High rates of HIV seroprevalence have been found among Puerto Rican drug users since early in the epidemic. Seroprevalence rates among injection drug users (IDUs) in Puer-to Rico (PR) and in New York City (NY) were estimated to be 50% by the late 1980s and early 1990s. 1,2 Seroprevalence among crack users in both locations was estimated at 15% in the early 1990s. 3,4
Declines in HIV prevalence among IDUs have been reported in both locations during the 1990s. HIV seroprevalence rates among IDUs in NY declined to approximately 25% by 1996, 5 and recent data indicate that these rates have further declined to below 20%. 6 The decline in NY has been reported for Hispanic IDUs (primarily Puerto Rican) as well as other ethnic groups. 7 Seroprevalence rates among IDUs in PR have also declined, to approximately 20% by the end of the 1990s. 8
Although trends in seroprevalence have shown encouraging declines, incidence is a more useful indicator of prevention and public health service needs. HIV incidence among IDUs in all NY ethnic groups declined to less than 1/100 person-years at risk (pyr) by the late 1990s. 9 No published studies of seroincidence among IDUs in PR are available. Holmberg 10 estimated the size and direction of the HIV epidemic, based on data collected primarily in the early 1990s, and reported that the highest incidence among IDUs (per pyr) was in NY (4.4/100 pyr) and PR (4.9/100 pyr; the overall US incidence for IDUs was estimated at about 1.5/100 pyr). Recent reviews of HIV incidence in the United States 11 report that although incidence decreased among IDUs since the mid 1980s, it remained high in the East, at 1 to 3/100 pyr, and that continued monitoring among this high-risk group is required.
A recent longitudinal study of Puerto Rican IDUs and crack smokers in NY and PR has reported significant differences between the 2 locations in terms of access to risk reduction services (eg, methadone maintenance treatment programs [MMTPs], 12 needle exchange programs [NEPs]), 13 and in levels of risk behaviors. 14,15 An examination of incidence rates in the 2 locations can be helpful in identifying potential public health and service needs, and in assessing the impact of the differential risk characteristics of the 2 environments on seroconversion. In addition, sexual transmission among drug users has been identified as a growing concern, 16 and this may be particularly important for crack users who trade sex for drugs or money. An examination of the impact of sexual risk behaviors on seroconversion among IDUs and crack users can enhance our understanding of the importance of sexual transmission in the future of the HIV epidemic among drug users.
Many variables have been found to be associated with HIV seroconversion or risky behaviors among drug users, including homelessness, 17 injecting in shooting galleries, 18 young age or recent onset of injection, 19 injection with equipment used by others, 20 and frequency of injection. 21 Protective factors related to HIV seroconversion have also been identified and include drug treatment 22 and participation in NEPs. 23 Crack use has been associated with HIV transmission due to its relationship with unprotected sex. 24 Sex-related risk behaviors found to be related to HIV transmission in drug users include number of sex partners, 25 same-gender sex, 26,27 unprotected sex, 28 and sex trading. 29
This paper reports seroconversion rates and examines factors related to seroconversion among Puerto Rican drug users in 2 very different environments, with different distributions of factors related to increasing seroconversion (eg, shooting gallery use) and decreasing seroconversion (eg, availability of drug treatment). The opportunity to identify the characteristics and behaviors of those drug users who are continuing to seroconvert, in what appear to be declining epidemics in both locations, is of great importance for identifying where prevention resources should be directed.
Sample and Recruitment
A total of 1199 participants were recruited between January 1998 and June 1999, 800 in NY and 399 in PR. Participants were recruited in East Harlem, NY, and Bayamón, PR, using targeted sampling methods, based on ethnographic mapping of the 2 communities to identify areas where drug users could be found. 30 Extensive procedures were instituted to ensure comparability of study methods and instrumentation in both sites. 14,15 Criteria for recruitment included age 18 or older, self-identification as Puerto Rican, injected drugs or smoked crack within the prior 30 days, and recent (48-hour) use of heroin or cocaine (verified by urinalysis using Roche OnTrak). Participants were brought to the local field site, informed consent was obtained, and a computer-assisted interview was administered. Participants were paid $15 for the baseline interview in NY and $20 in PR (consistent with prior studies conducted in the 2 communities). The baseline and follow-up interviews included sections on demographics, drug use, and HIV risk behaviors (lifetime and prior 30 days). All participants were provided with HIV counseling immediately after each interview. Voluntary HIV testing was provided to all participants at baseline and to seronegative participants at each subsequent follow-up. All project protocols and interview instruments were approved by the local institutional review boards.
There were differences in the services available for drug users in the 2 study sites. For example, in 1998, when fieldwork was initiated, there were 5 methadone drug treatment programs (with multiple clinics) in East Harlem, compared with only 1 clinic servicing the Bayamón area. There were also 4 NEPs serving the East Harlem IDUs compared with 1 in Bayamón, and the East Harlem NEPs operated for more hours per week and had less restrictive exchange policies. 13
At baseline (time 1 or T1), 28% of the sample in NY and 21% in PR were seropositive. For this seroconversion analysis, only those who tested HIV negative at T1 and received at least 1 additional HIV test at any follow-up interview were included. Three follow-up interviews were conducted, at approximately 6 (T2), 36 (T3), and 42 months (T4) after the baseline interview.
There were some differences in follow-up rates in the 2 locations. The T2 follow-up rates were similar: 77% in NY and 83% in PR. The T3 follow-up rates, however, differed, and the data used in this analysis (based on data through November 2002) are based on preliminary T3 rates of 48% in NY and 62% in PR. The lower rate in NY was in part due to the destruction of locator information, used to find people for the T3 follow-up, in the 9/11/2001 terrorist attacks on the World Trade Center (where NDRI had its central offices). In addition, increased police presence due to terrorist concerns (soon after the T3 follow-up interviews were initiated), may have also contributed to lower follow-up rates in NY. 31 The T4 interview was only conducted with those who had a T3 interview, and T4 follow-up rates were 63% in NY and 55% in PR (as of November 2002). Overall, 79% of the NY group and 84% of the PR group participated in at least 1 of the 3 follow-up interviews. Thus, this study includes participants with 2 to 4 interviews, and has a total sample of 723, consisting of 455 in NY and 268 in PR.
Examination of baseline characteristics of the original cohorts in each site, and the T2, T3, and T4 follow-up samples was undertaken (see Results) to assess potential bias in HIV incidence estimates related to losses to follow-up. Significance tests were conducted comparing those retained at T4 with those lost to T4 follow-up. Factors associated with both seroconversion and study retention were the basis for estimating expected incidence in those lost to the study, using observed incidence in participants with and without the factor in question.
Analysis of Seroconversion
Rates of HIV seroconversion per 100 pyr were calculated. Bivariate significance tests were calculated, by site and for the total, comparing seroconverters and nonseroconverters using baseline variables as predictors.
The variables selected for examination as potential predictors of seroconversion were based on the research literature. The questionnaire focused on identifying risk behaviors and their determinants in the 2 sites so that site differences could be examined, and detailed information was collected for the 30 days prior to the interview. All predictor variables are baseline variables, and thus the results can assist in identifying characteristics of those who are at highest risk, and in developing targeted intervention efforts. Three categories of predictors were examined
Gender, age, and homelessness (defined as living on the streets, in abandoned buildings, or in a shelter) have been related to differences in seroprevalence and incidence, and may represent unmeasured behaviors or susceptibilities.
Drug-Related Risk Behaviors
All injection and other drug risk behaviors were based on the 30 days prior to interview, and included any injection drug use, receptive syringe sharing (ie, sharing a syringe used by others), sharing of other injection equipment (includes receptive sharing of cookers, cotton, rinse water, and backloading [the squirting of drug solution from another syringe into the participant’s syringe]), and frequency of injection. Being a new injector (defined as injecting for 5 years or less) was also used as a predictor. Protective behaviors examined included NEP use, and enrollment in MMTP and other drug treatment programs. To include assessment of the adequacy of the NEP service, a measure of the percent of needles provided (calculated by dividing the number of needles obtained from an NEP in the prior 30 days by the number of injections in the prior 30 days) was also used. Injecting in shooting galleries, an environment that has been associated with high-risk injecting, was also included.
Sex-Related Risk Behaviors
Data on sex-related behaviors (prior 30 days) included number of sex partners, any unprotected sex (vaginal or anal), and sex trading (trading sex for money or drugs). Because a substantial proportion of participants reported not engaging in sex during that period, the gender of recent sex partners could not be ascertained. Thus, the measure of same-sex behavior (ie, men having sex with men [MSM] or women having sex with women [WSW]) was based on 3 related questions in the survey. The questions were: (1) How many of the persons you had sex with in the past 30 days were males? Were females? (2) How do you usually identify yourself to other people? (3) Privately, what do you consider your sexual orientation or identity to be? (Response categories were heterosexual, homosexual, gay, lesbian, or bisexual.) A response to any of these questions indicating same-sex behavior was coded as MSM or WSW.
Comparison of Baseline and Follow-up Samples
A comparison of baseline characteristics (including sociodemographics and risk behaviors) of the original cohorts of HIV-negative participants with baseline characteristics of those retained at each follow-up was undertaken (by site) to identify potential sources of bias in the follow-up samples (Table 1). Comparisons of baseline characteristics of the original and the follow-up samples indicated that the follow-up samples in each site were very similar to the baseline samples on most variables assessed.
Overall, the samples in both locations were primarily male (approximately three quarters of the baseline group), and the mean age of the East Harlem sample was older than the Bayamón sample (37.9 years vs. 32.5 years at baseline respectively).
Approximately two thirds of the NY sample had injected drugs (in the prior 30 days) compared with about three quarters of the sample in PR, and about three quarters of both samples reported lifetime injection drug use. The NY sample consisted of fewer new injectors (eg, 22% of the T1 group reported injecting ≤5 years compared with 38% of the Bayamón group). More of the NY sample reported recent crack use (approximately two thirds compared with about one half of the sample in PR).
As reported in prior studies, 2,15 injection-related risk behaviors were significantly higher (P < 0.05 or less) in PR. For example, about half (55%) the injectors recruited in PR reported injecting at least 150 times per month, compared with less than one fifth of the NY sample. The injectors in PR were more likely to report using shooting galleries (80% vs. 23% at T1), and sharing needles and other injection equipment (20% vs. 8% and 50% vs. 23% respectively). Current use of NEP was higher in NY (49% vs. 39%), as was the percent of needles used obtained from NEPs (45% vs. 10%). More of the NY sample reported being in methadone treatment (47% vs. 8%). In terms of current sex-related risk behaviors, there was a trend for higher sex risks in PR: At T1, a higher percent of the PR sample who engaged in sex reported 10 or more sex partners (8% vs. 2%, P < .001) and more reported recent sex trading (18% vs. 13%, P < .10).
Significance tests were conducted to compare those retained at T4 with those who were not retained in terms of baseline characteristics. In NY, the sample followed up at T4 was less likely to be male and was older than the T1 sample not retained at T4. In addition, they were more likely to have used an NEP, received a higher proportion of needles used from NEPs, and were more likely to have been enrolled in MMTP at T1. In PR, similar to the NY sample, those retained at T4 were more likely to have been enrolled in MMTP at T1. No other significant differences were found.
There were a total of 32 HIV seroconverters, 9 in NY (2.0% of all those initially seronegative who had at least 1 follow-up test, 9 of 455) and 23 (8.6%) in PR (23 of 268). The total number of person years at risk were 1019 pyr in NY and 683 pyr in PR. Total seroconversion rates by site were 0.88/100 pyr (95% CI, 0.31–1.45) in NY and 3.37/100 pyr in PR (95% CI, 2.02–4.72; P < 0.001). Approximately half of those who seroconverted in NY were current IDUs (5 of 9, 56%) compared with more than three quarters (19 of 23 or 83%) in PR. All those who seroconverted in NY were current crack smokers (100%) compared with less than half (44%) of seroconverters in PR (Table 2).
Variables Related to Seroconversion
Bivariate analyses were conducted by site and for the total sample (Table 2). None of the sociodemographic variables examined were significantly related to seroconversion in NY. In PR, however, the seroconverters were more likely to be younger (mean age, 29.9 years versus 32.5 years for nonseroconverters; P < 0.05). In NY, seroconversions occurred at all three age levels examined (<30 years, 30–39 years, and 40+ years). In PR, all the seroconversions were among the 2 younger categories. For the total sample, younger age remained significantly associated with seroconversion.
In NY, crack use was a significant predictor of seroconversion (P < 0.01); all the seroconverters in NY were crack smokers. Analysis of injection-related variables was based only on those who had recently injected (prior 30 days). Examination of the relationship between the number of years injecting and seroconversion approached significance in the total sample, with seroconverters injecting a mean of 10.7 years compared with 14.3 years among nonseroconverters (P < 0.10). This difference was primarily due to the results for PR (although not statistically significant), in which about half the seroconverters injected for 5 years or less. In NY, none of the seroconverters had injected for 5 years or less.
For the total sample, frequency of injection was related to seroconversion, with seroconverters having a mean frequency of 204/month versus 116/month among nonseroconverters (P < 0.001). Although this same direction of association was found in each site, it did not reach statistical significance (P < 0.05) by site. Due to the skewness of injection frequency, these tests were also conducted using logarithmic transformations. Similar results were obtained, and frequency of injection was significantly related to seroconversion for the total sample.
Injecting in shooting galleries was associated with seroconversion in the PR sample (P < 0.05) and in the total sample. Overall, 83% of the seroconverters reported using shooting galleries compared with 42% of the nonseroconverters (P < 0.001). Sharing needles or other injection equipment was not found to be associated with seroconversion, although there was a trend for sharing to be related to seroconverting in NY, and to be related to nonseroconversion in PR.
Potential protective factors (ie, NEP use and drug treatment) were also examined. In PR, being enrolled in MMTP was protective (P < 0.05)—that is, associated with a lower risk of HIV infection, and none of the seroconverters in PR reported being in MMTP. This association between MMTP enrollment and HIV infection was also found in the NY sample (although not statistically significant). Being in other types of drug treatment (excluding MMTP) was reported by few participants and was not significantly associated with seroconversion. There was a trend (not statistically significant) for trading sex for money or drugs to be related to seroconversion. Other sex-related variables, including engaging in same-gender sex or considering oneself gay/bisexual/lesbian, and engaging in any unprotected sex, were not related to seroconversion.
The relatively small number of seroconverters, especially in NY, and the fact that some of the analyses would involve 0% or 100% in individual cells (eg, impact of MMTP in PR, or impact of crack use in NY) precluded conducting multivariate logistic regression analyses.
One of the significant predictors of seroconversion—enrollment in MMTP—was significantly related to T4 follow-up status (Table 1) for both the NY and PR samples. Those who were retained at the T4 interview were more likely to have been enrolled in MMTP at baseline. To assess the potential impact of differential losses to follow-up on the observed incidence rates, estimated incidence rates in those lost to follow-up were calculated using observed HIV seroconversion rates for retained participants enrolled and not enrolled in MMTP. The incidence rates obtained for those not retained at T4, for NY and PR were similar (3.45 in NY and 0.77 in PR) to the observed rates.
The HIV incidence rate among the sample of drug users in Bayamón was almost 4 times greater than that found in the East Harlem sample (3.37 vs. 0.88/100 pyr). Although the overall incidence rates among IDUs and crack smokers in NY and in various locations in the United States has typically been below 2/100 pyr since the mid 1990s, 10 the rate in PR remains at an unacceptably high level.
Methadone treatment was found to be protective against seroconversion. Prior research by several investigators, including Metzger et al, 22 has shown a protective effect of methadone treatment, and this study indicates its importance in both resource-rich and in resource-poor environments. Enrollment in other types of drug treatment was not associated with seroconversion. Thus, the importance of increasing availability of methadone treatment as HIV prevention, especially in PR, was indicated.
In NY, all the seroconverters reported current crack use, and only about half had ever injected drugs, indicating that HIV transmission and sex-related risk behaviors among crack cocaine users should be addressed. Although none of the sex-related risk behaviors examined were significantly related to seroconversion, more detailed behavioral analyses may be needed, including risk network variables. Other studies have indicated that in more mature IDU epidemics, sex-related transmission emerges as an important factor. 16,28
Transmission in NY occurred across the age spectrum. In PR, transmission was mainly in younger, newer injectors, consonant with other findings 19 and indicating that interventions that target young users, or those who may be likely to transition into injection users (eg, noninjection heroin and cocaine users) may be particularly effective in preventing HIV infection and should be a focus in PR. A study undertaken to further our understanding of the transition from noninjection drug use to IDU was recently initiated in PR. 32 The higher level of many HIV risk behaviors in PR (including injection frequency, using shooting galleries, and sharing injection-related equipment), and the trends indicating a higher percentage reporting sex trading, points to the need for HIV prevention efforts in PR to focus on both injection-related and sex-related risk behaviors.
The analyses in this paper examined potentially significant variables from baseline data only, to identify those participants at highest risk of subsequent seroconversion. Examination of key variables at follow-up, however, indicated stability in these behaviors. For example, of the 9 seroconverters in NY, 4 reported never injecting (lifetime) at baseline and only 1 reported injecting at any subsequent follow-up interview preceding seroconversion. Similarly, in PR, none of the seroconverters reported being in MMTP at baseline, and none of them reported entering MMTP at subsequent follow-up interviews.
The drug users recruited for this study may not be representative of all IDUs and crack smokers in the 2 communities. Nonetheless, the targeted sampling plan and the geographic and temporal diversity used in recruiting helped ensure a more representative sample. Most of the data collected were self-reported (except for HIV test results). Differential losses to follow-up in the 2 communities may have biased the results; however, comparisons of those followed-up with the original cohorts in each site resulted in only 1 significant difference in the variables found to be related to seroconversion: enrollment in MMTP. Extrapolation from the observed HIV incidence in the retained sample to calculate incidence rates for those lost to follow-up yielded similar estimates. Thus it is unlikely that losses to follow-up related to MMTP enrollment biased the comparison of HIV incidence in NY and PR.
The urgency posed by the high seroconversion rate in PR leads to several recommendations for the island to reduce the level of risk behaviors and seroconversions:
1. Increase availability of MMTP: The drug treatment that was shown to be most effective in reducing HIV seroincidence was MMTP. The relative unavailability of this health service in PR, 12,33 in light of its demonstrated effectiveness in reducing HIV risk, underlines the need to increase this modality in PR.
2. Increase access to new syringes in PR: Our research in PR indicates that although there is an NEP in Bayamón, its operations, in terms of the hours available and the maximum number of syringes that can be obtained, is inadequate for the need. 13 At the initiation of the field component of this research project in 1997, the maximum number of syringes that could be exchanged was 2, and the current maximum is 50% of the needles exchanged, up to a total of 10 (Personal communication, A. Finlinson, 2002). This is in stark contrast with the East Harlem community in NY, where there are multiple NEPs, and generally there is no maximum number of needles that can be exchanged. This difference is particularly disturbing because in PR the mean number of injections per month was found to be approximately 3 times higher than in the NY sample, indicating a need for a larger number of syringes for each IDU.
3. Increase distribution of injection-related equipment: This was indicated by the high rates of sharing injection-related equipment (approximately twice as high in PR compared with NY). Although no significant relationship between sharing equipment and HIV seroconversion was found in this analysis, other research has shown the relationship between paraphernalia sharing and HCV transmission. 34 Work by Colón et al 35 indicating the relationship between joint purchasing of drugs and sharing of paraphernalia points to the need for innovative intervention efforts to reduce risky behaviors involved in drug preparation, including syringe-mediated sharing and paraphernalia sharing.
4. The trend toward an association between sex trading and seroconversion suggests a need to develop and test innovative methods to reduce unprotected sex, especially among those involved in the sex trade.
5. Enhance outreach efforts among drug users: In addition to MMTP and NEP, the third proven intervention related to risk reduction among drug users has been outreach efforts. 36 Funding for outreach should be expanded: to distribute materials; conduct risk assessments; and refer drug users to NEPs, drug treatment, etc. Intervention efforts in shooting galleries, crack prostitution strolls, and other high-risk venues are indicated.
Several recommendations regarding the NY site are also indicated:
1. Maintain NEP, MMTPs, and outreach efforts: The finding that injection-related risk behaviors are relatively low among the IDUs recruited in East Harlem indicates that prevention efforts have reduced risks. 5 The maintenance and even expansion of NEPs, MMTPs, and outreach efforts is critical in continuing and creating further declines in risk behaviors and incidence rates in NY. This is especially important in light of recent research in NY indicating that risk behaviors may be increasing. 6
2. Address sex risks among crack users: As indicated by the current study and others, 16,28 sex-related risk behaviors are becoming increasingly important in the HIV epidemic among drug users. Enhancing awareness of sex risks for drug users, especially among HIV and other service providers with whom users come in contact, and supporting the development of innovative outreach and risk reduction efforts, must be undertaken.
Recommendations regarding topics requiring additional research can also be made. The relationship between various risk reduction services (eg, NEP, MMTP, outreach) and the proper mix needed of these types of interventions in various resource-rich and resource-poor environments needs to be conducted. 37 The unexpected difference found in the relationships between syringe/paraphernalia sharing and seroconversion (positively related in NY, as expected, but in the opposite direction in PR) requires further investigation. Preliminary analysis of the relationship between sharing and sociodemographic characteristics indicates that in PR, injection-related sharing is less frequent among the younger, newer injectors and thus is negatively related to seroconverting. It may be that seroconversion in this younger group is more related to sex risks. Our data indicate that in PR, higher levels of unprotected sex were found in the younger seroconverters, and further research on the relationship between age- and sex-related risk in drug users should be conducted. Finally, continued monitoring of risk factors for seroconversion among high-risk populations such as those included in this paper are needed to identify subgroups and communities in which intervention efforts are needed.
It has been encouraging that reductions in prevalence and risk behaviors among high-risk drug users have been found in prior studies in NY and PR. The relatively high sero-incidence found among drug users in PR, however, among the highest in the United States and the Caribbean, imparts an ethical as well as a health mandate to action. The recent attention to enhancing HIV prevention efforts in the Caribbean 38 should be expanded to address the HIV–drug use epidemic in PR.
The authors acknowledge Drs. Holly Hagan, Shiela Strauss, and Peter Flom for their suggestions regarding statistical analysis.
1. Des Jarlais DC, Friedman SR, Sotheran JL, et al. Continuity and change within an HIV epidemic. JAMA
2. Robles RR, Colón HM, Matos TD, et al. Risk factors and HIV infection among three different cultural groups of injection drug users. In: Brown BS, Beschner GM, eds. Handbook on Risk of AIDS: Injection Drug Users and Sexual Partners
. Westport, CT: Greenwood Press; 1993:256–274.
3. Andia JF, Deren S, Robles RR, et al. HIV-Related Risk Behaviors for Puerto Rican Drug Users in New York and Puerto Rico
. Presented at the annual meeting of the American Public Health Association; Indianapolis, IN; 1997.
4. Matos TD, Robles RR, Marrero CA, et al. Crack Use in Puerto Rico: Evidence of a Recent Epidemic
. Presented at the annual meeting of the American Public Health Association; San Francisco, CA; 1993.
5. Des Jarlais DC, Perlis TE, Friedman SR, et al. Declining seroprevalence in a very large HIV epidemic: injecting drug users in New York City, 1991 to 1996. Am J Public Health
6. Des Jarlais DC. Trends in HIV Among IDUs in New York City
. Presented at the 14th International AIDS Conference; Barcelona, Spain; 2002.
7. Friedman SR, Chapman TF, Perlis TE, et al. Similarities and differences by race/ethnicity in changes of HIV seroprevalence and related behaviors among drug injectors in New York City, 1991–1996 . J Acquir Immune Defic Syndr
8. Deren S, Robles RR, Andía JF, et al. Trends in HIV seroprevalence and needle sharing among Puerto Rican drug injectors in Puerto Rico and New York: 1992–1999. J Acquir Immune Defic Syndr
9. Des Jarlais DC, Marmor M, Friedman P, et al. HIV incidence among injection drug users in New York City, 1992–1997: evidence for a declining epidemic. Am J Public Health
10. Holmberg S. The estimated prevalence and incidence of HIV in 96 large US metropolitan areas. Am J Public Health.
11. Quan VM, Steketee RW, Valleroy L, et al. HIV incidence in the United States, 1978–1999. J Acquir Immune Defic Syndr
12. Oliver–Velez D, Beardsley M, Deren S, et al. The Impact of Methadone Treatment on HIV Risk Behaviors Among Puerto Rican IDUs in East Harlem, New York and Bayamón, Puerto Rico
. Presented at the annual meeting of the American Public Health Association; Chicago, IL; 1999.
13. Finlinson HA, Oliver–Vélez D, Colón HM, et al. Syringe acquisition and use of syringe exchange programs by Puerto Rican drug injectors in New York and Puerto Rico: comparisons based on quantitative and qualitative methods. AIDS Behav
14. Colón HM, Robles RR. Deren, et al. Between-city variation in frequency of injection among Puerto Rican injection drug users: East Harlem, New York, and Bayamón, Puerto Rico. J Acquir Immune Defic Syndr
15. Deren S, Oliver–Velez D, Finlinson HA, et al. Integrating qualitative and quantitative methods: comparing HIV-related risk behaviors among Puer-to Rican drug users in Puerto Rico and New York. Subst Use Misuse
16. Kral AH, Bluthenthal RN, Lorvick J, et al. Sexual transmission of HIV-1 among injection drug users in San Francisco, USA: risk-factor analysis. Lancet
17. Rockwell R, Friedman SR, Sotheran JL, et al. Medical access for injecting drug users. In Singer M, ed. The Political Economy of AIDS
. Amityville, NY: Baywood Press; 1998:131–147.
18. Des Jarlais DC, Friedman SR, Stoneburner R. HIV infection and intravenous drug use: critical issues in transmission dynamics, infection outcomes, and prevention. Rev Infect Dis
19. Fennema JSA, van Ameijden EJC, Van Den Hoek A, et al. Young and recent onset injecting drug users are at higher risk for HIV. Addiction
20. Koester S, Hoffer L. Indirect sharing: additional HIV risks associated with drug injection. AIDS Public Policy J.
21. Marmor M, Des Jarlais DC, Cohen H, et al. Risk factors for infection with human immunodeficiency virus among intravenous drug abusers in New York City. AIDS
22. Metzger DS, Navaline H, Woody GE Drug abuse treatment as AIDS prevention. Public Health Rep.
23. Normand J, Vlahov D, Moses L, eds. Preventing HIV Transmission: The Role of Sterile Needles and Bleach
. Washington, DC: National Academy Press; 1995.
24. Edlin BR, Irwin KL, Faruque S, et al. Intersecting epidemics—crack cocaine use and HIV infection among inner-city youth. N Engl J Med
25. Elifson KW, Boles J, Darrow WW, et al. HIV seroprevalence and risk factors among clients of female and male prostitutes. J Acquir Immune Defic Syndr
26. Friedman SR, Ompad DC, Maslow C, et al. HIV prevalence, risk behaviors, and high-risk sexual and injection networks among young women injectors who have sex with women. Am J Public Health
27. Maslow C, Friedman SR, Perlis TE, et al. Changes in HIV seroprevalence and related behaviors among male injection drug users who do and do not have sex with men: New York City, 1990–1999. Am J Public Health
28. Strathdee SA, Galai N, Safaiean M, et al. Sex differences in risk factors for HIV seroconversion among injection drug users: a 10-year perspective. Arch Intern Med
29. Astemborsk J, Vlahov D, Warren D, et al. The trading of sex for drugs or money and HIV seropositivity among female intravenous drug users. Am J Public Health
30. Oliver–Velez D, Finlinson HA, Deren S, et al. Mapping the air-bridge locations: the application of ethnographic mapping techniques to dual-site study of HIV risk behavior determinants in East Harlem, New York and Bayamón, Puerto Rico. Hum Org
31. Deren S, Shedlin M, Hamilton T, et al. Impact of the September 11th attacks in New York City on drug users: a preliminary assessment. J Urban Health
32. Colón HM. HIV Risks and Transitions to Injection among Non-IDUs
[proposal]. Grant no. RO1DA015621, 2002. Funded by the National Institute on Drug Abuse.
33. Robles R, Mastro TD, Colón HM, et al. Determinants of health care use among Puerto Rican drug users in Puerto Rico and New York City. Clin Infect Dis
34. Hagan H, Thiede H, Weiss NS, et al. Sharing of drug preparation as a risk factor for hepatitis C. Am J Public Health
35. Colón HM, Finlinson HA, Robles RR, et al. Joint drug purchases and drug preparation risk behaviors among injection drug users. AIDS Behav.
36. National Institute on Health. Interventions to Prevent HIV Risk Behaviors: Consensus Statement.
Bethesda, MD: NIH, Office of the Director; 1997:1–41.
37. National Institute on Drug Abuse. NIDA Community-Based Outreach Model: A Manual to Reduce the Risk of HIV and Other Blood-Borne Infections in Drug Users
. Rockville, MD: National Institute on Drug Abuse, 2000.