The AIDS Risk Reduction Model: Examining Intentions to Use Condoms Among Injection Drug Users1


  • 1

    This work was supported by NIDA Grants DA05589 and DA00146, and Training Grant DA07272.

  • Special thanks to Kevin D. Henson and to the anonymous reviewer for comments on earlier drafts of this paper, and to Shih-chao Hsieh for technical assistance.

Correspondence concerning this article should be addressed to Mark R. Kowalewski, UCLA Drug Abuse Research Center, Neuropsychiatric Institute, 1100 Glendon Avenue, #763, Los Angeles, CA 90024.


This paper uses the AIDS Risk Reduction Model (ARRM) to examine psychosocial factors involved in adopting safer sex practices in a sample of Los Angeles injection drug users (IDUs; n= 161) who reported having more than one sex partner in the year preceding the interview. The ARRM hypothesizes that behavior change is a process occurring in three stages: (a) labeling one's behavior as problematic, (b) making a commitment to behavior change, and (c) taking action to accomplish that change. We test the first two stages of the model using a measure of perceived risk of HIV infection (Stage I), and intentions to use condoms always during vaginal or anal sex in the next year (Stage 2). We examine differences in the predictive value of the ARM between IDUs who reported using condoms in the year prior to the interview and those who reported not using them. We identify leverage points in the model-factors which appear to have a major influence on intentions to use condoms and which may be amenable to change through educational or other types of intervention. For both condom users and non-users, susceptibility to AIDS predicted perceived infection risk (Stage I). For condom users, knowledge about AIDS also predicted perceived risk. For both groups, self efficacy, peer norms concerning condom use, and the perceived pleasure of using condoms predicted intentions to use condoms (Stage 2). Our findings do not support either direct or indirect relationships between the Stage 1 and Stage 2 outcome variables for either group.