SIMILARITIES IN SEXUAL ACTIVITY AND CONDOM USE AMONG FRIENDS WITHIN GROUPS BEFORE AND AFTER A RISK-REDUCTION INTERVENTION.(Statistical Data Included)

Continued expansion of the AIDS epidemic reinforces the need to develop effective risk-reduction interventions, particularly those targeting potentially high-risk populations. Minority adolescents constitute one of these demographically important high-risk populations (Centers for Disease Control and Prevention, 1993). With increasing evidence that interventions can result in some reduction of adolescent risk behavior, an especially compelling challenge is the identification of mechanisms to accentuate and prolong this effect (Kim, Stanton, Li, Dickerson, & Galbraith, 1997; Kirby et al., 1994; St. Lawrence et al., 1995).

Inherent in the normal social development of early and middle adolescents is the separation from family and the identification with the peer social group (Felice & Friedman, 1982). Adolescents manifest that identification by similarities in dress and physical appearance as well as similarities in behaviors (Petersen, Leffert, & Graham, 1995). For many years, researchers and other professionals involved with adolescents have noted the similarity of risk behaviors among friends (Brooks-Gunn & Furstenberg, 1989; Cohen, 1977; Dolcini & Adler, 1994; Hollingshead, 1949; Newcomb, Huba, & Bentler, 1986). In particular, the most proximate measure of adolescent risk behavior is the perceived or actual behavior of one's close friends (Jessor & Jessor, 1977). Pressures to conform to group behaviors can be found across ages and grade levels, although peer pressures toward misconduct increase as grade level increases (Clasen & Brown, 1985). Group social pressures can lead to increasing prosocial/protective behaviors or antisocial/risk behaviors. The salience of peer groups, and therefore the influence on adolescents, varies with age (Brown, Eicher,& Petrie, 1986).

One promising avenue for modifying risk behaviors and intensifying intervention effect is the use of peer influence to increase the levels of protective behaviors. This observation has contributed to the widespread practice of delivering risk-reduction interventions to groups of youth and to the use of peer educators (Klepp, Halper, & Perry, 1986; Perry, Klepp, Halper, Hawkins, & Murray, 1986). Many risk-reduction interventions have been school-based so as to deliver interventions to a captured audience of same-aged youth. Taken one step further, intervention delivery to a youth along with his or her actual friends (as opposed to an investigator-formed peer group consisting of individuals who do not have any particular social value or salience to the youth) could be a powerful vehicle for intervention delivery. Such an intervention approach, delivered in community settings, might be expected to activate an iterative reinforcement of positive behaviors and perceptions between any individual within the group and the remainder of the group (see Figure 1).

[Figure 1 ILLUSTRATION OMITTED]

However, as provocative as these observations and corollary hypotheses are, much remains unknown regarding this phenomenon of similarity of sexual risk behaviors among friends. First, how similar are sexual risk and protective behaviors and perceptions among friends and how persistent over time is the similarity? Second, what effect would the delivery of an HIV risk-reduction intervention to groups of friends have on the similarity of risk and protective behaviors among youth within these groups? A related question is whether intervention delivery to the group would have a variable impact on group behavior depending on the baseline prevalence of risk behavior within the group. We hypothesize that for groups of youth with a low prevalence of risk behavior, a risk-reduction intervention might be reinforcing and thus, after the intervention, prevalence of risk behavior among group members would remain low. For groups of youth with a high prevalence of risk behavior, the intervention, which would be in opposition to group effect, might not be effective. If it were effective, one might anticipate that it would lower risk behavior among only some of the group members such that most groups would remain high risk, although a few might convert to low risk (see Figure 2).

[Figure 2 ILLUSTRATION OMITTED]

To address these questions and hypotheses, we analyzed data collected from 76 groups of adolescent friends enrolled in a longitudinal HIV risk-reduction intervention.

MATERIALS AND METHODS

OVERVIEW

A recent randomized controlled evaluation of an HIV risk-reduction intervention targeting low-income, African American urban adolescents found the intervention to be effective in increasing condom use among sexually active youth 6 months post-intervention (e.g., condom-use rates were 21% higher among intervention compared to control youth, p [is less than] .05). Intervention impact was waning by 12 months post-intervention (e.g., rates were only 9% higher among intervention youth), but condom use was again 21% higher among intervention youth at 18 months (p [is less than] .05) (Stanton, Kim, Galbraith, Feigelman, & Kaljee, 1996; Stanton, Li, Ricardo, et al., 1996). At the baseline, 36% of the youth had ever had sex (Stanton, Li, Ricardo, et al., 1996). The intervention was not effective in delaying the onset of sexual activity at subsequent assessments. As the intervention was delivered through 76 naturally occurring groups of friends (vide infra), the data from this study provide a unique opportunity to examine the similarity of risk behaviors and perceptions among friends and their persistence over time with and without a risk-reduction intervention.

STUDY PARTICIPANTS

Participants were African American youth, 9 through 15 years of age at the time of recruitment, identified from nine recreation centers associated with three public housing developments in a large Eastern city. Research staff conducted three introductory sessions at each recreation center to describe the purpose, general design, and enrollment criteria to potentially eligible youth. It was explained that, because the intervention was to be delivered to naturally formed friendship groups (vide infra), only groups of friends (rather than single individuals) could be enrolled. Potentially interested youth were instructed to approach their friends (defined as other youth with whom …