substance use, report greater need for social services, and make greater use of work, substance use, and health-related services. Homeless youth have different behavioral profiles in Australia and the United States, reflecting differences in the effectiveness of service systems in the two countries in keeping youth with fewer problems out of homelessness.
Homelessness among youth is a significant problem in both the United States (US) and Australia (AU) (Hammer, Finkelhor, & Sedlak, 2002; MacKenzie & Chamberlin, 2002; Prime Ministerial Homeless Report, 1998). Because AU and the US are similar in several key dimensions, data gathered in AU may inform US policies regarding homeless youth. Both countries are developed nations with large immigrant populations; both have strong national norms for individualism (in contrast to collectivism) (Kim, Triandis, Kagitcibasi, Choi, & Yoon, 1994); and services for homeless youth are similarly organized with a system of community-based shelters, primary health care clinics, foster care, and, to a lesser extent, comprehensive service agencies.
Yet, the cross-national differences in service and policy contexts offer important opportunities to examine national variations in youth's pathways into and out of homelessness. In AU, services for homeless youth are organized so that a large number of relatively small community-based agencies are geographically dispersed throughout the city (Harrison & Dempsey, 1997). In contrast, US services are more likely to be provided at a few large shelters located in the inner city. Further, AU has adopted a harm-reduction approach to substance use (Nguyen & Fox, 2002), has had multiple national social marketing plans for HIV prevention (Ross, Rigby, Rosse, Anagnostou, & Brown, 1990), has adopted a sexual health approach (not the US disease prevention model) (Harrison, Hillier, & Walsh, 1996), offers universal access to health care (Doctors Reform Society, 2003), provides some school-based prevention services to divert youth from homelessness, and provides a weekly stipend for homeless youth. Moreover, AU has less socioeconomic diversity than the US. Given these contextual differences, our first goal was to document the similarities and differences among homeless youth in Melbourne, AU and Los Angeles, CA, US.4
We anticipated that there may be different pathways into and out of homelessness in the two settings; therefore, we compared newly homeless and experienced youth (those on the streets a long time) in each country.5 Homeless youth show very high rates of problem behaviors in both countries (Booth & Zhang, 1996; Inciardi & Surratt, 1998; Kipke, Montgomery, Simon, & Iverson, 1997; Sleegers, Spijker, van Limbeek, & van Engeland, 1998; Greene & Ringwalt, 1996; Kral, Molnar, Booth, & Watters, 1997; Rosenthal, Moore, & Buzwell, 1994; Buhrich & Teesson, 1990; Kamieniecki, 2001; Teesson, Hodder, & Buhrich, 2003). Yet, it is unclear whether newly homeless youth in both countries would have exhibited similar behavioral profiles whether they were in the US and AU. Because the methodologies are not consistent across previous studies, it is difficult to compare the behavioral profiles of youth in the two countries. Therefore, using similar measures, we compared newly homeless and experienced homeless youth in two major regions in each country.
Representative samples of homeless youth were recruited in Melbourne, AU and Los Angeles County (LA), US through a three step process of: (1) sampling, (2) screening for eligibility and recruitment, and (3) assessment.
A homeless youth was defined as: (1) ranging in age from 12 to 20 years; and (2) spending at least two consecutive nights away from home, either without his or her guardian's permission or after being told to leave home. Youth aged 18 to 20 were included because service providers noted that some young people who had recently left home were over 17 years and the range of 12 to 17 years did not accurately reflect the population that they were serving. Two subgroups of homeless youth were identified: (1) newly homeless youth who had been living away from home for less than 6 months in total; or (2) experienced homeless youth who had been living away from home for more than 6 months.6
The sampling procedure varied slightly across countries, reflecting differences in the types, number, and geographical distribution of agencies serving homeless youth and policies in each setting. In the US, sites were selected through a systematic process. First, all of the potential recruitment sites for homeless adolescents in Los Angeles County were identified by interviewing line and supervisory staff in agencies that served homeless adolescents throughout the county (Brooks, Milburn, Witkin, & Rotheram-Borus, 2004). Thirty sites were identified, including 17 shelters and drop-in centers and 13 street hangout sites. Next, the 30 sites were audited at pre-selected times over three different week-long time periods to determine the number of homeless adolescents that could be found at each site. All of these locations were included as recruitment sites. Interviewers were sent out in pairs to screen and recruit eligible homeless adolescents. The interviewers approached youth at the sites, identifying themselves as researchers, and asked youth if they would take a few minutes for an interview about their homelessness.
In AU, recruitment sites were systematically selected from a database of all young people and homeless services across the five metropolitan regions of Melbourne, including shelters (refuges), housing referral sites, employment training and government assistance programs, and preventive service agencies. Service managers were approached to seek permission for their service's involvement in the project. Of the 114 eligible services, 112 agreed to participate. Participants were subsequently recruited from 95 agencies. Each service was given an information session that outlined the project and informed staff of the eligibility criteria for participants. Potentially eligible homeless young people were referred to the study by staff working at these services, either by giving the young person the project's free-call number or telephoning the number on their behalf.
Interviewers in both countries used a 13-item screen to determine whether youth met the eligibility criteria as homeless and establish whether the youth was newly homeless or experienced. All newly homeless youth were recruited. Experienced youth were randomly selected for recruitment (of youth who were screened and found to be experienced, a random sample was selected for assessment). Following screening, voluntary informed consent was obtained from each youth, with the caveats that physical or sexual abuse, suicidal and homicidal feelings would be reported. Informed consent was obtained directly from all participants 18 years and older. For minors, loco parentis consent was obtained from a member of the outreach (recruitment) team present, and assent was obtained from the minor. The study fulfilled all human subjects guidelines and was approved by the appropriate Institutional Review Boards at both the US and AU research centers.
The interviewers received approximately 40 hr of training, including lectures, role-playing, mock surveys, ethics training, and emergency procedures. All interviews were conducted face-to-face in convenient locations (e.g., private space at service sites, coffee shops, etc.) using an audio computer assisted interview (ACASI) and lasted between 1 and 1.5 hr. Some of the interview was conducted by the interviewer while more sensitive sections were self-administered. All participants received $20 (local currency) compensation for their time. Baseline refusal rates were less than 7% in the US and not evaluated in AU.
The following self-report measures were completed by all youth. “Recent” refers to the past three months.
Sociodemographic Characteristics included age, gender, having a non-English speaking parent, mother's and father's education (less than high school, high school graduate and/or some college, or college graduate), sexual orientation (heterosexual or non-heterosexual), pregnancy history including whether had ever given birth (pregnant: lifetime, and pregnant: recent), educational experience including attending school, the recent number of days of school missed, and being a high school graduate, guardianship (both parents, a single parent, extended family, foster parents, self, or other), and, for Australian youth, government assistance (youth allowance or no youth allowance).
Risk behaviors consisted of high risk sexual behaviors for HIV transmission, including recent unprotected sex (not using condoms during vaginal and/or anal sex acts), recent sex work (vaginal, anal sex in exchange for money, drugs, food or a place to stay), recent number of partners (the sum of vaginal and anal sex partners), and four or more lifetime partners (a count of the anal and vaginal sex partners ever); disease testing and status including lifetime testing for HIV and hepatitis C and whether tested positive and self-reported health (excellent, good, fair, or poor). Substance use included recent drug use (used marijuana, cocaine, heroin, methamphetamines or alcohol), recent alcohol use, recent marijuana use, recent marijuana/alcohol use only, recent hard drug use (used heroin, cocaine, crack, methamphetamines, or ecstasy), and recent injected drug use. Mental health included whether the youth had attempted suicide (lifetime, recent attempted suicide) and overall mental health (scores on the Brief Symptom Inventory (Derogatis, 1993) [α = .96], wherein higher scores indicate more emotional distress). Jail/juvenile detention was reported over lifetime incidence.
Service utilization consisted of whether youth needed help for various problems (need for service) and whether they had used a community-based service to address that need (use of service). Need and use were rated for seven problem domains: (1) homeless services (housing/shelter); (2) mental health services (depression/anxiety); (3) general health services (hepatitis C, HIV treatment/service, pregnancy or general health); (4) work-related services (work and income); (5) interpersonal services (family, boyfriend/girlfriend, and gender identity/sexual attraction); (6) drug/alcohol services; and (7) school-related services. Number of needs was a continuous variable that summed across all seven domains, such that youth who had no needs scored 0, whereas youth who had needs in seven domains scored 7. Number of unmet needs was also continuous. For each identified need domain, if youth had not used a service in that domain, their number of unmet needs increased by 1, such that youth who had needs in all domains but used no services scored 7.
Sociodemographic characteristics, risk behaviors, and service utilization were compared among homeless youth based on country differences (US or AU) and homelessness experience (newly or experienced). Chi-square tests examined proportional significant differences for categorical variables. For continuous variables, t-tests determined the significance of mean differences. To account for all country-by-experience interactions, simple logistic regressions were conducted for categorical variables, resulting in odds ratios for each pair-wise comparison. For continuous variables, adjusted means were calculated and the significance of pair-wise differences in adjusted means was reported. Significance was set at p < .05. Two multivariate logistic regression models were conducted. Significance was set at p < .05 and only variables that met these criteria were retained in the final model. Both models were associational, not causal. These models identify the important independent factors associated with country differences and experience differences, respectively.
Table I compares youth's sociodemographic characteristics by country and homelessness experience. American youth (n = 618) were mostly ethnic minorities (e.g., 34.4% Hispanic, 21.6% African American, and 26.1% Caucasian), with 261 newly homeless youth (42%; 105 males) and 357 experienced youth (58%; 194 males). The AU sample (n = 673) included 166 newly homeless youth (25%; 77 males) and 507 experienced youth (75%; 257 males). Youth ranged in age from 12 to 20 years (M = 17.3; SD = 1.9); American youth were younger than their Australian counterparts. Most youth came from households where both parents were English-speakers (70%). More American than Australian youth had at least one non-English speaking parent. More American youth had parents who had graduated from high school or college, whereas more Australian youth had parents who had not completed high school. More American youth claimed the state or foster parents as their legal guardian and came from single-parent households. Likewise, more American than Australian youth were high school graduates. More American than Australian youth had given birth. Only 80% of youth identified as heterosexual; more American than Australian youth self-identified as gay or bisexual.
Overall, 47% of youth had recent unprotected sex; significantly more Australian than American youth had such sex. More Australian than American youth claimed to have had four or more lifetime sexual partners. More Australian than American youth had taken drugs of any kind, used alcohol, used hard drugs, or injected drugs. Overall, 10% of youth in either country had been jailed or held in a juvenile detention center. Significantly more American youth had been jailed than Australian youth, by a factor of three. More American youth had been tested for HIV than Australians, but the HIV prevalence was similar in both countries. Although testing for hepatitis C was similar across countries, more AU youth were positive for hepatitis C. American youth reported better health status than AU youth. American youth reported better overall mental health than Australian youth. Fewer American than Australian youth had ever or recently attempted suicide.
Without exception, use of services was less than need. More Australian youth reported need/use than did American youth; specifically homeless, mental health, health services, work-related and drug/alcohol abuse services. The one type of service used more by American youth was interpersonal services.
Table II presents the results of a logistic regression model that identifies the important independent factors associated with cross-national differences. Overall, 17 of the variables presented in Table I are included in this model; all these independent factors are significant at the p < .05 level. Controlling for other factors, AU youth tended to be older, more had four or more sexual partners, more were hepatitis C positive, more had overall needs for social services, and more used general health and work-related services. Controlling for all other factors, more American youth had a mother or father graduate from high school, more gave birth, more were high school graduates, more came from foster parent households, more had been tested for HIV, more had been incarcerated, more needed school-related services, and more used interpersonal services as compared with Australian youth.
Table III presents characteristics of newly homeless compared to experienced homeless youth. Experienced youth in both countries were older and more likely to be male than newly homeless youth. In both countries, more newly homeless had at least one non-English speaking parent than experienced youth. More newly homeless youth were still attending school, and they had missed fewer days of school recently than their experienced peers. Significantly more experienced youth were gay or bisexual, compared to the newly homeless youth. More experienced homeless youth claimed the state or foster parents as their legal guardian than did newly homeless youth.
Experienced homeless youth were more likely to use drugs of any kind and, more specifically, to use alcohol, marijuana, hard drugs, and injected drugs. Overall, the experienced youth in both countries participated in more risky sexual behaviors than newly homeless youth: they engaged in more unprotected sex and sex work, and had more current or lifetime sexual partners. Compared with the newly homeless youth, more experienced youth were drug users of any kind, alcohol drinkers, marijuana users, hard drug users, and injection drug users. More newly homeless youth were exclusive users of marijuana and alcohol compared to their experienced peers. More experienced than newly homeless youth had been jailed or held in juvenile detention centers.
More experienced homeless youth claimed the state or foster parents as their legal guardian than did newly homeless youth. Experienced youth reported worse overall physical and mental health than newly homeless. More experienced youth had ever attempted suicide than newly homeless youth. In AU, more experienced youth received a government stipend (78%) compared to newly homeless (57%); no US youth received such stipends.
Overall, newly homeless youth reported having more unmet service needs than experienced youth. Some services were used more and were in greater demand by experienced than newly homeless youth: homeless, health, work-related, and drug/alcohol abuse services. Other services such as mental health, school-related, and interpersonal services were more often used and needed by newly homeless than experienced youth.
Table IV presents the results of a multivariate model that explores differences by experience on the streets. There were 12 independent factors associated with experience differences in this model; all 12 variables were significant at the p < .05 level. Controlling for all other factors, experienced youth were older, more had been pregnant (or got a partner pregnant), more came from foster homes, more were doing sex work, more had at least four sexual partners, more used intravenous drugs, and more had a need for general health services as compared to newly homeless youth. Likewise, more newly homeless youth were currently attending school, more came from single parent households, more had overall unmet service needs, and more were in need of school-related social services as compared to experienced homeless youth.
Table V presents results of logistic regressions (odds ratios, confidence limits, and significant differences) in the behavioral profiles of homeless youth based on country. Table VI presents results on length of homelessness (new or experienced) for dichotomous variables for each pair-wise comparison. Table VII presents the adjusted mean scores for comparisons based on the country and homelessness experience for continuous variables.
In the United States, newly homeless youth were significantly more likely to be children of single parents, to be female, and to have non-English speaking parents than experienced homeless youth. Similarly, newly homeless US youth were significantly less likely to have ever been pregnant, to have become a parent, or to have engaged in sexual risk acts or drug and alcohol use, but were more likely to have recently attempted suicide than experienced homeless youth in the US. Newly homeless youth were less likely to be gay or bisexual and less likely to have been in jail than experienced US youth. However, experienced US youth were also more likely than AU experienced youth to have been in jail and to be gay or bisexual, and not to be attending school.
Within Australia, newly homeless youth were more likely than experienced homeless AU youth to have non-English speaking parents, to have been pregnant, to be parents, and to attend school, and were less likely to have received government assistance, to have engaged in sexual risk acts, to have been tested for HIV and hepatitis C, and to have used hard drugs.
Newly homeless youth in the US were more likely to need and use interpersonal, school-related, and mental health services, and less likely to need shelter and healthcare services compared to their experienced homeless US peers. AU newly homeless youth were more likely to need interpersonal, work-related, and shelter services compared to their counterparts in the US. Experienced AU homeless youth were significantly more likely than their US experienced counterparts to use substance, general health, work, mental health, and shelter services.
There are cross-national differences and differences as a function of time being homeless among these youth. It would have been desirable to recruit in identical fashion across sites to ensure that there was no selection bias in recruitment. However, the cross-national differences in the organization of the system-of-care prohibited identical recruitment strategies. The inclusion criteria, definitions of newly and experienced homeless, and measures were identical and allow us to compare similar samples from the two sites.
Consistent with other research on homeless youth, youth in both countries were found to engage in more risk behaviors than youth in the general population. Comparing rates of risk behaviors among homeless youth in the US to the rates for youth from the general population (Centers for Disease Control and Prevention, 2000; Centers for Disease Control and Prevention, 2002) on the Youth Risk Behaviors Survey (YRBS), homeless youth reported having four or more sexual partners at 3 to 4 times the national norm; teenage parenthood was 2–3 times more likely among homeless youth; and marijuana and hard drugs were used twice as often. Similarly, rates reported by the Australian Institute of Health and Welfare show that homeless youth in AU report about 4–6 times the rate of hard drug use compared to non-homeless peers in national data sets, as well as 1.5–1.75 times the rate unprotected sex, and 4–5 times the rate suicide attempts (Australian Institute of Health and Welfare, 2000; Whitbeck, Hoyt, Yoder, Cauce, & Paradise, 2001). While the focus of most of these results is on the differences between the two countries, there were several notable similarities between the two samples: pregnancy history, the number of HIV-positive youth, the number of youth who only used alcohol and marijuana, and the number of youth attending school.
When comparing across countries, homeless youth in the US were younger, more likely to be in school and in jail, demonstrated fewer sexual and substance use risk acts, had committed fewer suicidal acts, and reported less need for social services. In addition, US homeless youth had parents who were more highly educated, and who were more likely to be non-English speaking and to be single. As the multivariate model reveals, there are a larger number of important independent factors that differentiate the two nation's populations of homeless youth. Moreover, the multivariate model reporting the independent factors associated with experience on the streets shows that there are also a large number of independent factors that distinguish the newly homeless youth from experienced homeless youth. These data suggest there is a window of opportunity to intervene with newly homeless youth, prior to youth's exposure to experienced homeless peers who are likely to introduce the youth to antisocial and risky behavior patterns (Whitbeck et al., 2001). As the literature suggests, as youth spend more time on the streets they become increasingly involved in risky drug use and risky sex, and become distanced from the normal trajectories of adolescence, such as attending high school (e.g., Whitbeck et al., 2001).
Across sites, experienced homeless youth were more likely than newly homeless youth to be older and male. In addition, experienced youth engaged in a greater number of risky behaviors in contrast to newly homeless youth; they were more likely to have been pregnant or to be a parent, and to engage in sexual acts and activities that resulted in incarceration. Experienced homeless youth had a greater need for social services, and utilized more work, substance use, and health-related services. Newly homeless youth in AU were much more likely to be older, to have become substance abusers, to have engaged in risky sexual practices, and to be in greater need of a range of social services (e.g., physical health, mental health, shelters, substance abuse, and employment) than their newly homeless American peers. Although this study cannot assert that the service system is responsible for these differences, it appears that youth may not become homeless in AU until a more serious risk profile has developed.
In the US, experienced homeless youth were more likely to be gay-identified and to have been in jail, and less likely to be from single parent families than either newly homeless youth in the US or Australian homeless youth, new or experienced. Almost one in three experienced youth in this US (29%) sample identified as gay or bisexual, double the rate among newly homeless US youth. Gay or bisexual youth may be more likely to have been ejected from their home or to have had their sexual orientation present more barriers to returning; a longitudinal follow-up is currently examining this hypothesis. Similarly, based on the results reported in Table VI, newly homeless youth were more likely to come from single parent families, to have non-English speaking parents, and to be female. Based on the comparison of newly homeless and experienced youth, there appears to be a subgroup that remains homeless. Again, only a longitudinal design can answer if there is bias in the rates of returning.
Australian homeless youth receive a government stipend as a living allowance; more than half of newly homeless youth in this study and four of five experienced youth were receiving a stipend. The types of measures monitored in this study did not identify the extent to which the government stipend for homeless youth in Australia helped those youth contend with the problems they faced.
Homeless youth in the US rated their physical health and mental health better than homeless youth in AU. However, homeless youth in AU have universal access to health care, in contrast to healthcare services that are very difficult to arrange for homeless youth in the US (Hart-Shegos & Ray, 1999). The rates of testing for HIV, hepatitis C, pregnancy, and childbearing reflect cross-national differences in access to healthcare for adolescents in the two countries. It is not clear from this study whether these differences exist only for homeless youth or for all adolescents in each country. American experienced youth were more likely to have been tested for HIV than experienced Australian youth. Although the pregnancy rates were similar for newly homeless youth, experienced youth in the US were far more likely to have given birth to a child, consistent with cross-national data (Myers, Rossiter, & Rosenthal, 2001).
There is a narrow window of opportunity, especially in the US, to intervene with newly homeless youth and to deter them from a pathway of chronic homelessness. The differences in risk behaviors and service utilization between newly and experienced homeless youth suggest that social service programs may be designed that can effectively interrupt the pathways toward chronic homelessness. Newly homeless youth from this study are currently being followed longitudinally to identify such pathways.
National Institute of Mental Health grants #1ROI MH49958-04, R01 MH061185-05, and P30MH58107 supported this study.
4There are two equally compelling yet contradictory general hypotheses regarding cross-national differences between homeless youth in AU versus US: (1) Australian youth may have better overall health and behavioral profiles because they live in a society where the social service sector is more accommodating of youth and stresses a harm reduction approach; (2) Australian youth may have worse overall health and behavioral profiles than American youth because, living in a society where the social service sector is better developed and more supportive of youth, being homeless is an indicator of more troubled youth who have fallen through the cracks.
5Most research on homeless adolescents has focused on chronic experienced homeless adolescents who have been out of home for extended periods of time or have had multiple episodes of leaving home. The distinction between newly homeless young people who have been out of home for a short period of time and chronic experienced homeless young people has been often overlooked in the research literature (Milburn, Rotheram-Borus, Brumback, Mallett, Rosenthal, Witkin, & May, Unpublished data). The distinction that has been made most often among homeless adolescents has been between young people who are literally on the streets and young people in shelters (Kipke, Montgomery, Simon, & Iverson, 1997; Greene, Ringwalt, & Iachan, 1997). Time out of home is implied within a classification of homeless young people based upon the location but it is not the primary classifying criterion. Classifying homeless young people by time out of home provides a greater understanding of the heterogeneity of homeless young people.
6Based on a previous study which audited homeless youth in Los Angeles (Witkin et al., 2005), the six month cut-off point for “newly” homeless youth emerged as a naturally occurring division in this population.