Predictors of depression in street youth.

Smart, Reginald G. ; Walsh, Gordon W.

The past few years have seen an increased interest in studies of mental illness among homeless populations. Some mental illness may be attributed to homelessness (Morrissey & Dennis, 1986), and the deinstitutionalization of the mentally ill has led to many becoming homeless (Reich & Siegel, 1978). The stresses of homelessness and the lack of support systems mean that the homeless are constantly threatened psychologically as well as physically (La Gory et al., 1990; Rossi et al., 1987; Radford et al., 1989). Several studies have determined that the rate of depression is greater among the homeless in general (Burnam & Koegel, 1988; Rossi et al., 1987), but few studies have been undertaken on homeless or street youth. None of those that are available provide detailed information on the predictors of depression. This paper presents the results of a study of factors related to depression among street youth in Toronto, Canada.

Depression and other psychiatric problems among the homeless or street population have been studied in several cities. These studies show that homeless, bowery, or shelter populations have high rates of psychiatric problems. Fischer (1989) reviewed 75 studies of psychiatric status in homeless populations conducted since 1980 and found that the rates of psychiatric problems varied between 10% and 90%, but the mean seemed to be around 50%. Almost all of these studies involved adults, and few focused on depression, the most common psychiatric problem. Vernez et al. (1988) found that 22% of the homeless adults in three areas of California had a major affective disorder. Hier et al. (1990) researched social adjustment and symptomatology among 52 runaway and throwaway adolescents in Brisbane, Australia. Clinical levels of depression were found for both types of homeless adolescents. La Gory et al. (1990) studied depression among homeless adults in Alabama. Almost 60% showed the signs of "probable clinical caseness" on the Center for Epidemiological Studies Depression Scale. Depression was greater among younger persons, those who had experienced more undesirable life events and more hospitalizations for mental illness, and those with fewer social supports. Alcohol and drug problems were not included as variables in that study, although many studies have shown them to be common among the homeless (Fischer, 1989).

Several studies have shown that depression and substance abuse are related among adolescents in general and in college and school populations. Crumley (1990) reviewed studies of adolescent substance abuse and suicide and found that the degree of association varied from one study to another, but up to 43% of actual suicides among adolescents were related to substance abuse. Levy and Deykin (1989) found that major depression and substance abuse were independent and interactive risk factors for suicide in college students.

In some studies, high levels of depression and suicidal ideation have been found among homeless youth. Robertson et al. (1989) noted that 26% of street youth in Hollywood were depressed, based on DSM-III criteria as assessed by the Diagnostic Interview Schedule. Radford et al. (1989), in a study in Canada, also found that about a quarter of street youth had suicidal thoughts. Neither of these studies analyzed the factors related to depression, such as drug abuse. However, Stiffman (1989) studied lifetime suicide attempts among runaway youth seen at shelters in St. Louis, finding that suicide attempts were related to sex (female), substance abuse, negative life events, running away, behavior problems, and family instability, but not to coping ability. No measures of alcohol or drug problems were used, but scores on a composite drug-use variable were positively related to suicide attempts.

The present paper reports the results of a study of current depression among street youth in Toronto. It attempts to extend knowledge of this group by examining the association of depression with alcohol and drug use and related problems, social supports, self-esteem, family background, and alcohol and drug use among family members. It was expected that depression would be greatest among street youth who (1) used drugs frequently and had alcohol and drug problems, (2) had family members who used drugs, (3) had the weakest social supports and self-esteem, (4) were on the street the longest, and (5) came from the most unstable family backgrounds.

METHOD

Sample Criteria

A most perplexing problem faced by researchers in studying street youth is setting criteria for selection. Rather than being a categorical entity, we see street involvement as a continuum, with youth drifting in and out at various times. Thus, we chose to capture a broad spectrum of the population by setting the following criteria for participation in the study. The primary criterion was that all participants had to be 24 years old or younger. This reflected a commonly used age requirement for youth-specific social service facilities. American surveys of urban centers suggest that this age group represents from 12% to 20% of the entire homeless population (Piliavin, Westerfelt, & Eliott, 1989; Snow, Baker, & Anderson, 1989). An index approach was used for establishing the secondary criteria, which were as follows:(1) participants must have used at least one social-service facility directed toward street youth in their lifetime; (2) they must have left school before completing Grade 12; (3) they must have lived away from their family (or guardian) at least two days during the past year; (4) they must have run away or been thrown out of their home at least once; and (5) they must have been homeless (i.e., without a place to stay) at least once.

Youths were interviewed if they responded affirmatively to the first of the five secondary criteria, or if they responded affirmatively to three or more of the other four. This procedure ensured that people such as students hanging out temporarily were excluded, and that street youth who used the social-service system directly or indirectly were included. In fact, almost all those interviewed easily satisfied most of the criteria. Geographically, the study was restricted to youth located within, or agencies serving, the downtown core of Toronto.

Sample Design and Selection

Ideally, a study such as this should attempt to obtain a representative sample of street youth. Since no survey has scientifically estimated its size, the universe of street youth is unknown. In addition, since the population is highly transitory and street youth use many agencies, it is impossible to know the probability of selecting a given youth. Although we were unable to derive a probability sample, where possible we used randomization in selecting the 145 youth whom we interviewed.

To select agency-derived youth, a two-stage process was used. With the guidance of the Coalition of Youth Work Professionals, a sampling frame of about 45 agencies serving the Toronto downtown core was constructed. In the first stage, 11 youth agencies from this list were randomly selected. "Youth agency" was defined as any facility providing social and other services to young people, including both residential (e.g., shelters and hostels) and nonresidential (e.g., drop-ins, outreach, education/referral, counseling, and needle exchange) programs. Excluded from consideration were facilities for drug treatment exclusively, crisis phone lines, and those for special needs (e.g., battered women). Although no selected agency refused to cooperate in the study, three had few youths in their client population and were replaced with random selections.

The second stage involved selecting, preferably at random, a minimum of 10 youth from each agency. In many cases, random selection was not possible; thus, all who volunteered were interviewed. A total of 108 youth (75% of the total) were interviewed from agencies.

Although the agency-derived sample provides invaluable information on facility usage, it may not represent street youth who fail to use these services. Both the size and proportion of those who do not use services are not well-established. American studies have found that anywhere from 17% to 63% spend their nights on the street (Rossi et al., 1987; Bart & Cohen, 1989; La Gory, Ritchey, & Mullis, 1990). The size of this group appears to vary depending upon gender and season. In Chicago for instance, it was found that 59% of the homeless were street dwellers during the fall versus 26% during the winter months (Rossi et al., 1987).

We used several methods to draw a sample of youth derived from the street (as opposed to agencies). First, a seasoned street worker was employed to approach and interview apparent street youth by "cold" contacts within the study's geographical area. Second, agency personnel at mobile outreach vans informed youth of our study. Finally, a word-of-mouth snowball sample was obtained by asking participants from the van-derived sample to inform others of our interest in interviewing youth like themselves. In total, 37 street-derived youth (25% of the total) were interviewed (10 via cold contact, 18 van-derived, and 9 by word-of-mouth).

Interviews lasted an average of 75 minutes and were conducted during a six-week period in February and March, 1990. We expect that the cold weather during this time minimized the size of the unhoused street population and increased the proportion of agency users. All youth were interviewed individually by male and female interviewers trained in the objectives of the study. All participants received $20 for completing the interview.

Questionnaire and Variables

Prior to formal interviewing, the questionnaire was pretested among agency-derived youth and evaluated by professional youth workers. We believe that the responses of the youth were, on the whole, honest and forthright. In many ways, these youth had nothing to lose by confiding in our interviewers. The $20 payment influenced participation for some, although many stated afterward that they would have done so without it.

The primary variable of interest, depression, was measured using 4 items from the Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977; Mechanic & Hansell, 1987) for the three-month period prior to the interview: (1) How often have you felt sad? (2) How often have you felt lonely? (3) How often have you felt depressed? and (4) How often have you felt like crying? Responses were made on a 4-point scale: (1) rarely, (2) sometimes, (3) often, or (4) always. The range of possible summated scores extended from 4 to 16. Although we employed a three-month interval (versus a seven-day period) for the CES-D, the reliability coefficient (Cronbach's alpha) for the depression scale was .77.

Five other scales were used to measure the independent variables. Social support was calculated by summing the responses (1 = no, 2 = yes) to the following:(1) Do you have anyone you feel you can confide in? and (2) Is there anyone you could depend upon if you became ill? The possible summated scores ranged from 2 to 4. The reliability coefficient was .57.

Self-esteem was measured by summing the answers (0 = false, 1 = true) to two questions from the Rosenberg Self-Esteem Scale (Rosenberg, 1989): (1) I feel good about myself, and (2) I feel I am a person of worth. The possible range of summated scores was from 0 to 2, and the reliability coefficient was .63.

Alcohol problems were measured by the 4-item CAGE scale (Mayfield et al., 1974). Subjects were asked to respond positively or negatively to the following questions: (1) Have you ever felt you should drink less? (2) Have others ever bothered you by complaining about your drinking? (3) Have you ever felt bad or guilty about your drinking? and (4) Have you ever drank in the early morning to get rid of a hangover? Positive responses were summed, resulting in values from 0 to 4. Clinical evidence suggests that a score of 2 or above is indicative of problem drinking (Mayfield et al., 1974).

Drug problems were measured by summing positive responses to the following seven questions, including four items from the Drug Abuse Screening Test (Skinner, 1982), for the 12 months prior to the interview: (1) Do you think you could use drugs less than you do now? (2) Are you always able to stop using drugs when you want to? (3) Have you ever gone to anyone for help for a drug problem? (4) Have you ever seen a doctor or been in a hospital because of your drug use? (5) Have you ever had "blackouts" or "flashbacks" due to your drug use? (6) Have you ever had any medical problems as a result of your drug use? and (7) Have you ever been arrested or warned by police because of your drug use? Summated scores could range from 0 to 7. The reliability coefficient for drug problems was found to be .62.

Family instability was computed by summing the positive responses to the following possible living arrangements: (1) Have you ever lived in a group home? (2) Have you ever lived in a detention center? (3) Have you ever lived in a foster home? and (4) Have you ever lived in a hostel/shelter? Possible values ranged from 0 to 4. The reliability coefficient for family instability was .48.

RESULTS

Table 1 presents selected characteristics of the Toronto sample. Two-thirds (64%) were male, and they ranged in age from 13 to 24 years (average age = 19 years). As noted earlier, 75% were derived from social-service agencies and 25% were derived from the street. The number of social-service facilities used in the youths' lifetime ranged from 0 to 16. Three percent used no services, while 26% reported having used 10 or more facilities in their lifetime. The prevalence of "literal" homelessness (i.e., the percentage who did not have access to conventional, permanent housing when interviewed) was 45% (35% resided with agencies and 10% lived on the street). During the 13 months prior to the survey, the youth reported spending an average of 1.4 months on the street. Sixty-three percent did not live on the street during this 13-month interval, while 8% lived on the street less than one month, 11% from one to three months, 11% from four to six months, and 6% seven or more months.

Table 1 also shows that a substantial proportion had lived in a variety of settings other than with their biological parents, had left home at a relatively early age, and had run away from home more than once. About one-third of the sample reported feelings of depression often or always during the three months prior to the interview, and 42% had attempted suicide at least once in their lifetime.

Table 1. Selected Sample Characteristics of Toronto Street
Youth (N = 145)
 %
Gender
 Male 64
 female 36
Age
 |is less than~ 16 years 5
 16-18 years 41
 19-21 years 42
 22-24 years 12
Sample
 Agency 75
 Street via Agency Van 12
 Street 7
 Word of Mouth 6
Number of Social Services
Used in Lifetime
 0 3
 1-2 15
 3-5 23
 6-9 34
 10+ 26
Current Homelessness
 Agency 35
 Street 10
Grade Left School
 |is less than~ Grade 9 12
 9-11 54
 12 + 12
 Currently Enrolled 22
Have Ever Lived...
 with biological parents 86
 with adoptive parents 21
 in a foster home 32
 in a group home 41
 with other relatives 48
 in detention center 50
Times Left Biological
Parents(a)
 1 time 34
 2 times 11
 3 times 14
 4 + times 41
Age First Left Biological
Parents(a)
 10 years or less 19
 11-15 years 53
 16 + years 28
Depression(b)
 Felt sad 37
 Felt lonely 36
 Felt depressed 30
 Felt like crying 26
Ever Attempted Suicide 42
Worked During Past Two Weeks 43
a Among the 18% who had left home.
b Percentage reporting item often or always during the prior
three months.

Table 2 presents mean depression scores for selected variables. Contrary to expectations, depression was significantly related to only three of the twelve variables tested and marginally significant for one of the variables. The correlation coefficient for self-esteem (r = -.397) indicates a significant inverse relationship with depression. As anticipated, those with high self-esteem were significantly less depressed than those with moderate and low self-esteem. The findings also showed that social support (r = -.188) had a significant inverse relationship with depression. Those with a high level of social support were significantly less depressed than those with medium and low levels of social support. Furthermore, the number of months lived in a hostel TABULAR DATA OMITTED (r = .206) was found to have a significant direct relationship with depression. Seemingly, those who had spent the least amount of time in a hostel were the least depressed. Finally, a one-way ANOVA indicated that parents' drug use as related to the respondent's leaving home had a marginally significant association with depression (p = .075). Those youth for whom parents' drug use was related to leaving home were more depressed.

Table 3. Regression of Depression Score on Selected Variables
Regressors b b(*) p
Age .034 .032 .692
Gender -.599 -.103 .229
Social Support -.353 -.096 .267
Self-Esteem -1.287 -.352 .000
Alcohol Problems (CAGE) .138 .066 .460
Drug Problems .082 .053 .558
Annual Number of Drugs Used -.029 -.021 .824
Own Drug Use Related to Leaving -.310 .045 .613
Parents' Drug Use Related to Leaving 1.428 .133 .101
Family Instability -.261 -.110 .182
Months Lived on the Street .048 .047 .567
Months Lived in Hostel .189 .171 .040
Constant = 11.169
|R.sup.2~ = .242
F = 3.370
p = .000
N = 140
Note: b, regression coefficient; b(*), standard coefficient.

Table 3 shows the results of regressing the depression score on selected variables. Taking into account the other factors in the regression, the relationship between social support and depression was no longer significant (p = .267). However, parents' drug use as related to the respondent's leaving home still bordered on signficance, although weaker (p = .101). Consistent with earlier results, self-esteem again had a strong inverse relationship with depression (p = .000). Finally, the regression confirmed the statistically significant relationship between months having lived in a hostel and depression (p = .040).

DISCUSSION

The results show that, among street youth, only self-esteem, social support, and time spent in a hostel were clearly related to depression, with parents' drug use as related to the respondent's leaving home being marginally significant. When the regression analysis was performed, social support was no longer important, and depression was found to be greatest among those with low self-esteem and those who had spent the greatest amount of time in hostels.

We had predicted that depression would be greatest among those who used drugs heavily or had alcohol and drug problems, because drug abuse and suicide have been found to be related among college students and the general population. Street youth have high rates of both drug use and depression. However, no relationships between depression and any of the alcohol or drug-use variables were found. Since no previous study of street youth and depression has been done, an explanation for these results is difficult. It may be that drug use and problems are so ubiquitous among street youth that there is too little variation and a ceiling effect is found. For example, 92% of street youth used cannabis, 70% used LSD, and 64% used cocaine in the past year. Only 7% of 145 used no drugs in the past year and only 12% had no drug problems. We might note here that even though street youth had considerable drug problems, they did not see them as their main problems, which were money, food, housing, clothes, and employment.

Stiffman's (1989) study of runaways examined predictors of suicide attempts, not depression among street youth. She found that a composite drug-use score did predict lifetime suicide attempts, but alcohol and drug problems were not assessed. Our study examined current depression (past three months) and found that the correlation with lifetime suicide attempts was only .34. Although significant, this correlation is not so high as to preclude some different predictors from being important for lifetime suicide attempts and current depression.

Our results for factors related to psychological and parental support variables are somewhat in keeping with the predictions. There are suggestions that street youth are more depressed if parents' drug use caused them to leave home. Also, high self-esteem and a high level of social support protect against depression, as expected. These findings are similar to those of Stiffman (1989) with runaways and La Gory et al. (1990) with the homeless of all ages. However, family stability was not an important factor in current depression, nor was the amount of time youth spent on the street.

The best indicators of current depression among street youth, taking other variables into account with the regression analysis, were self-esteem and amount of time spent in hostels. These findings may result from complex relationships. Seriously depressed street youth probably have low self-esteem because they are unable to succeed in school or at work due to their depression; this may lead to even lower self-esteem. Depressed street youth are probably less able to cope with the problem of accommodations and more often require hostels. Staying in hostels, which are sometimes dirty, noisy, dangerous, and overcrowded, would lead to more depression.

Further research on depression among street youth should examine whether depression leads to entering a life on the street, and whether reducing depression is related to leaving the street. It would also be helpful to know more about depression and suicide attempts in family members to determine whether there is a familial pattern.

The views expressed in this paper are those of the authors and do not necessarily reflect those of the Addiction Research Foundation.

Gordon W. Walsh, Addiction Research Foundation, Toronto, Ontario, Canada.

Reprint requests to Reginald G. Smart, Ph.D., Head, Social Epidemiology, Addiction Research Foundation, 33 Russell Street, Toronto, Ontario, Canada M5S 2S1.

REFERENCES

Bart, M. R., & Cohen, B. E. (1989). Differences among homeless single women, women without children, and single men. Social Problems, 36, 508-524.

Burnam, A., & Koegel, P. (1988). Methodology for obtaining a representative sample of homeless persons: The Los Angeles Skid Row Study. Evaluation Review, 12(2), 117-152.

Crumley, F. E. (1990). Substance abuse and adolescent suicidal behavior. Journal of the American Medical Association, 236, 3051-3056.

Fischer, P. J. (1989). Estimating the prevalence of alcohol, drug and mental health problems in the contemporary homeless population: A review of the literature. Contemporary Drug Problems, 16, 333-389.

Hier, S. J., Korboot, P. J., & Schweitzer, R. D. (1990). Social adjustment and symptomatology in two types of homeless adolescents: Runaways and throwaways. Adolescence, 25(100), 761-771.

La Gory, M., Ritchey, F. J., & Mullis, J. (1990). Depression among the homeless. Journal of Health and Social Behavior, 31, 87-101.

Levy, J. C., & Deykin, E. Y. (1989). Suicidality, depression, and substance abuse in adolescence. American Journal of Psychiatry, 146(11), 1462-1467.

Mayfield, D., McLeod, G., & Hall, P. (1974). The CAGE questionnaire: Validation of a new alcoholism screening instrument. American Journal of Psychiatry, 131, 1121-1123.

Mechanic, D., & Hansell, S. (1987). Adolescent competence, psychological well-being and self-assessed physical health. Journal of Health and Social Behavior, 28, 364-374.

Morrissey, J. P., & Dennis, D. L. (1986). NIMH-funded research concerning homeless mentally ill persons: Implications for policy and practice. Washington, DC: Alcohol, Drug Abuse and Mental Health Administration.

Piliavin, I., Westerfelt, H., & Elliott, E. (1989). Estimating mental illness among the homeless: The effects of choice-based sampling. Social Problems, 36, 525-531.

Radford, J. L., King, A. J. C., & Warren, W. K. (1989). Street youth and AIDS. Ottawa: Health and Welfare Canada.

Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401.

Reich, R., & Siegel, L. (1978). The emergence of the Bowery as a psychiatric dumping ground. Psychiatric Quarterly, 50, 191.

Robertson, J. J., Koegel, P., & Ferguson, L. (1989). Alcohol used and abuse among homeless adolescents in Hollywood. Contemporary Drug Problems, 16, 415-453.

Rosenberg, M. (1989). Society and the adolescent self image (rev. ed.). Middletown: Wesleyan University Press.

Rossi, P. H., Wright, J. D., Fisher, G. A., & Willis, G. (1987). The urban homeless: Estimating composition and size. Science, 235, 1336-1341.

Skinner, H. A. (1982). The drug abuse screening test. Addictive Behaviors, 7, 363-371.

Snow, D. A., Baker, S. G., & Anderson, L. (1989). Criminality and homeless men: An empirical assessment. Social Problems, 36, 532-549.

Stiffman, A. R. (1989). Suicide attempts in runaway youths. Suicide and Life-Threatening Behavior, 19(2), 147-159.

Vernes, G., Burnam, M. A., McGlynn, E. A., Trude, S., & Mittman, B. S. (1988). Review of California's program for the homeless mentally disabled. Santa Monica: RAND Corporation.