Mental Health Social Env And Sexual Risk Behaviors

Journal of Child and Family Studies, Vol. 6, No. 1, 1997, pp. 9-25

Mental Health, Social Environment and Sexual Risk Behaviors of Adolescent Service Users: A Gender Comparison Yu-Wen Chen, Ph.D.,1 Arlene Rubin Stiffman, Ph.D.,2,5 Li-Chen Cheng, Ph.D.,3 and Peter Dore, MA.4

We examined individual mental health problems (depression, conduct disorder, and substance abuse) and social environment (family, peer, and neighborhood) factors associated with the sexual risk behaviors of male and female adolescents. Interviews with 778 adolescents, aged 14 to 18, showed that both mental health problems and social environment were related to adolescents' involvement in sexual risk behaviors. Conduct disorder symptoms, substance abuse or dependence symptoms, and the interaction between peer misbehavior and neighborhood problems were significantly associated with risky sexual behaviors. Peer misbehavior was a particularly strong factor related to sexual risk behaviors for youths who lived in neighborhoods with multiple problems. The only gender differences were found in age, with older males more likely to report engaging in high risk sexual behaviors. This study suggests the utility of multidimensional intervention strategies to deal with various adolescent problem behaviors, including risky sexual behaviors, within the context of their social environment. KEY WORDS: adolescents; sexual risk behaviors; gender differences; mental health problems; social environment.

1Assistant

Professor, Department of Sociology, National Taiwan University, Taipei, Taiwan. Professor, George Warren Brown School of Social Work, Washington University, St. Louis, Missouri. 3Assistant Professor, Department of Social Work, SooChow University, Taipei, Taiwan. 4Statistician, George Warren Brown School of Social Work, Washington University, St. Louis, Missouri. 5Correspondence should be directed to Arlene Rubin Stiffman, George Warren Brown School of Social Work, Campus Box 1196, Washington University, St. Louis, Missouri 63130.

2Associate

9 1062-1024/97/0300-0009$12.50/0 c 1997 Human Sciences Press, Inc.

10

Yu-Wen Chen et al.

The health, economic and social consequences of adolescents' sexual risk behaviors have highlighted the need to find associated behavioral and environmental indicators in order to target early and effective identification of risk and intervention for that risk. This study examines the differential indicators for male and female adolescents' sexual risk behaviors. Specifically, we examine how, and to what extent, individuals' mental health problems and their social environments are associated with their involvement in risky sexual behaviors. Risky sexual behaviors can lead to several major problems: unplanned pregnancy, HIV infection, and other sexually transmitted diseases. Early pregnancy is costly to the adolescents, their infants, and our society. For example, early pregnancy often truncates the adolescents' educational attainments and employment opportunities (Hayes, 1987; Luster & Small, 1994). Infants born to unmarried adolescent mothers suffer higher rates of mortality and mental retardation, and have lower IQs and birth weight (DiBlasio & Benda, 1990; Gerber & Resnick, 1988). In the United States, the societal costs of teen pregnancy are estimated to be about $200 million a year (Edelman, 1987; Weddle, McKenney & Leigh, 1988). With the threat of the AIDS epidemic, concern about adolescent risky sexual behavior is no longer limited to unplanned pregnancy (Hein, 1993). About 21% of AIDS cases are found among people in their early 20s. Because HIV has an incubation period of 8 to 10 years, many young adults may have been infected in their adolescence (Botvin & Dusenbury, 1992; Centers for Disease Control, 1993). Risky sexual behaviors continue to be the primary source of HIV transmission for adolescents and young adults (Catania et al, 1992). In addition to AIDS, the incidence of other sexually transmitted diseases (e.g., chlamydia, genital warts, gonorrhea, syphilis, and trichomonas) in the United States has also been growing rapidly in the last few years, with adolescents experiencing the greatest increases (Cates & Stone, 1992). It is reported that by age 21, approximately one of every five youths has required treatment for at least one type of sexually transmitted disease (National Center for Health Statistics, 1994). The prevalence and consequences of adolescents' sexual risk behaviors discussed above have highlighted the need to find variables correlated with such behaviors in order to target early and effective intervention. Although numerous studies have been conducted on adolescents' sexual behaviors in the United States, few studies focus on at-risk populations. Sexual risk behaviors are known to be more frequent in certain populations or neighborhoods than in others (Barlett, Keller, Eckholdt, & Schleifer, 1995; Centers for Disease Control, 1993). In fact, many of the adolescents who engage in risky sexual behaviors also have had some contact with potential

Mental Health Social Environment

11

helping professionals, such as primary care physicians, school counselors/social workers, juvenile court officers, or child welfare workers. These adolescents are generally at a higher risk for mental health problems and other problem behaviors, and are, therefore, most urgently in need of professional attention (Elliott & Morse, 1989; Ensminger, 1990; Jessor & Jessor, 1977; Rosenbaum & Kandel, 1990). Research findings suggested by rigorous empirical data on this population can help maximize the development of effective prevention and intervention. Previous studies which support a relationship between mental health problems, social environment and adolescent sexual risk behaviors provide the basic rationale for this study. Many studies have shown that adolescents' risky sexual behaviors are related to mental health problems. For example, risky sexual activities are strongly associated with alcohol and drug use (Biglan et al, 1990; Booth, Watters, & Chitwood, 1993; Edlin et al., 1992; Harvey & Spigner, 1995; Hudgins, McCusker, & Stoddard, 1995; Leigh, 1990; McEwan et al., 1992); depression (Flavin, Franklin, & Francis, 1986; Ostrow et al., 1989; Simpson, Knight, & Ray, 1993; Stiffman et al., 1992,1987); conduct disorder (Baker & Mossman, 1991; Greenblatt & Robertson, 1993); and suicidality (Kipke, Montgomery, & MacKenzie, 1993; Yates et al, 1991). A number of theories (e.g., social learning theory, social-ecological theory, and reference group theory) and many studies support the relationship between environment and risky sexual behaviors (Bandura, 1977,1992; Bronfenbrenner, 1979; Mirande, 1968). The social context can play a more important role than the individual context in influencing human behavior. One's behavior may be the result of interactions with other people or other influences from their social environment (Billy & Udry, 1985; Brown et al, 1986; Mirande, 1968; Nathanson & Becker, 1986; Oskamp & Mindick, 1983). A few studies address the broader impact of community or neighborhood traits on adolescents' involvement in high-risk sexual behaviors (Billy, Brewster, & Grady, 1994; Brewster, Billy, & Grady, 1993; Ku, Sonensten, & Pleck, 1993). In general, associated community or neighborhood traits include socioeconomic resources, employment opportunities, and social disorganization (i.e., single-parent families, unemployment, and crime rates in a community). Ineffective schools, the lack of social services, and community deprivation limit poor adolescents' life chances and their expectations for future success. This enhances the likelihood that they might engage in sexual risk behaviors. In order to develop additional avenues of intervention in adolescents' sexual risk behaviors, one must understand these contextual or social variables. Until recently, few studies examined multiple correlates of sexual risk behaviors. Some of these studies found that factors associated with adolescent sexual risk behaviors include depression, antisocial behaviors, illicit drug or alcohol use, anxiety, suicidality, and variables related to adoles-

Yu-Wen Chen et al.

12

cents' family, peers or friends (Biglan et al., 1990; Stiffman et al, 1995). An earlier study by Stiffman et al. (1995) showed that both individual mental health problems and neighborhood drug dealing and murders were positively associated with change in HIV risk behaviors between adolescence and young adulthood. Previous studies supported the relationship between gender differences in perceptions of sexual risk. Calvert and Stanton (1992) found that male adolescents tended to think that having a baby would make them feel good, but only a few female adolescents thought so. Carver, Kettelson and Lacey (1990) also found significant gender differences in attitudes and beliefs. Our study expands on these studies by examining differences in the determinants of male and female adolescents' actual sexual risk behaviors. Although numerous studies have been conducted on adolescents' sexual risk behaviors, most have focused on only one or two categories of variables; few studies have examined the simultaneous multivariate relationship of variables at the individual and social levels. A major limitation of this narrow approach is that the relative explanatory power of different variables cannot be examined. In addition, though gender differences are usually examined in studies of adolescent risk or problem behaviors, such as conduct disorder and drug or alcohol use, gender comparisons of factors associated with sexual risk behaviors have been ignored. Given what is known and is not known from previous research, this study uniquely assesses adolescents' engagement in sexual risk behaviors as a function of the problems they experience at the personal and social levels. Similarities and differences across gender groups were also studied. Our sample was comprised of inner-city adolescents who were using gateway public services at the time of our interview. As discussed above, these adolescents are generally at a higher risk for mental health problems and other problem behaviors, and therefore, most urgently in need of effective intervention (Elliott & Morse, 1989; Ensminger, 1990; Jessor & Jessor, 1977; Rosenbaum & Kandel, 1990). Moreover, our study includes neighborhood effects and examines the interactive effects of various environmental variables. To date, ours is the only study to examine the differential multivariate association of these factors with male versus female adolescent sexual risk behaviors. METHOD Design Adolescents from the city of St. Louis, Missouri were interviewed between April and October of 1994. All had used services (i.e., seeing by

13

Mental Health Social Environment

social workers, counselors, or health care workers) provided by gateway service sectors: child welfare, primary health care, juvenile justice, and education. The term 'gateway' refers to services within non-mental health service sectors which are provided by trained professionals. These professionals could potentially, but often do not, identify the mental health problems of youths they serve, and then either provide them with mental health services or link them, through referrals, to mental health services available elsewhere. Data were collected through face-to-face interviews by 15 trained professional interviewers, most were social work graduate students. All interviewers were trained in a three-day workshop, practicing with teen volunteers during training. Respondents were recruited with the aid of service providers, by having interviewers approach them in the service sector waiting rooms, and through letters and posters requesting volunteers from each sector's service users. Over 95% of the youths approached directly assented to the interview. Informed consent was obtained from all subjects and their guardians prior to interviews. When possible, the interviewer completed the interview on site, immediately before or after services were obtained. Otherwise, the interviewer arranged an appointment for a future interview in the adolescent's home, or at a mutually acceptable site. Subjects

A total of 796 subjects were interviewed, and after deleting cases with missing information, 778 were used in this study. As shown in Table 1, all subjects were between 14 and 18 years of age, with a mean age of 15.3. Thirteen percent of the adolescents interviewed were White (n = 102) and 87% African American (n = 676). Forty-two percent (« = 328) were male, and 58% (n = 450) female. Comparisons with tallies of the demographics of adolescents using the four gateway sectors found that the sample was quite representative of adolescent service sector users. Note, however, that they were not representative of the population of St. Louis city, which has a population that is closer to 50% white and 50% African American. Measures

Dependent Variable: Sexual Risk Behaviors In the present study, adolescents' sexual risk behaviors were operationalized as the behaviors that would put an adolescent at risk for unplanned pregnancy, or contracting AIDS and other STDs. On the basis

Yu-Wen Chen et al.

14

Table 1. Means, Standard Deviations, and Ranges of Variables (N = 778) Variable Sexual risk behaviors

Age

Standard Deviation

Range

0.36 15.3

0.58 1.18

14-18

9.90 3.34 2.24

6.16 2.97 4.97

0-25 0-17 0-33

9.29 1.88 19.07 6.17

5.24 1.95 4.00 3.84

0-25 0-10 5-25 0-14

Mean

Mental Health Problems Depression symptoms Conduct disorder symptoms Substance abuse symptoms Social Environment Peer misbehavior Stressful life events Family support Negative neighborhood

0-3

of this operationalization, the measure of sexual risk behaviors was comprised of four discrete behaviors. These included having multiple partners in the last 6 months, getting pregnant or impregnating someone, contracting sexually transmitted diseases, and engaging in prostitution. Each of these four behaviors may occur independently, but each adds to potential risk for the youth. Independent Variables Mental Health Problems Measures of depression, conduct disorder, and substance (drug and alcohol) dependence or abuse came from the Diagnostic Interview Schedule for Children-Revised (DISC-R: Shaffer, Schwab-Stone, Fisher, Cohen et al., 1993). The DISC-R allows two separate operationalizations of these constructs: 1) a diagnosis of a specific disorder based on computer algorithms that combine symptoms according to the criteria in DSM-IV; and 2) a count of symptoms lasting 2 weeks or more, or which interfere significantly with the adolescent's life (a = 0.78 for depression, a = 0.89 for conduct disorder, and a = 0.91 for substance abuse). In this paper, we use the total number of symptoms. Social Environment Parent child relationships were assessed with an adaptation of the Family Satisfaction Scale (Hudson, 1982). Subjects were asked to rate how

Mental Health Social Environment

15

much of the time, over the course of the last 6 months, their family got on their nerves, they really enjoyed their family, they could depend on their family, their family argued too much, and they felt like a stranger in their family. Each item was rated on a 5 point Likert scale with 1 indicating never and 5 indicating almost or all of the time. Negatively worded questions were reverse scored prior to summing all scores. The sum, which ranged from 5 to 25, provided an index of the quality of family support and relationships (Cronbach's a = 0.68). Stressful life events were measured by a scale from the Diagnostic Interview for Children and Adolescents (DICA: Welner et al., 1987). Respondents were asked if they experienced any of the following ongoing stressful situations in the last 6 months: fighting and quarreling in the home, a family member's illness, poverty, intrafamily violence, a family member's death, family alcoholism/drug abuse, police problems, threats to self or family members, and beating or killing of friends or family members. The extent of stressful life events experienced by the youth was indicated by the sum of the above situations. This scale demonstrated construct validity in prior studies through its association with depression and suicidality (Stiffman & Earls, 1990; Welner et al., 1987), relationship problems (Stiffman, 1989), and AIDS risk behaviors (Stiffman et al., 1991; Stiffman et al., 1995). To measure peer misbehaviors, the youths rated, on a 5 point scale (none, a few, about half, most, or all), how many of their peers had trouble with the police, used drugs or marijuana, were both unemployed and out of school, used drugs/drank alcohol daily, and/or had become pregnant. Responses to these 5 questions were summed (Cronbach's a = 0.77). These peer misbehaviors have been found to be correlated with other problem behaviors experienced by youths, such as violence (Stiffman et al., 1996) and change in AIDS risk behaviors (Stiffman et al., 1992). Negative neighborhood environment was measured through adolescents' rating of their neighborhood problems on a 3-point scale ("none," "some," "a lot"). The summed items included shootings, murders, abandoned buildings, neighbors on welfare, prostitution, drug dealing, and/or homeless individuals (Cronbach's a = 0.84). Construct validity of this neighborhood scale was evidenced by its association with census tract information on population density, poverty, and housing deterioration. For example, the youths' perceptions of the number of neighbors on welfare were significantly correlated with census tract data concerning the proportion of households on public assistance (r = .38, p < .001), and the youths' perceptions of the number of abandoned buildings in their neighborhood

Yu-Wen Chen et al.

16

were significantly correlated with census tract data on the proportion of vacant housing units (r = .29, p < .001). In summary, most measures were adopted from standardized instruments. With respect to measures that were developed for the present study, about 30 iterations were completed prior to the collection of data and each measure demonstrated consistent construct validity in other studies. The whole questionnaire was also pilot-tested with teenagers to assure that it was clear and comprehensible.

Data Analyses Univariate analyses examined each dependent and independent variable. Several variables (e.g., stressful life events, substance abuse symptoms) were skewed. However, transforming the skewed variables did not change the model's significance. Therefore, the original untransformed variables were used in the analyses presented in this paper. Multiple regressions, using simultaneous ordinary least squares (OLS) procedures, examined the effect of each independent variable on sexual risk behaviors while controlling for the effects of other independent variables in the model. Previous studies found that people behave differently in different social contexts; that is, there are significant interactions between various environmental traits in influencing human behavior (Mason, Wong, & Entwisle, 1983). In general, peers and the family usually have an alternative effect on adolescent behaviors. When the adolescent's family life is in turmoil, their peer group will have stronger direct and indirect effect on adolescents' behavior (Coleman, 1981; Nathanson & Becker, 1986). Given such empirical support, we examined all possible interaction effects between the independent variables measuring the adolescents' peer, familial, and neighborhood environments.

RESULTS Univariate Analyses As shown in Table 1, adolescents in this study engaged in an average of 0.36 sexual risk behaviors in the last 6 months, with a range of 0 to 3 and a standard deviation of 0.58. The means, standard deviations, and ranges of all independent variables are listed in Table 1.

17

Mental Health Social Environment Table 2. Bivariate Relationships with Sexual Risk Behaviors (N = 778) Sexual Risk Behaviors Age Mental Health Problems Depression symptoms Conduct disorder symptoms Substance abuse symptoms Social Environment Peer misbehavior Stressful life events Negative neighborhood Family support

Variance Explained (R2)

.08*

.01*

.11*** .44*** .35***

.01* .19* .12*

.29*** .21*** .21*** -.10**

.09* .04* .01* .04*

*p < .05, **p < .01, ***p < .001.

Bivariate Analyses

Bivariate analyses demonstrated that sexual risk behaviors were significantly associated with all independent variables (i.e., age, depression symptoms, conduct disorder symptoms, substance abuse symptoms, peer misbehavior, stressful life events, family support, and negative neighborhood environment) except race (see Table 2). The more behavior or mental health problems that the youth had, the more likely the youth was to engage in sexual risk behaviors. Also, the more life problems (stress, negative neighborhood environment, and misbehaving peers), the more likely the youth was to engage in sexual risk behaviors. In contrast, the more family support the youth had, the less likely the youth was to engage in sexual risk behaviors. On average, males (X = 0.55) engaged in more sexual risk behaviors than females (X = 0.21; t = 8.07, p < 0.0001). Multivariate Analyses

A series of multivariate analyses were performed to examine the relationship of mental health problems and social environment with sexual risk behaviors. We first analyzed the overall model for both genders combined, and then analyzed it including gender-based interactions. Before multiple regression analyses were undertaken, we examined the potential problems of extreme values and multicollinearity. No extreme values were detected, and the tolerance estimates for each independent variable ranged from 0.51 to 0.92, indicating that multicollinearity is not a problem.

Yu-Wen Chen et at.

18

Both Genders The overall multivariate model included measures of subjects' gender, age, conduct disorder symptoms, substance abuse symptoms, peer misbehaviors, stressful life events, family support, negative neighborhood environments, and all potential interactions between independent variables. The final overall model accounted for 27% of the variance in sexual risk behaviors [F(10, 767) = 27.70, p < 0.0001] (see Table 3). Unique variance was explained by age (r2 = 0.004), conduct disorder symptoms (r2 = 0.038), substance abuse symptoms (r2 = 0.005), and negative neighborhood environments (r2 = 0.001). In contrast, depression symptoms, family support, and stressful life events provided no unique significance. The only significant interaction was between peer misbehavior and negative neighborhood environment (r2 = .01). The regression results shown in Table 3 excluded the nonsignificant variables. Our results suggest that older adolescents who have more conduct disorder and substance abuse symptoms and who live in neighborhoods with multiple problems are more likely to engage in high risk sexual behaviors. In addition, peer misbehavior is associated most strongly with adolescents' high risk sexual behaviors when adolescents live in neighborhoods with more problems. Possible reasons for this interaction effect are discussed in the final section of this paper.

Table 3. Multiple Regression Results for Factors Significantly Associated with Adolescent Sexual Risk Behaviors (N - 778) Variables

beta

Unique r2

Gender

-.24

.03

.04*** .004*

.05 .01

.038*** .005*

ns .02

.001*

.02

.01*

Age Mental Health Problems Conduct disorder symptoms Substance abuse symptoms Social Environment Peer misbehavior Negative neighborhood environment Interaction Peer misbehavior x negative neighborhood environment Total model statistics: Model SS = 69.69 F(10,767) = 27.70***

ns

Overall Adjusted R2 = .26

Note: unique r2 reported here were obtained before the interaction term was added to the model. *p < .05, **p < .01, ***p < .001.

19

Mental Health Social Environment Table 4. Multiple Regression Results for Factors Significantly Associated with Adolescent Sexual Risk Behaviors by Gender beta

Unique r2a

Age

ns ns

ns ns

Mental Health Problems Conduct disorder symptoms Substance abuse symptoms

.05 .01

.04*** .005*

ns .02

.005*

.00 .06

.01** .004*

Independent Variables Gender

Social Environment Peer misbehavior Negative neighborhood environment

ns

Interactions Peer misbehavior X negative neighborhood environment Gender x Ageb Total model statistics: Model SS = 70.50 F(8, 769) = 35.27***

Overall Adjusted R2 = .26

*p < .05, **p < .01, ***p < .001. Note a: unique r2 reported here were obtained before the interaction term was added to the model. Note b: Only the significant interaction by gender was included in the table.

Males vs. Females To examine whether there were differences between males and females in the factors associated with adolescent sexual risk behaviors, multiple regression models were estimated examining the interactions of all independent variables with respondent's gender. A stepwise approach was adopted to deal with the potential problem of multicollinearity which would be created when many interaction terms were included in a multivariate regression model. However, the stepwise approach was only used for interactions rather than main effects. All main effects were entered in the model first and estimated simultaneously. The overall model including interactions was statistically significant F(8, 769) = 35.27, p = 0.001, and explained about 27% of the variance in sexual risk behaviors, the same percentage as explained by the model without including gender interactions (see Table 4). Like the other model, when controlling for the effects of all other independent variables in the model, variables that were significantly associated with adolescents' sexual risk behaviors included conduct disorder symptoms (r2 = 0.04), substance abuse or dependence symptoms (r2 = 0.005), and negative neighborhood envi-

Yu-Wen Chen et al.

20

ronments (r2 = 0.005). Also, the interaction between peer misbehavior and negative neighborhood environments continued to be significant (r2 = 0.01). In this model, the main effect of age dropped out, but the age by gender interaction was found to be significant (r2 = 0.004). The significant positive interaction of gender by age despite the nonsignificant main effect of age suggested that age only had an effect on male adolescents' engagement in sexual risk behaviors. Older males, but not older females, were more likely to engage in sexual risk behaviors than their younger counterparts. No other gender differences were found. DISCUSSION

Our study shows that mental health problems and social environment did not differ in their association with the sexual risk behaviors of male and female adolescents: in fact, the only difference between the genders is that age is more closely associated with males' sexual risk behaviors than with females' risk behaviors. Like other studies, our study found associations between sexual risk behaviors and mental health problems (e.g., Harvey & Spinger, 1995; Simpson et al., 1993; Stiffman et al, 1992); sexual risk behaviors and social environment (e.g., Billy et al., 1994, Brown, Clasen, & Eicher, 1986;); and sexual risk behaviors and a combination of both mental health and social environment (Biglan et al., 1990; Stiffman et al., 1995). Our study expanded on other studies by also examining gender differences and interactions between various social environment factors. For both male and female adolescents, conduct disorder symptoms, substance abuse or dependence symptoms, and an interaction between peer misbehaviors and negative neighborhood environment were significantly associated with sexual risk behaviors. Although previous studies have documented differences in factors associated with mental health and behavior problems (e.g., depression, drinking problems, cigarette smoking, and illicit drug use) across gender groups (Avison & Mcalpine, 1992; Peterson, Sarigiani, & Kennedy, 1991; Robbins & Martin, 1993; Swanson, Dibble, & Trocki, 1995), and some studies also have found gender differences in HIV risk behaviors among adult heterosexuals (e.g., Bloor et al., 1992; Padian, Shiboski, & Jewell, 1991; Scheidt & Windle, 1995), our study did not find differences in the mental health and social environment variables associated with adolescent sexual risk behaviors across gender. The only gender differences were found in age, with older males more likely to report engaging in high risk sexual behaviors, which was consistent with previous studies on adolescents' use

Mental Health Social Environment

21

of contraceptives and involvement in sexual risk behaviors (DiClemente, 1992; Hingson et al., 1990). The similar significant interaction between peer misbehavior and neighborhood problems for both males and females suggests that neighborhood problems moderate the association between peer behaviors and both genders' sexual risk behaviors. Peer misbehavior alone is not related to the engagement in sexual risk behaviors. However, sexual risk behaviors are more frequent when the adolescent lives in a community with multiple problems and has misbehaving peers. Although our findings are informative, some limitations should be noted. The first limitation lies in the nonrepresentative nature of the sample in the present study. All adolescents were inner-city gateway service users in one midwestern city and most were from lower socioeconomic families; therefore, they might not be representative of the general adolescent population. However, youths receiving public services are generally at high risk of unwanted pregnancy and infections of HIV and other sexually transmitted diseases (DiClemente, 1991; Rotheram-Borus et al., 1991), and are therefore, most urgently in need of professional assistance. Research done with this population can help provide insight for interventions to this high risk population. Second, some major indices of sexual risk behaviors (such as condom use) are not included in our scale. It is expected that the number of youths engaging in sexual risk behaviors would increase with the inclusion of measures on condom use. Third, all information for our study was gathered through self-report. Thus, social desirability, privacy, embarrassment, and fear of reprisals may make youths conceal or underreport many sensitive and personal questions (e.g., sexual behaviors, conduct disorder, alcohol and drug use). Fourth, the study used a crosssectional research design. Additional studies using a longitudinal design will be necessary to evaluate the significance of the predictors in influencing adolescent sexual risk behaviors. Only a longitudinal design would be able to examine any causal relationships between mental health problems, social environment variables and sexual risk behaviors. Despite these limitations, our findings suggest several major points for program development and future research. First, engagement in sexual risk behaviors among at-risk adolescents is related to a combination of mental health and social environmental problems. Our findings underscore the role of the social environment in high-risk sexual behaviors among adolescents, and reinforce the need for greater involvement of family, peer, and community in prevention programs. In other words, we need to develop multidimensional interventions that deal with various associated adolescent problem behaviors, such as conduct disorder, sexual risk behaviors, and substance use within the context of their social environment. Second, older

Yu-Wen Chen et al.

22

male adolescents are identified as being in greater need for interventions than those who are younger. Identification and prevention strategies should be based on all of those indicators. Third, future research should work on the development of a better measure of sexual risk behaviors, that is, a measure that encompasses a broader spectrum of behaviors that are considered to be risky. Finally, we need to continue our efforts to the development of comprehensive theories (i.e., theories incorporating individual and social factors and are specific to the adolescent population) applicable to explain adolescent sexual risk behaviors. The combined use of a more comprehensive theory and a longitudinal research design will enable us to find direct and indirect causes of adolescent sexual risk behaviors. Prevention and intervention strategies based on such research findings promise to be more effective in reducing adolescent high-risk sexual behaviors than traditional approaches. ACKNOWLEDGMENTS

This research was supported by grant #R24 MH50857-02 from the National Institute for Mental Health. A earlier version of this paper was presented at the 5th Annual Virginia Beach Conference, 1995. REFERENCES Avison, W. R., & Mcalpine, D. D. (1992). Gender differences in symptoms of depression among adolescents. Journal of Health and Social Behavior, 33, 77-96. Baker, D. G., & Mossman, D. (1991). Potential HIV exposure in psychiatrically hospitalized adolescent girls. American Journal of Psychiatry, 148, 528-530. Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1992). A social cognitive approach to the exercise of control over AIDS infection. In R. J. DiClemente (Ed.). Adolescents and AIDS: A generation in jeopardy (pp. 89-116). CA: Sage Publication. Barlett, J. A., Keller, S. E., Eckholdt, H., & Schleifer, S. J. (1995). HIV-relevant issues in adolescents. In N. Boyd-Franklin, G. L. Steiner, and M. G. Boland (Eds.), Children. families. and HIV/AIDS: Psychosocial and therapeutic issue (pp. 78-89), New York: Guilford Press. Biglan, A., Metzler, C. W., Wirt, R., Ary, D., Noell, J., Ochs, L., French, C, & Hood, D. (1990). Social and behavioral factors associated with high-risk sexual behavior among adolescents. Journal of Behavioral Medicine, 13, 245-261. Billy, J. O .G., & Udry, J. R. (1985). The influence of male and female best friends on adolescent sexual risk behavior. Adolescence, 20, 21-32. Billy, J. O. G., Brewster, K. L, & Grady, W. R. (1994). Contextual effects on the sexual behavior of adolescent women. Journal of Marriage and the Family, 56, 387-404. Bloor, M. J., McKeganey, N. P., Finlay, A., & Barnard, M. A. (1992). The inappropriateness of psycho-social models of risk behavior for understanding HIV-related risk practices among Glasgow male prostitutes. AIDS Care, 4, 131-137.

Mental Health Social Environment

23

Booth, R. E., Watters, J. K., & Chitwood, D. D. (1993). HIV risk related sex behaviors among injection drug users, crack smokers and injection drug users who smoke crack. American Journal of Public Health, 83, 1144-1148. Botvin, G. J., & Dusenbury, L. (1992). Substance abuse prevention: Implications for reducing risk for HIV infection. Psychology of Addictive Behaviors, 6, 70-80. Brewster, K. L., Billy, J. O. G., & Grady, W. R. (1993). Social context and adolescent behavior: The impact of community on the transition to sexual activity. Social Forces, 71, 713-740. Bronfenbrenner, U. (1979). The ecology of human development. Cambridge, MA: Harvard University Press. Brown, B. B., Clasen, D. R., & Eicher, S. A. (1986). Perceptions of peer pressure, peer conformity dispositions, and self-reported behavior among adolescents. Developmental Psychology, 22, 521-530. Calvert, B., & Stanton, W. R. (1992). Perceptions of parenthood: similarities and differences between 15-year-old girls and boys. Adolescence, 27, 315-328. Carver, V. C, Kittleson, M. J., & Lacey, E. P. (1990). Adolescent pregnancy: A reason to examine gender differences in sexual knowledge, attitudes, and behavior. Health Values Health Behavior, Education and Promotion, 14, 24-29. Catania, J. A., Coates, T. J., Kegeles, S., Fullilove, M. T., Peterson, J., Martin, B., Siegel, D., & Hulley, S. (1992). Condom use in multi-ethnic neighborhoods of San Francisco: The population-based AMEN (AIDS in multi-ethnic neighborhoods) study. American Journal of Public Health, 82, 284-287. Cates, W., Jr., & Stone, K. M. (1992). Family planning, sexually transmitted diseases and contraceptive choice: a literature update—Part I. Family Planning Perspectives, 24, 75. Centers for Disease Control. (1993). U.S. AIDS cases reported through September 1993. HIV/AIDS Weekly Surveillance Report, 5(3), 10-11. Coleman, J. S. (1981). The Adolescent Society. New York: Free Press. DiBlasio, F. A., & Benda, B. B. (1990). Adolescent sexual behavior: Multivariate analysis of a social learning theory. Journal of Adolescent Research, 5, 449-466. DiClemente, R. J. (1991). Predictors of HIV-preventive sexual behavior in a high-risk adolescent population: The influence of perceived peer norms and sexual communication on incarcerated adolescents' consistent use of condom. Journal of Adolescent Health, 12, 385-390. DiClemente, R. J. (1992). Psychological determinants of condom use among adolescents. In R. J. DiClemente (Ed.) Adolescents and AIDS: A generation in jeopardy (pp. 34-51). CA: Sage Publication. Edelman, M. R. (1987). Families in Peril. Massachusetts: Harvard University Press. Edlin, B. R., Irwin, K. L., Ludwig, D. D., McCoy, H. V., Serrano, Y., Word, C, Bowser, B. P., Faruque, S., McCoy, C. B., Schilling, R. F., Holmberg, S. D., & the Multicenter Crack Cocaine and HIV Infection Study Team. (1992). High-risk sexual behavior among young street-recruited crack cocaine smokers in three American cities: An interim report. Journal of Psychoactive Drugs, 24, 363-371. Elliot, D. S., & Morse, B. J. (1989). Delinquency and drug use as risk factors in teenage sexual activity. Youth and Society, 21, 32-60. Ensminger, M. E. (1990). Sexual activity and problem behaviors among black, urban adolescents. Child Development, 61, 2032-2046. Flavin, D., Franklin, J., & Francis, R. (1986). AIDS and suicidal behavior in alcohol-dependent homosexual men. American Journal of Psychiatry, 143, 1440-1442 Gerber, G., & Resnick, M. D. (1988). Family functioning of adolescent who parent and place for adoptions. Adolescence, 23, 417-428. Greenblatt, M., & Robertson, M. J. (1993). Life-styles, adaptive strategies, and sexual behaviors of homeless adolescents. Hospital and Community Psychiatry, 44, 11771180. Harvey, S. M., & Spigner, C. (1995). Factors associated with sexual behavior among adolescents: A multivariate analysis. Adolescence, 30, 253-264. Hayes, C. D. (1987). Risking the future: Adolescent sexuality, pregnancy, and childbearing. Washington: National Academy Press.

24

Yu-Wen Chen et al.

Hein, K. (1993). "Getting real" about HIV in adolescents. American Journal of Public Health, 83, 492-494. Hingson, R. W., Strunin, L., Berlin, B. M, & Heeren, T. (1990). Beliefs about AIDS, use of alcohol and drugs, and unprotected sex among Massachusetts. American Journal of Public Health, 80, 295-299. Hudgins, R., McCusker, J., Stoddard, A. (1995). Cocaine use and risky injection and sexual behaviors. Drug and Alcohol Dependence, 37, 7-14. Hudson, W. W. (1982). The clinical measurement package. Homewood, IL: Dorsey Press. Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. New York: Academy Press. Kipke, M. D., Montgomery, S., & MacKenzie, R. G. (1993). Substance use among youth seen at a community-based health clinic. Journal of Adolescent Health, 4, 289-294. Ku, L., Sonenstein, F. L., & Pleck, J. H. (1993). Neighborhood, family, and work: Influences on the premarital behaviors of adolescent males. Social Forces, 72, 479-503. Leigh, B. C. (1990). The relationship of substance use during sex to high-risk sexual behavior. Journal of Sex Research, 27, 199-213. Luster, T., & Small, S. A. (1994). Factors associated with sexual risk-taking behaviors among adolescents. Journal of Marriage and the Family, 56, 622-632. Mason, W. M., Wong, G., & Entwisle, B. (1983). Contextual analysis through the multi-level linear model. In Samuel Leinhardt (Ed.), Sociological Methodology 1983-1984 (pp. 72-103). Jossey-Bass. McEwan, R. T., McCallum, A., Bhopal, R. S., & Madhok, R. (1992). Sex and the risk of HIV infection: The role of alcohol. British Journal of Addition, 87, 577-584. Mirande, A. M. (1968). Reference group theory and adolescent sexual behavior. Journal of Marriage and the Family, 30, 572-577. Nathanson, C A., & Becker, M. H. (1986). Family and peer influence on obtaining a method of contraception. Journal of Marriage and the Family, 48, 513-525. National Center for Health Statistics. (1994) Healthy People 2000 Review, 1993, Hyattsville, Maryland: Public Health Service. Oskamp, S., & Mindick, B. (1983). Personality and attitudinal barriers to contraception. In D. Byrne & W. A. Fisher (Eds.), Adolescents, sex, and contraception, (pp. 65-107). Hillsdale, New Jersey: Lawrence Erlbaum Associates, Pub. Ostrow, D. G., Monjan, A., Joseph, J., VanRaden, M., Fox, R., Kingsley, L., Dudley, J., & Phair, J. (1989). HIV-related symptoms and psychological functioning in a cohort of homosexual men. American Journal of Psychiatry, 146, 737-742. Padian, N. S., Shiboski, S. C., & Jewell, N. P. (1991). Female-to-male transmission of human immunodeficiency virus. Journal of American Medical Association, 266, 1664-1667. Peterson, A. C, Sarigiani, P. A., & Kennedy, R. E. (1991). Adolescent depression: Why more girls? Journal of Youth and Adolescence, 20, 247-271. Robbins, C. A., & Martin, S. S. (1993). Gender, styles of deviance, and drinking problems. Journal of Health and Social Behavior, 34, 302-321. Rosenbaum, E., & Kandel, D. B. (1990). Early onset of adolescent sexual behavior and drug involvement. Journal of Marriage and the Family, 52, 783-798. Rotheram-Borus, M. J., Koopman, C, Haignere, C., & Davies, M. (1991). Reducing HIV sexual risk behaviors among runaway adolescents, Journal of American Medical Association, 266, 1237-1241. Scheidt, D. M., & Windle, M. (1995). The alcoholics in treatment HIV risk (ATRISK) study: Gender, ethnic and geographic group comparisons. Journal of Studies on Alcohol, 56, 300-308. Shaffer, D., Schwab-Stone, M., Fisher, P., & Cohen, et al., (1993). The Diagnostic Interview Schedule for Children-Revised Version (DISC-R): Preparation, field testing, interrater reliability, and acceptability. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 643-650. Simpson, D. D., Knight, K., & Ray, S. (1993). Psychosocial correlates of AIDS-risk drug use and sexual behaviors. AIDS Education and Prevention, 5, 121-130.

Mental Health Social Environment

25

Stiffman, R. A. (1989). Physical and sexual abuse in runaway youths. Child Abuse and Neglect, 13, 417-426. Stiffman, A. R., Cunningham, R., Earls, R, & Dore, P. (1991). Change in AIDS risk behaviors from adolescence to adulthood. In Giraldo G. (Ed.), Science challenging AIDS: Conference proceedings, Seventh International Conference on AIDS. Basel, Switzerland: Karger. Stiffman, A. R., Dore, P., Cunningham, R., & Earls, F. (1995). Person and environment in HIV risk behavior change between adolescence and young adulthood. Health Education Quarterly, 22, 211-226. Stiffman, A. R., Dore, P., Cunningham, R., Earls, F., Farber, S. (1996). Adolescent violence. In DiClemente R., Hansen W., & Ponton L. (Eds.), Handbook of adolescent risk behavior. (pp. 289-312) New York: Plenum. Stiffman, A., R., Dore, P., Earls, F., & Cunningham, R. (1992). The influence of mental health problems on AIDS-related risk behaviors in young adults. Journal of Nervous and Mental Disease, 180, 314-320. Stiffman, A. R. & Earls, F. (1990). Behavioral risks for HIV infection in adolescent medical patients. Pediatrics, 85, 303-310. Stiffman, A. R., Earls, F., Robins, L. N., Jung, K. G., & Kulbok, P. (1987). Adolescent sexual activity and pregnancy: Socioenvironmental problems, physical health, and mental health. Journal of Youth and Adolescence, 16, 497-509. Swanson, J. M., Dibble, S. L., & Trocki, K. (1995). A description of the gender differences in risk behaviors in youth adults with genital herpes. Public Health Nursing, 2(2), 99-108. Weddle, K. D., McKenry, P. C, & Leigh, G. K. (1988). Adolescent sexual behavior: Trends and issues in research. Journal of Adolescent Research, 3, 245-257. Welner, Z., Reich, W., Herjanic, B., Amado, H., & Jung, K. (1987). Reliability and validity, and parent-child agreement studies of the diagnostic interview for children and

adolescents (DICA). Journal of the American Academy of Child and Adolescent Psychiatry, 26, 649-953. Yates, G. L,, Mackenzie, R. G., Pennbridge, J., & Swofford, A. (1991). A risk profile comparison of homeless youth involved in prostitution and homeless youth not involved.

Journal of Adolescent Health, 12, 545-548.