THE ASSOCIATION BETWEEN METHAMPHETAMINE use and sexual risk behaviors among men who have sex with men (MSM) has been well documented.1–7 These studies and others have shown increased rates of unprotected anal intercourse and multiple sexual partners among MSM during prolonged and potentially less-inhibited sexual encounters involving methamphetamine.
Less well-documented, however, are the context and effects of methamphetamine use during heterosexual encounters. One study of a predominantly heterosexual sample of methamphetamine users in Australia concluded that use of methamphetamine did not place these users at heightened sexual risk.8 Although other studies have found heterosexual methamphetamine users to be at higher risk for contracting HIV through sex than users of other drugs,9,10 the majority of these studies11–19 did not analyze event-level data and thus they were unable to establish that the use of methamphetamine and sexual risk behaviors occurred within the same sexual encounter.
Although event-level analyses of substance abuse and attendant sexual risk have become more common in the literature addressing both homosexual and heterosexual behavior, to date there have been relatively few event-level studies that have specifically examined the co-occurrence of methamphetamine use and sexual practices.20 Such studies are particularly rare among heterosexuals. In addition, the event-level studies of substance use and sexual risk that have been done typically have focused their analyses exclusively on the occurrence of 1 type of sexual risk.
In an effort to expand the current understanding of the context of methamphetamine use among heterosexuals, we analyzed event-level data from 1213 heterosexual encounters reported by 703 out-of-treatment injecting drug users. We were interested in more fully examining event-level associations between methamphetamine use and a range of sexual behaviors in heterosexual encounters.
Materials and Methods
A total of 855 out-of-treatment injecting drug users were enrolled in the study between July 2003 and January 2006 in the Raleigh-Durham metropolitan area (North Carolina) using a targeted sampling approach that has been used in other cities.21,22 This approach included street outreach methods, where recovering drug users went into high drug use communities to recruit active drug users and distribute risk reduction materials (e.g., condoms, bleach, water, and educational materials).23 After preliminary screening in the field, prospective participants were referred to a project office where the study was described to them and they provided informed consent. To minimize underreporting of sensitive behaviors, data collection was performed using Audio Computer Administered Self Interviews. Upon completion of the initial interview, participants were randomized to either a 6-session motivational intervention or a 6-session educational intervention designed to modify behaviors which may increase the probability of infection, transmission, or progression of HIV, hepatitis C virus (HCV), or hepatitis B virus (HBV). Following the first intervention session, participants were offered counseling and testing for antibodies to HIV, HCV, and HBV. Baseline data were collected across 2 visits, which were completed about 1 week apart. Participants received their test results (HIV, HCV, and HBV) and posttest counseling following completion of data collection at the second visit. Follow-up interviews were scheduled for 6 and 12 months after enrollment to provide data for evaluating intervention effects. This report is limited to 1213 heterosexual encounters reported by 703 participants. Encounters involving women having sex with women or men having sex with men were excluded.
Eligibility criteria for the study included a minimum age of 18 years; self-reported injecting drug use in the previous 30 days; visible tracks (injection marks) and/or a urine specimen positive for heroin (morphine), cocaine, or methamphetamine; no formal substance abuse treatment in the previous 30 days; and current residence in 1 of the 2 counties in which the study was conducted. This study was approved by RTI International’s Office for Research Protection.
The interviews included sections on sociodemographics, alcohol and other drug use, injecting practices, substance abuse treatment, sexual behavior, and health. A series of questions were also asked about the last time the survey respondent had sex. This section was introduced by the statement, “the next questions are about the last time you had sex.” We are using the term sexual encounter to refer to all of the behaviors reported as having occurred during the last time a person had sex. We chose the term “encounter” rather than “event” because many participants reported engaging in more than 1 type of sex (e.g., oral, vaginal, and/or anal intercourse) the last time that they “had sex.” Questions that were asked about the encounter included the partner’s demographic characteristics (gender, race, and estimated age), number of times the participant had sex with the partner previously, and participant’s relationship to the partner. Additional questions included use of specific drugs by the study participant and by the partner, types of sex (e.g., oral, vaginal, anal), and whether a condom was used during each type of sex. Methamphetamine use during the encounter was assessed with the questions; “Did you use speed or crank before or during the last time you had sex?” and “Did your partner use speed or crank before or during the last time you had sex?” Use of other drugs during the encounter was assessed with similar questions.
We combined types of sex and use of protection to come up with the following 6 different risky or unsafe sexual behaviors: unprotected vaginal intercourse, anal intercourse, unprotected anal intercourse, vaginal and anal intercourse during the same encounter, sex with a new partner (i.e., someone the participant had sex with for the first time) and unprotected intercourse with a new partner. To adjust for within-person correlations, a generalized estimating equations (GEE) approach was used to assess associations between methamphetamine use and each of the 6 sexual risk behaviors.24 Separate analyses were performed to examine the effects of any methamphetamine use during an encounter, and subsequently to examine methamphetamine use by just 1 partner and by both partners during the encounter. Multivariate logistic regression models were developed which adjusted for age of each partner, race of the male partner, and cocaine (crack or powder—alone or in combination with heroin [speedball]) use during the encounter. We included this variable because in separate analyses (not shown) powder cocaine, crack, and speedball were each associated with some of the outcomes. Type of partner (main or casual) was also controlled for in all models except those examining sex with a new partner or unprotected intercourse with a new partner. Since both partners were of the same racial or ethnic group in 80% of encounters, only race of the male partner was entered as a control variable. Analyses were performed using the GEE procedure in PROC GENMOD SAS 9.1.
Participants reported on their own and their partners’ substance use and the couples’ sexual behaviors during as many as 3 separate sexual encounters at different points in time. These include 633 encounters reported by participants at the baseline interviews, 344 encounters reported at the 6-month follow-up interview, and 236 encounters reported at the 12-month follow-up interview. Forty-seven percent of participants only provided a report of 1 heterosexual encounter, 33% reported 2 encounters, and 20% reported 3 encounters.
Of the 703 participants analyzed for this report, the majority were male (73%), black (62%), and unemployed (71%). The mean age of participants was 41 years. Twenty percent were married or living as married and 49% had completed high school. At the time of the intake interview, 35% considered themselves to be homeless, 65% reported a history of substance abuse treatment, and 57% reported having been in prison. Seven percent tested positive for HIV and 54% tested positive for HCV. Crack was the most common drug participants reported having used in the last 30 days (72%), followed by alcohol (71%), heroin (69%), powder cocaine (63%), and marijuana (62%). Eleven percent of the sample reported using methamphetamine in the last 30 days. Twenty-seven percent of participants reported having more than 1 sexual partner in the previous 30 days, 39% reported engaging in unprotected vaginal intercourse, and 9% reported engaging in anal intercourse during that period.
Participants who reported using methamphetamine in the previous 30 days at baseline differed significantly from those who did report using it in several ways. Methamphetamine users were more likely to be non-Hispanic white and less likely to be black than other participants. They were also significantly more likely to be homeless, less likely to have completed high school, more likely to be unemployed, and their income was lower. They were also more likely to smoke marijuana, use crack cocaine, powder cocaine, and heroin. Participants who reported using methamphetamine were also significantly more likely than other participants to report having more than 1 sexual partner, having an IDU sexual partner, and engaging in anal intercourse in the last 30 days. These comparisons are shown in Table 1
Characteristics of Encounters
The distribution of age and race among the sexual partners was similar to the distribution among the study participants. The male partners in this study sample were on average 5 years older than the female partners (41 years old vs. 36 years old) in this sample. In 64% of encounters the male partner was black, in 28% the male was non-Hispanic white, and in 8% the male was of other race/ethnicity (e.g., Hispanic, Native American, Asian). In 62% of encounters the female partner was black, in 33% the female was non-Hispanic white, and in 5% the female was of other race/ethnicity. In 80% of encounters both partners were of the same race/ethnicity.
Forty-nine percent of encounters involved only 1 type of sex, 45% involved 2 types, and 6% involved 3 types. Most encounters involved vaginal intercourse (89%)—alone (39%), combined with oral sex (42%), combined with oral and anal sex (6%), or combined with anal sex (2%). Nine percent (n = 111) of encounters involved anal intercourse. Of these, 61% (68 of 111) also included vaginal and oral sex, 18% (20 of 111) included vaginal but not oral, 10% (11 of 111) included oral but not vaginal sex, and 11% (12 of 111) involved anal intercourse by itself. Condoms were not used in more than half of the encounters involving vaginal intercourse (54%) and more than half of those involving anal intercourse (52%). Of the 88 encounters that involved vaginal and anal intercourse, a condom was used for both acts in 39% (34 of 88), for vaginal intercourse but not anal intercourse in 6% (5 of 88), and for anal intercourse but not vaginal intercourse in 7% (6 of 88). In 49% (43 of 88) of these encounters, no condom was used for either act. Seventeen percent (n = 210) of encounters involved sex with a new partner, and 34% of these encounters (n = 72) were unprotected. Frequencies of each type of sex with and without a condom are shown in Table 2.
At least 1 partner used alcohol or other drugs in 82% of encounters, and both partners used at least 1 substance in 54% of encounters. A majority of encounters involved combinations of substances; the male partner used more than 1 substance in 56% of encounters and the female partner used more than 1 substance in 47% of encounters. Methamphetamine was used in 7% (n = 80) of encounters. Some form of cocaine (crack, powder—alone or in combination with heroin [speedball]) was used by at least 1 partner in 59% of encounters. As noted above, most encounters involved polydrug use. Frequencies for use of each drug by either partner during the encounter are shown in Table 2.
Methamphetamine and Heterosexual Practices
Any Methamphetamine Use.
Bivariate and multivariate logistic regression models for the association between engaging in 6 different sexual practices and the use of methamphetamine by either or both partners during the encounter are shown in Table 3. In bivariate GEE analyses, encounters in which either or both partners used methamphetamine were significantly more likely to include anal intercourse (odds ratio [OR] = 3.54, 95% confidence interval [CI] = 2.03–6.15), unprotected anal intercourse (OR = 2.91, 95% CI = 1.40–6.06), vaginal and anal intercourse (OR = 3.86, 95% CI = 2.16–6.89), and sex with a new partner (OR = 2.26, 95% CI = 1.30–3.93) but significantly less likely to include unprotected vaginal intercourse (OR = 0.56, 95% CI = 0.35–0.91). After adjusting for partner type (main or casual), age of each partner, race of the male partner, and use of cocaine (crack, powder, or mixed with heroin as speedball), methamphetamine use by either or both partners during a sexual encounter was associated with 3 of the 6 sexual behaviors. Specifically, encounters in which methamphetamine was used by either or both partners were more than twice as likely to involve anal intercourse (OR = 2.41, 95% CI = 1.29–4.53), and vaginal and anal intercourse (OR = 2.41, 95% CI = 1.22–4.77), and close to twice as likely to involve sex with a new partner (OR = 1.98, 95% CI = 1.09–3.61).
Methamphetamine Use by Partner and Sexual Risk Behaviors.
We also examined participation in different types of sex acts by which partner used methamphetamine (i.e., no one used, male used, female used, both partners used) during the encounter (Fig. 1). Both partners used methamphetamine in 28 encounters, just the male partner used it in 21 encounters, and just the female partner used it in 31 encounters. In general, encounters in which methamphetamine was used were more likely to involve anal intercourse and sex with a new partner than encounters in which it was not used. In encounters in which just 1 partner used methamphetamine, the gender of the partner using it appeared to have limited impact on the sexual behaviors we examined. Since encounters in which both partners used methamphetamine seemed to be substantially more likely than other encounters to involve anal intercourse and sex with a new partner, we developed separate logistic regression models to examine the effects of methamphetamine on sexual practices when just 1 partner used it and when both partners used it.
As shown in Table 4, in the bivariate GEE analyses, in encounters in which only 1 partner used methamphetamine, methamphetamine use was significantly associated only with an increased odds of engaging in vaginal and anal intercourse (OR = 2.45, 95% CI = 1.13–5.31). In comparison, in the bivariate GEE analyses, encounters in which both partners used methamphetamine were significantly more likely to include 5 of the 6 sexual risk behaviors, the exception being unprotected vaginal intercourse.
In models that adjusted for partner type (main or casual), age of each partner, race of the male partner, and use of cocaine (crack, powder, or mixed with heroin as speedball), methamphetamine use by only 1 partner was not significantly associated with any of the 6 sexual practices whereas methamphetamine use by both partners was significantly associated with 5 of the 6 sexual practices, the exception again being unprotected vaginal intercourse. Specifically, encounters in which methamphetamine was used by both partners were more than 6 times as likely to involve anal intercourse (OR = 6.88, 95% CI = 3.03–15.64), more than 5 times as likely to involve vaginal and anal intercourse (OR = 5.95, 95% CI = 2.53–14.03) and unprotected intercourse with a new partner (OR = 5.20, 95% CI = 2.09–12.93), and more than 4 times as likely to involve unprotected anal intercourse (OR = 4.63, 95% CI = 1.69–12.70) and sex with a new partner (OR = 4.73, 95% CI = 2.22–10.09).
The findings from this analysis indicate that among a sample of polydrug users, heterosexual encounters in which methamphetamine is used are more likely than other encounters to involve sexual behaviors which may place individuals at increased risk for HIV and other sexually transmitted infections (STIs). Specifically, we looked at the association between methamphetamine and the following 6 different sexual behaviors: unprotected vaginal intercourse, anal intercourse, unprotected anal intercourse, vaginal and anal intercourse during the same encounter, sex with a new partner, and unprotected intercourse with a new partner. In bivariate analyses, we found the use of methamphetamine by either or both partners increased the odds of engaging in each of these behaviors with the exception of unprotected vaginal intercourse. Similarly, in multivariate models which controlled for participant demographics and other drug use, methamphetamine use by either or both partners was associated with increased odds of engaging in anal intercourse, vaginal and anal intercourse, and sex with a new partner. Finally, in multivariate analyses modeling behaviors when both partners used methamphetamine either before or during the encounter, methamphetamine use was associated with increased odds of all of these behaviors with the exception of unprotected vaginal and unprotected anal intercourse. The lack of association between methamphetamine use and unprotected vaginal intercourse is not surprising given that methamphetamine use was associated with sex with a new partner, which tends to increase condom use.
These analyses add to the current understanding of methamphetamine use and sexual risk among heterosexuals in several significant ways. For one, the use of encounter-level data allows us to confirm a temporal relationship between methamphetamine use and sexual risk among heterosexual couples; a relationship that had not been firmly established in many previous analyses. In addition, the ability to look at multiple types of sexual risk and the co-occurrence of these sexual risk behaviors expands previous analyses which have commonly just focused on 1 type of sexual risk behavior. For instance, while studies have documented an association between methamphetamine use and heterosexual anal intercourse,9,19,25 few have documented the co-occurrence of vaginal and anal intercourse in the same encounter. In this population, most encounters involving anal intercourse also involved vaginal intercourse (79%). Receptive anal intercourse is known to increase the risk of HIV acquisition and this risk is likely to increase when vaginal and anal sex occur during the same encounter inasmuch as there is a greater likelihood of condom failure or improper use of condoms during these encounters. In the case that anal intercourse precedes vaginal intercourse and the penis or condom is not cleaned adequately between acts, the co-occurrence of vaginal and anal intercourse also may increase the risk of vaginal infections through the disruption of vaginal flora, possibly increasing susceptibility to HIV.26–28
The association between methamphetamine use during an encounter and having sex with someone for the first time, particularly without condoms, is another type of sexual risk behavior that has rarely been explored in relation to methamphetamine use. Our data are insufficient to allow us to characterize the possible causal link between methamphetamine use and sex with a new partner; although studies of methamphetamine use among MSM and heterosexuals suggest that methamphetamine use facilitates sexual partnership formation by lowering inhibitions.4,7,12,18
Finally, most research on the effects of methamphetamine on sexual behavior (e.g., duration, vigorousness, types of sex) has not reported whether 1 or both partners used it. Our findings suggest that methamphetamine use by only 1 partner in a heterosexual couple does not significantly increase sexual risk-taking. Rather, we find that heterosexual risk-taking in a sample of polydrug users is substantially elevated in encounters where both partners are using methamphetamine compared with using any other combination of drugs or no drugs. This finding is not surprising when one considers that extended duration encounters (e.g., >12 hours) such as those described by Kall29 are unlikely to occur if only 1 person were using methamphetamine. It is also possible that encounters where both partners use methamphetamine may involve greater risk as they may be the result of some type of formal or informal exchange such as those described by Kall and Nilsonne30 who noted that when a man invites a woman to inject his amphetamine, although not explicit, it is understood that he is inviting her to have sex as well.
Several limitations must be considered when interpreting findings from this study. First, we cannot rule out the possibility that the associations between methamphetamine use and the sexual behaviors we examined are the result of differences between users of methamphetamine and users of other drugs rather the result of methamphetamine use during the encounter. However, the strongest associations were in encounters where both partners used methamphetamine, suggesting that methamphetamine use itself is important. Second, although we asked participants whether they used each drug before or during the last time they had sex, we did not specify exact limits on how long before the encounter a drug could have been used. However, this was intentional, since there is so much variation in the duration of effects of different drugs (e.g., the effects of crack cocaine may last 15 minutes while the effects of methamphetamine may last 8 hours or more) and in rapidity of onset depending on mode of administration (e.g., it may take more than 30 minutes for the effects of methamphetamine ingested orally to be felt, while the effects of injected methamphetamine are felt almost instantaneously). Third, although targeted sampling was used to ensure that the sample was demographically and geographically diverse, the representativeness of the sample is unknown and therefore caution should be used in generalizing the results to other samples of drug users. In addition, since the targeted sample was recruited through street outreach, it seems likely that in this sample, unemployed, homeless, and marginally-housed drug users may be overrepresented, relative to employed middle class drug users. Another concern with these data is that, as with most studies of sexual behavior, they rely on self-reports, which may be inaccurate due to faulty memory or intentional misrepresentation. Intentional misrepresentation may have been minimized by the Audio Computer Administered Self Interviews data collection procedures, which have been shown to reduce the reporting of socially desirable responses and increase reporting of potentially embarrassing sexual behaviors.31 Caution should also be used in interpreting these findings because of the relatively low power to detect statistically significant associations between methamphetamine use and some risk behaviors due to the small numbers of encounters that involved those behaviors. Specifically, unprotected anal intercourse was reported in only 58 encounters and there were only 28 encounters in which both partners reported using methamphetamine.
In conclusion, in this sample of polydrug users, methamphetamine use during heterosexual encounters appeared to be associated with potentially higher risk behaviors and greater variation in sexual practices. These effects appear to be magnified when both partners are using methamphetamine. Additional research is needed to confirm and elucidate the effects on sexual risk when both heterosexual partners are using methamphetamine. The findings from this analysis coupled with additional information on the biologic and contextual reasons why methamphetamine use contributes to potentially higher-risk heterosexual practices will be critical to developing appropriate interventions to minimize the spread of HIV and other STIs among heterosexual methamphetamine users. In particular, further event-level analyses capturing information on the full repertoire of sexual and drug use behaviors occurring in any one encounter will be crucial to designing more effective HIV and STI interventions which address the possibility of multiple and varied sex acts—thus, multiple and varied infection risks—during any single encounter. In the meantime, health care providers, HIV counselors, and other professionals working with people that use methamphetamine should begin work on developing strategies and messages (e.g., using a new condom for each act) for reducing risks associated with sex while using methamphetamine.
1. Paul JP, Stall R, Davis F. Sexual risk for HIV transmission among gay/bisexual men in substance-abuse treatment. AIDS Educ Prev 1993; 5:11–24.
2. Frosch D, Shoptaw S, Huber A, et al. Sexual HIV risk among gay and bisexual male methamphetamine abusers. J Subst Abuse Treat 1996; 13:483–486.
3. Gorman EM, Barr BD, Hansen A, et al. Speed, sex, gay men, and HIV: Ecological and community perspectives. Med Anthropol Q 1997; 11:505–515.
4. Reback, CJ. The social construction of a gay drug: Methamphetamine use among gay and bisexual males in Los Angeles. City of Los Angeles, LA: AIDS Coordinator. 1997.
5. Reback CJ, Larkins S, Shoptaw S. Changes in the meaning of sexual risk behaviors among gay and bisexual male methamphetamine abusers before and after drug treatment. AIDS Behav 2004; 8:87–98.
6. Lambert E, Normand J, Stall R, et al. Introduction: New dynamics of HIV risk among drug-using men who have sex with men. J Urban Health 2005; 82:i1–i8.
7. Halkitis PN, Shrem MT. Psychological differences between binge and chronic methamphetamine using gay and bisexual men. Addict Behav 2006; 31:549–552.
8. Hando J, Hall W. HIV risk-taking behaviour among amphetamine users in Sydney, Australia. Addiction 1994; 89:79–85.
9. Molitor F, Truax SR, Ruiz JD, et al. Association of methamphetamine use during sex with risky sexual behaviors and HIV infection among non-injection drug users. West J Med 1998; 168:93–97.
10. Zule WA, Desmond DP. An ethnographic comparison of HIV risk behaviors among heroin and methamphetamine injectors. Am J Drug Alcohol Abuse 1999; 25:1–23.
11. Gibson DR, Leamon MH, Flynn N. Epidemiology and public health consequences of methamphetamine use in California’s Central Valley. J Psychoactive Drugs 2002; 34:313–319.
12. Rawson RA, Washton A, Domier CP, et al. Drugs and sexual effects: Role of drug type and gender. J Subst Abuse Treat 2002; 22:103–108.
13. Wohl AR, Johnson DF, Lu S, et al. HIV risk behaviors among African American men in Los Angeles County who self-identify as heterosexual. J Acquir Immune Defic Syndr 2002; 31:354–360.
14. Somlai AM, Kelly JA, McAuliffe TL, et al. Predictors of HIV sexual risk behaviors in a community sample of injection drug-using men and women. AIDS Behav 2003; 7:383–393.
15. Semple SJ, Grant I, Patterson TL. Female methamphetamine users: Social characteristics and sexual risk behavior. Women Health 2004; 40:35–50.
16. Semple SJ, Patterson TL, Grant I. The context of sexual risk behavior among heterosexual methamphetamine users. Addict Behav 2004; 29:807–810.
17. Semple SJ, Patterson TL, Grant I. Determinants of condom use stage of change among heterosexually-identified methamphetamine users. AIDS Behav 2004; 8:391–400.
18. Brown AH, Domier CP, Rawson RA. Stimulants, sex, and gender. Sexual Addiction & Compulsivity 2005; 2:169–180.
19. Centers for Disease Control and Prevention (CDC). Methamphetamine use and HIV risk behaviors among heterosexual men: Preliminary results from five northern California counties, December 2001-November 2003. MMWR Morb Mortal Wkly Rep 2006; 55:273–277.
20. Mansergh G, Shouse RL, Marks G, et al. Methamphetamine and sildenafil (Viagra) use are linked to unprotected receptive and insertive anal sex, respectively, in a sample of men who have sex with men. Sex Transm Infect 2006; 82:131–134.
21. Watters JK, Biernacki P. Targeted sampling: Options for the study of hidden populations. Soc Probl 1989; 36:416–430.
22. Carlson RG, Wang J, Siegal HA, et al. An ethnographic approach to targeted sampling: Problems and solutions in AIDS prevention research among injection drug and crack-cocaine users. Hum Organ 1994; 53:279–286.
23. Cunningham-Williams RM, Cottler LB, Compton WM, et al. Reaching and enrolling drug users for HIV prevention: A multi-site analysis. Drug Alcohol Depend 1999; 54:1–10.
24. Diggle PJ, Heagerty P, Ling K-Y, et al. Analysis of Longitudinal Data. 2nd ed. New York: Oxford, 2002.
25. Molitor F, Ruiz JD, Flynn N, et al. Methamphetamine use and sexual and injection risk behaviors among out-of-treatment injection drug users. Am J Drug Alcohol Abuse 1999; 25:475–493.
26. Tchoudomirova K, Mardh PA, Kallings I, et al. History, clinical findings, sexual behavior and hygiene habits in women with and without recurrent episodes of urinary symptoms. Acta Obstet Gynecol Scand 1998; 77:654–659.
27. Taha TE, Gray RH, Kumwenda NI, et al. HIV infection and disturbances of vaginal flora during pregnancy. J Acquir Immune Defic Syndr Hum Retrovirol 1999; 20:52–59.
28. Bukusi EA, Cohen CR, Meier AS, et al. Bacterial vaginosis: Risk factors among Kenyan women and their male partners. Sex Transm Dis 2006; 33:361–367.
29. Kall KI. Effects of amphetamine on sexual behavior of male IV drug users in Stockholm: A pilot study. AIDS Educ Prev 1992; 4:6–17.
30. Kall KI, Nilsonne A. Preference for sex on amphetamine: A marker for HIV risk behaviour among male intravenous amphetamine users in Stockholm. AIDS Care 1995; 7:171–188.
31. Macalino GE, Celentano DD, Latkin C, et al. Risk behaviors by audio computer-assisted self-interviews among HIV-seropositive and HIV-seronegative injection drug users. AIDS Educ Prev 2002; 14:367–378.