David Buchanan a,, Janet A. Tooze b, Susan Shaw c, Mark Kinzly d, Robert Heimer d, Merrill Singer c

Transcription

1 Drug and Alcohol Dependence 81 (2006) Demographic, HIV risk behavior, and health status characteristics of crack cocaine injectors compared to other injection drug users in three New England cities David Buchanan a,, Janet A. Tooze b, Susan Shaw c, Mark Kinzly d, Robert Heimer d, Merrill Singer c a School of Public Health, University of Massachusetts, 306 Arnold House, Amherst 01002, MA 01003, USA b School of Medicine, Wake Forest University, Winston-Salem, NC 27157, USA c Hispanic Health Council, Hartford, CT 06106, USA d School of Public Health, Yale University, New Haven, CT 06520, USA Received 7 June 2004; received in revised form 11 July 2005; accepted 11 July 2005 Abstract Objectives: To compare demographic, HIV risk behaviors, and health status characteristics of injection drug users (IDUs) who have injected crack cocaine with IDUs who have not. Methods: Nine hundred and eighty-nine IDUs were recruited in New Haven, CT, Hartford, CT and Springfield, MA from January 2000 to May Participants were administered a modified version of the National Institute on Drug Abuse Risk Behavior Assessment Questionnaire. Results: Nine percent (n = 89) of participants reported ever injecting crack cocaine and 4.2% (n = 42) reported injecting crack in the past 30 days. Lifetime and current crack injectors did not differ significantly on any demographic characteristics. Lifetime and current crack injectors did not differ on gender, age or marital status from IDUs who have never injected crack. Significant differences were found on race, education, employment and residence, with crack injectors more likely to be white, employed, better educated and living in New Haven than IDUs who have never injected crack. After adjusting for current (past 30 day) speedball and powder cocaine injection, crack injectors reported higher rates of risky drug use behaviors and female crack injectors reported higher rates of risky sexual behaviors. Crack injectors reported higher rates of abscesses, mental illness and Hepatitis C infection, but not Hepatitis B or HIV infection. Conclusions: The emergence of crack cocaine injection requires urgent attention, as this new drug use behavior is associated with elevated rates of high risk behaviors Elsevier Ireland Ltd. All rights reserved. Keywords: Injection drug use; HIV/AIDS risk behaviors; Crack cocaine 1. Introduction Ethnographic studies from Britain and the United States have reported evidence of crack cocaine injection over the last decade (Hunter et al., 1995; Johnson and Ouellet, 1996; Carlson et al., 2000; Sterk et al., 2000). This report presents data on the demographic, behavioral and health status characteristics related to HIV risk among injection drug users Corresponding author. Tel.: address: buchanan@schoolph.umass.edu (D. Buchanan). (IDUs) who have injected crack cocaine compared to IDUs who have not injected crack cocaine. In one of the earliest reports of the phenomenon, Johnson and Ouellet described the emergence of crack cocaine injection in the US in Dissolving crack cocaine for injection is different than dissolving powder cocaine or heroin for injection since crack is not soluble in water. Hence, crack injectors must learn new behaviors. Golub and Johnson (1996) suggested that the adoption of crack cocaine injection might thus follow a classic diffusion of innovation pattern (Rogers, 1983), with innovators disseminating the new behavior to early adopters, and so on /$ see front matter 2005 Elsevier Ireland Ltd. All rights reserved. doi: /j.drugalcdep

2 222 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) In a study on emerging patterns of drug use, Sterk and Elifson (2000) found an increasing prevalence of crack cocaine injection among 40 active female drug users. They reported that a number of primary heroin injectors began injecting crack cocaine because of difficulties in obtaining powder cocaine for speedballs (mixtures of heroin and cocaine). In addition, they found that diffusion processes relating to the initiation of crack cocaine injection varied by neighborhood, with informal channels playing a key role in one neighborhood, while drug dealers played an active role in instructing new users in another neighborhood. In an ethnographic study of 16 crack cocaine injectors, participants reported that they smoked crack cocaine before they began injecting it and most also had a history of speedball injection (Carlson et al., 2000). While learning this new behavior thus appears to diffuse through different channels in different locales, these and other studies suggest two major routes to crack injection. Drug users experienced in smoking crack may decide to try injecting it, possibly to overcome the effects of tolerance (Fuller et al., 2002; Irwin et al., 1996). Alternatively, drug users who are experienced speedball users may seek out crack as a substitute for powder cocaine, for reasons, such as availability and cost. Previous studies have reported higher rates of crack cocaine smoking among African Americans than other population groups (Carlson et al., 1999; Kral et al., 2000; Diaz et al., 2001; Substance Abuse and Mental Health Services Administration, 2002; Drug and Alcohol Services Information System, 2002). Crack cocaine smoking, in turn, has been associated with higher rates of HIV infection among African American drug users (Warner and Leukefeld, 1999), although the evidence is mixed. Iguchi and Bux (1997) found lower HIV prevalence in places where older IDUs already infected with HIV were reluctant to associate with younger IDUs who also smoked crack, which appears to have shielded the crack-smoking population from infection. A number of studies have reported higher levels of HIV risk associated with crack cocaine smoking among women (Edlin et al., 1994; Carlson et al., 2000; Erickson et al., 2000; Sterk et al., 2000; Inciardi and Surratt, 2001; Evans et al., 2002). In a study examining the role of sex-for-drugs exchanges in the economy of crack cocaine use, Sterk et al. (2000) found that 43% of 150 female crack smokers reported exchanging sex-for-drugs and that exchangers were younger and more likely to have been homeless than non-exchangers. In in-depth interviews with 30 crack-addicted women working in the sex trade, Erickson et al. (2000) found that crack intensified involvement in drug use and sex trade and led to more dangerous sexual activities. Inciardi and Surratt (2001) reported that crack remains the primary drug of choice in their sample of 851 cocaine-dependent women, of whom 83% reported having traded sex for drugs in last 30 days. Evans et al. (2002) reported that there are gendered pathways into crack cocaine use, with women more likely than men to be introduced to crack cocaine through intimate relationships with male sexual partners. In the only quantitative study to date, Santibanez et al. (2005) conducted secondary analysis of data collected between 1997 and 1999 as part of the second Collaborative Injection Drug Users Study. Based on a cohort of 2198 injection drug users between the ages of 18 and 30 years, they found that 15% reported crack cocaine injection, with large local variations ranging % across the six study sites. The results reported here are part of a larger study of ecological factors influencing HIV risk among IDUs (Singer et al., 2000; Buchanan et al., 2003a,b; Singer and Clair, 2003). Here, we present data on HIV/AIDS risk behaviors among IDUs who have injected crack cocaine compared to IDUs who have not injected crack. The findings add to the growing body of knowledge about HIV/AIDS risks in this new drug user group. This information will be useful to strengthen programs to prevent HIV/AIDS infection in this high-risk population. 2. Methods The Syringe Access, Use and Discard (SAUD) research project was conducted in New Haven, CT, Hartford, CT and Springfield, MA from January 2000 to May The primary purpose of the research was to examine whether the presence or absence of a syringe exchange program and/or a pharmacy that sells over-the-counter syringes was associated with lower levels of risk at the neighborhood level. A detailed discussion of the methods of research has been reported previously (Singer et al., 2000). The research was reviewed and approved by the Institutional Review Boards at the three primary research institutions. Data were collected from a targeted sample (Weibel, 1990; Carlson et al., 1994; Clatts et al., 1995; Singer, 1999). Based on ethnographic descriptions of the neighborhoods in each city, we purposefully selected neighborhoods with higher densities of drug users. In the Connecticut sites, we selected a total of eight neighborhoods in each city: two that had both a pharmacy and a syringe exchange program; two with just a pharmacy; two with just a syringe exchange program and two with neither. Springfield served as comparison site, as neither syringe exchange programs nor pharmacy sales were legally permitted there at the time of the study. Study participants were recruited by two outreach workers and one ethnographer at each site, using outreach protocols standardized across all three sites. The outreach workers were people in recovery and were reflective of the race and ethnicity of populations at high risk in these cities. Inclusion criteria consisted in being 18 years of age or older, not currently in drug treatment, resident in targeted neighborhood and injection drug use within the past 30 days; current use was confirmed by physical examination of site of injections. Participants were paid a small stipend (US$ 20) in compensation for their time. Epidemiological interviews were completed using a structured questionnaire adapted from the Risk Behavior

3 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) Assessment instrument (National Institute on Drug Abuse, 1991). To enhance the reliability of participant recall, questions concerning risk behaviors were presented for lifetime (ever) and for the past 30 days. Questions were presented in an oral, face-to-face interview format, in English or Spanish at the participant s preference. Standard demographic data were collected on all respondents in a forced-choice format. Questions on the number of injections per month, the amount spent on drugs each month and the number of times one had engaged in various sexual behaviors in the past 30 days were open-ended ( How many times have you... in the past 30 days? ). Questions on injecting with others, needle sharing, and various health outcomes were posed in a forced-choice format (e.g., Have you...ever/in the past 30 days, yes/no? ). Data on all health status indicators are based on self-reports, without independent serologic tests. The SAUD questionnaire started with a section on demographic characteristics, then health history and then drug use history, collected as background information prior to the main part of the questionnaire. The drug use history section asked whether the respondent had injected any of the following seven drugs, presented in the following order, as worded here: heroin by itself; powder cocaine by itself; opiates (nonprescription only, for example, morphine); speedball (heroin and cocaine mixed together); amphetamines ( speed, crystal meth ); crack/rock and other (please specify). For each drug, participants were asked, Have you ever injected heroin by itself? If they answered no, the interviewer went onto the next drug. If they answered yes, the interviewer then asked, Have you injected heroin in the past 30 days? If yes, they were then asked, How many times did you inject heroin in the past 30 days? The results are based on cross-sectional analysis of data collected at one point in time. Chi-square tests were used to compare demographic characteristics of lifetime (ever) crack injectors to IDUs who had never injected crack; t-tests were used to make comparisons for continuous variables (age and age of first use). The same analyses were also done comparing current (past 30 day) crack injectors to IDUs who had never injected crack. For drug use behaviors, sexual behaviors, and health status characteristics, we report medians instead of means for those items where the data were highly skewed. Crude and adjusted odds ratios and corresponding 95% confidence intervals were computed for drug use behaviors, sexual behaviors and health status characteristics using logistic regression on both current and lifetime crack cocaine injection. To compute the odds ratios, data were grouped in intervals to achieve sufficiently large and roughly equal numbers of respondents in each category/cell. In the adjusted analyses, race, education, employment status, city of residence, speedball injection in past 30 days, and cocaine injection in past 30 days were included as covariates in the logistic regression models. All analyses were two-sided and performed with an alpha level of Data were analyzed using SAS Software (SAS Institute, Cary, NC, Version 8.2). 3. Results Data were collected from 989 active IDUs 337 from Hartford, 320 from New Haven and 332 from Springfield. In the total sample, 89 participants (9%) reported ever injecting crack cocaine and 42 participants (4.2%) reported crack injection in the past 30 days. Among current users, the frequency of crack cocaine injection in the past 30 days ranged from 1 to 180 times, with a median of 7.5 times. For the analyses that follow, complete data on drug use history were available on 938 subjects, including 89 lifetime and 42 current crack injectors Demographic characteristics Table 1 presents the demographic characteristics of current (past 30 day) and lifetime (ever) crack injectors, compared to IDUs who report that they have never injected crack. No statistically significant differences were found in the ages of the different user groups, with a mean age of about 38 years, nor for the age of first injection drug use, with a mean age of 19.8 for current crack injectors, 20.3 years for lifetime crack injectors and 21.7 for IDUs who have never injected crack. We also found no statistically significant differences in the gender distribution of current and lifetime crack injectors compared to IDUs who have never injected crack, with males comprising approximately three-quarters of the participants in each group. We found no differences in the marital status of current and lifetime crack injectors compared to IDUs who have not injected crack. As shown in Table 1, we did, however, find significant differences in the racial/ethnic distribution of both current and lifetime crack injectors compared to IDUs who have never injected crack. Both current and lifetime crack injectors were significantly more likely to be white than IDUs who have never injected crack. Approximately half of the current and lifetime crack injectors were white, compared to only 17.1% of IDUs who have never injected crack in this sample. Although the numbers are small, and hence must be interpreted with caution, these differences were consistent across all three study sites, with 51.5% of lifetime crack injectors in New Haven being white, 50.0% of lifetime crack injectors in Hartford and 42.9% in Springfield. We also found significant differences in levels of education, with more than two-thirds of both current and lifetime crack injectors reporting having completed high school (or GED), while only slightly more than half of the IDUs who have not injected crack reported having completed a high school education or more. In addition, we found significant differences in employment status, with 33.3% of current crack injectors and 25.8% of lifetime crack injectors reporting full or part-time employment, compared to 16.3% of IDUs who have not injected crack. The number of lifetime crack cocaine injectors was not evenly distributed across the three study sites. Among lifetime crack injectors, 66 (74.2%) were living in New Haven, 16 (17.9%) in Hartford and 7 (7.9%) in Springfield at the time of this

4 224 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) Table 1 Demographic characteristics of current and lifetime crack cocaine injectors compared to injection drug users who have never injected crack cocaine Current ( past 30 day ) crack cocaine injection (n = 42) Lifetime (ever) crack cocaine injection (n = 89) Age (years) Mean age Age of first drug injection Mean age Gender Male 34 (80.9%) 66 (74.2%) 596 (70.2%) Female 8 (19.0%) 23 (25.8%) 253 (29.8%) Race a White 21 (50%) 45 (50.6%) 145 (17.3%) Black 12 (28.5%) 27 (30.3%) 273 (32.5%) Latino 9 (21.4%) 17 (19.1%) 422 (50.2%) Unknown 0 (0.0%) 0 (0.0%) 9 (1.1%) Education a <High school 13 (30.9%) 29 (32.6%) 382 (45%) High school 29 (69.0%) 60 (67.4%) 466 (54.7%) Unknown 0 (0.0%) 0 (0.0%) 1 (0.3%) Employment a Full/part-time 14 (33.3%) 23 (25.8%) 138 (16.3%) Unemployed 28 (66.7%) 66 (74.2%) 711 (83.8%) Marital status Single 27 (64.3%) 59 (66.3%) 561 (66.1%) Married 2 (4.7%) 6 (6.7%) 67 (7.9%) Separated/divorced 12 (28.6%) 21 (23.6%) 200 (23.5%) Widowed 1 (2.3%) 3 (3.4%) 21 (2.5%) Site a Hartford 4 (9.5%) 16 (17.9%) 292 (34.4%) New Haven 38 (90.5%) 66 (74.2%) 234 (27.6%) Springfield 0 7 (7.9%) 323 (38.0%) a Chi-square test p < Injection drug users who have never injected crack cocaine (n = 849) study. The residence of current crack injectors roughly followed the same distribution across study sites Drug use As shown in Table 2, both current and lifetime crack cocaine injectors were significantly more likely to have injected powder cocaine by itself and to have injected a speedball (a mixture of heroin and cocaine) than IDUs who have not injected crack. Among current crack injectors, 97.6% have injected powder cocaine at some point in the lives, and 86.7% have ever injected a speedball, compared to 60.9% for powder cocaine and 53.8% for speedball among IDUs who have not injected crack. Current crack injectors were about twice as likely to have injected powder cocaine and speedball in the past 30 days than IDUs who have not injected crack. That is, 59.5% of current crack injectors reported injecting powder cocaine in the past 30 days, and 66.6% reported injecting a speedball in the past 30 days, compared to 31.2% for current cocaine injection and 27.4% for current speedball injection among IDUs who have not injected crack. Among lifetime crack injectors, only 4 (4.5%) reported that they have never injected powder cocaine by itself, and 15 (16.9%) reported that they have never injected a speedball. Table 2 Other injection drug use of crack cocaine injectors compared to injection drug users who have never injected crack cocaine Current ( past 30 day ) crack cocaine injection (n = 42) Lifetime (ever) crack cocaine injection (n = 89) Powder Cocaine by itself (ever) a Yes 41 (97.6%) 85 (95.5%) 517 (60.9%) No 1 (2.4%) 4 (4.5%) 328 (38.6%) UK b 0 (0.0%) 0 (0.0%) 4 (0.5%) Powder cocaine by itself (past 30 days) a Yes 25 (59.5%) 49 (55.1%) 265 (31.2%) No 17 (40.5%) 39 (43.8%) 573 (67.5%) UK 0 (0.0%) 1 (1.1%) 11 (1.3%) Speedball (ever) a Yes 36 (85.7%) 74 (83.2%) 457 (53.8%) No 6 (14.3%) 15 (16.9%) 385 (45.5%) UK 0 (0.0%) 0 (0.0%) 7 (0.8%) Speedball (past 30 days) a Yes 28 (66.6%) 48 (53.9%) 233 (27.4%) No 14 (33.3%) 38 (42.7%) 600 (70.7%) UK 0 (0.0%) 3 (3.4%) 16 (1.9%) a Chi-square or Fisher s exact Test p < b Unknown. Injection drug users who have never injected crack cocaine (n = 849)

5 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) Table 3 High risk drug use behaviors of crack cocaine injectors compared to injection drug users who have never injected crack cocaine Ever injected crack (n = 89) Never injected crack (n = 849) Crude OR (95% CI) Adjusted OR a (95% CI) Number of injections in past 30 days b (12.4%) 209 (24.6%) (20.2%) 217 (25.6%) 1.58 (0.73, 3.42) 2.18 (0.95, 5.02) (21.4%) 153 (18.0%) 2.36 (1.09, 5.10) 2.81 (1.16, 6.80) > (46.1%) 270 (31.8%) 2.89 (1.45, 5.75) 3.13 (1.45, 6.74) Money spent on drugs in past 30 days c US$ (13.5%) 172 (22.1%) US$ (20.2%) 177 (22.7%) 1.46 (0.68, 3.12) 1.87 (0.79, 4.45) US$ (23.6%) 235 (30.2%) 1.28 (0.61, 2.67) 1.61 (0.72, 3.56) >US$ (42.7%) 195 (25.0%) 2.79 (1.41, 5.52) 2.17 (1.02, 4.64) Inject alone in past 30 days c Never 18 (20.7%) 89 (10.5%) Sometimes 17 (19.5%) 239 (28.3%) 0.35 (0.17, 0.71) 0.55 (0.24, 1.25) About half of the time 8 (9.2%) 68 (8.0%) 0.58 (0.24, 1.42) 0.51 (0.19, 1.38) Most of the time 27 (31.0%) 138 (16.3%) 0.97 (0.50, 1.86) 1.00 (0.46, 2.16) Always 17 (19.5%) 311 (36.8%) 0.27 (0.13, 0.55) 0.39 (0.18, 0.86) Used needle used by others in past 30 days b Never 62 (70.5%) 707 (83.8%) Ever 26 (29.6%) 137 (16.2%) 2.16 (1.32, 3.54) 1.91 (1.09, 3.35) a Adjusted for: race, education, employment, project site (Hartford, New Haven, Springfield), cocaine (past 30 days), speedball (past 30 days). b Overall p < 0.05 for crude and adjusted analysis using logistic regression. c Overall p < 0.05 for crude analysis using logistic regression; adjusted analysis overall p > 0.05 using logistic regression Risky drug use behaviors Lifetime crack injectors reported higher numbers of injections per month, a median of 104 injections per month versus 78 injections for IDUs who have never injected crack, or about 3.5 injections per day for crack injectors versus 2.6 injections per day for other IDUs. As shown in Table 3, the adjusted odds ratio for injecting more than 120 times per month for lifetime crack injectors relative to IDUs who have not injected crack was 3.13 (95% CI: 1.45, 6.74, p < 0.05). Lifetime crack injectors also reported spending significantly higher amounts of money on drugs each month than other IDUs, a median of US$ 800 versus US$ 500, respectively (or approximately US$ 27 per day for crack injectors versus US$ 17 per day for IDUs who have never injected crack). As shown in Table 3, in the adjusted model, lifetime crack injectors were more than twice as likely (AOR: 2.17, 95% CI: 1.02, 4.64, p < 0.05) to have spent more than US$ 1000 per month than other IDUs. Lifetime crack injectors were almost twice as likely to have reported that they never inject alone (20.7% versus 10.5%) and conversely, half as likely to say that they always inject alone (19.5% versus 36.8%) as other IDUs. Lifetime crack injectors are also significantly less likely than other IDUs to have reported that they have never injected with a needle used by others in the past 30 days. The adjusted odds ratio shows that crack injectors are nearly twice as likely to report having injected with a used needle than non-crack injectors (OR: 1.91, 95% CI: 1.09, 3.35, p < 0.05). When these same models were run to predict current (past 30 day) crack cocaine injection, both the crude and adjusted models became non-significant for number of injections, money spent on drugs and using a used needle; for number of injections and money spent on drugs, the trends in odds ratios were similar, thus the loss of significance may be due to insufficient power for the analysis High risk sexual behaviors With respect to sexual behaviors, a higher percentage of crack injectors reported having had sex in the past 30 days than IDUs who have never injected crack, 71.9% versus 57.6%, respectively. Since there were large differences in the frequencies of sexual behaviors for males and females, we present data only for women in Table 4. As shown in Table 4, female lifetime crack injectors (n = 23) were at eight times the risk of having had 21 or more partners in last 30 days than other female IDUs, after adjusting for current powder cocaine injection, current speedball use and other demographic variables. Female lifetime crack injectors reported a median of four sexual partners in the past 30 days, compared to one partner among female IDUs who have never injected crack. Female lifetime crack injectors reported a median of having had vaginal sex 15 times in the last month, compared to 10 times among other female IDUs; this difference, however, was not found to be statistically significant. Female crack injectors reported giving oral sex a median 20 times in the past 30 days, compared to six times among other female IDUs. Female lifetime crack injectors were found to be at more than four times the risk of giving oral sex 20 or more times in the past 30 days than other female IDUs (OR: 4.78, 95% CI: 1.35, 16.98, p < 0.05). Female crack injectors reported having had sex while high on drugs more frequently than other female IDUs, a median of 15 times versus 10 times, respectively, but this difference was not found to be statistically significant. Finally, we found no statisti-

6 226 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) Table 4 High risk sexual behaviors of female crack cocaine injectors compared to female injection drug users who have never injected crack cocaine Ever injected crack (n = 23) Never injected crack (n = 253) Crude OR (95% CI) Adjusted OR a (95% CI) No. partners in past 30 days c 1 7 (41.2%) 104 (64.2%) (29.4%) 49 (30.3%) 1.52 (0.46, 5.02) 2.08 (0.49, 8.87) 21 (max = 151) 5 (29.4%) 9 (5.6%) 8.25 (2.17, 31.35) 8.03 (1.35, 47.9) No. times had vaginal sex (23.3%) 45 (28.0%) (29.4%) 67 (41.6%) 0.84 (0.21, 3.30) 1.04 (0.21, 5.16) 20 (max = 750) 8 (47.1%) 49 (30.4%) 1.84 (0.52, 6.52) 2.02 (0.44, 9.15) No. times gave oral sex b (47.1%) 123 (75.9%) (52.9%) 39 (24.1%) 3.55 (1.28, 9.82) 4.78 (1.35, 16.98) No. times had sex while high (52.9%) 112 (69.6%) (47.1%) 49 (30.4%) 2.03 (0.74, 5.58) 2.12 (0.63, 7.12) Traded sex for drugs No 10 (58.8%) 124 (76.5%) Yes 7 (41.2%) 38 (23.5%) 2.28 (0.81, 6.41) 2.30 (0.63, 8.36) a Adjusted for: race, education, employment, project site (Hartford, New Haven, Springfield), cocaine (past 30 days), speedball (past 30 days). b Overall p < 0.05 for crude and adjusted analysis using logistic regression. c Overall p < 0.05 for crude analysis using logistic regression; adjusted analysis overall p > 0.05 using logistic regression. cally significant differences in the rates at which female crack injectors report having traded sex for drugs compared to other female IDUs, although the trend is towards higher risk for crack injectors. Among males, lifetime crack injectors were found to be at nearly four times higher risk for having traded sex for drugs (OR: 3.92; 95% CI: 1.65, 9.29), although these trends were not statistically significant in the adjusted model. We found no other significant differences in sexual behaviors in comparisons between male lifetime crack injectors versus male IDUs who have never injected crack Health status indictors As shown in Table 5, lifetime crack injectors reported several indicators of poorer health status than IDUs who have never injected crack. IDUs who had ever injected crack were at almost twice the risk of having had an STD (AOR: 1.86, CI: 1.07, 3.23, p < 0.05). Lifetime crack injectors were found to be at higher risk for having had an abscess and having had an overdose, although these differences disappeared in the adjusted model. In the adjusted models, lifetime crack injectors were found to be at twice the risk of a history of mental Table 5 Self-reported health status indicators of crack cocaine injectors compared to injection drug users who have never injected crack cocaine Ever injected crack (n = 89) Never injected crack (n = 849) Crude OR (95% CI) Adjusted OR a (95% CI) Had STD b Never 57 (64.0%) 659 (78.7%) Ever 32 (36.0%) 178 (21.3%) 2.08 (1.31, 3.30) 1.86 (1.07, 3.23) Had abscess c Never 56 (63.6%) 622 (74.4%) Ever 32 (36.4%) 214 (25.6%) 1.66 (1.05, 2.63) 0.91 (0.53, 1.57) Had overdose c Never 41 (46.6%) 533 (63.0%) Ever 47 (53.4%) 313 (37.0%) 1.95 (1.25, 3.04) 1.41 (0.85, 2.34) History of mental illness d No 67 (77.0%) 707 (84.2%) Yes 20 (23.0%) 133 (15.8%) 1.59 (0.93, 2.70) 2.12 (1.08, 4.16) History of drug treatment No 4 (4.5%) 88 (10.4%) Yes 84 (95.5%) 759 (89.6%) 2.43 (0.87, 6.80) 1.73 (0.56, 5.37) Hepatitis B Never 66 (75.0%) 643 (77.0%) Ever 22 (25.0%) 192 (23.0%) 1.11 (0.67, 1.86) 0.82 (0.45, 1.49) Hepatitis C b Never 51 (58.0%) 570 (68.4%) Ever 37 (42.0%) 264 (31.7%) 1.57 (1.00, 2.45) 1.85 (1.09, 3.13) HIV positive No 63 (75.0%) 593 (74.4%) Yes 25 (25%) 255 (26.6%) a Adjusted for: race, education, employment, project site (Hartford, New Haven, Springfield), cocaine (past 30 days), speedball (past 30 days). b Overall p < 0.05 for crude and adjusted analysis. c Overall p < 0.05 for crude analysis using logistic regression; adjusted analysis overall p > 0.05 using logistic regression. d Overall p > 0.05 for crude analysis using logistic regression; adjusted analysis overall p < 0.05 using logistic regression.

7 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) illness, and at almost twice the risk of having being told that they have Hepatitis C. We also found no significant differences in history of drug treatment (with very high rates in both groups) or rates of Hepatitis B or HIV infection. When we ran these same tests with current crack injectors, all of the differences became non-significant. 4. Discussion The results of this research indicate that IDUs who have ever injected crack cocaine exhibit significantly higher rates of behaviors that put them at high risk of contracting HIV and other blood borne infections than IDUs who have never injected crack. High risk sexual behaviors were found to be especially prevalent among female crack cocaine injectors. We found higher self-reported rates of adverse health outcomes, such as STDs, Hepatitis C and abscesses, among crack injectors, although no differences in self-reported rates of HIV infection. These findings should be interpreted cautiously, due to the limitations of the study methodology. Only three questions were asked about the participants history of crack cocaine injection. It is theoretically conceivable that an individual may have injected crack only once in the distant past, but given the demographic similarities between current and lifetime crack injectors, and the consistent discrimination of these same variables from the characteristics of other IDUs, lifetime crack cocaine injection appears to be a sufficiently discriminant indicator of a new drug use group. That is, the decision to combine all crack injectors was justified by the greater similarity between current and lifetime crack injectors than between lifetime and never crack injectors, even though the decision introduced the possibility for misclassification of exposure. In addition, although there is clearly a large degree of overlap between the group of crack injectors and speedball injectors, the results show that crack injection carries an independent risk that cannot be explained totally by speedball use, and may perhaps be indicative of a willingness to experiment and take risks. Furthermore, the study is based on cross-sectional analysis, so the direction of causality whether crack cocaine injection leads to risky behaviors, or whether individuals who are prone to risk are more likely to inject crack cocaine is unknown. Finally, all health status characteristics were based on self-report and were not confirmed through any means of independent corroboration, such as blood tests or medical reports. Despite the limitations, to our knowledge, the results presented here are one of only two quantitative studies of factors associated with the emergence of this phenomenon. Three findings in particular merit additional comment. Looking at the number of crack injectors identified in the three study sites, there appears to be a step-wise progression in the prevalence of crack cocaine injection along the I-91 Interstate highway. Like the flow of narcotics from New York City (Singer and Mirhej, 2005), it appears that this new drug use practice is gradually being introduced to cities farther and farther away from NYC. Based on these data, it appears that the phenomenon may be following the classic diffusion of innovations curve (Rogers, 1983), as suggested by Golub and Johnson (1996) and Sterk et al. (2000). Given the possibility that this innovative drug use behavior may continue to diffuse, outreach programs need to target the populations and social networks at high risk for crack cocaine injection and provide materials addressed to the particular risks, especially for women, associated with this new drug use practice. Second, from previous studies, there appears to be two major routes into crack cocaine injection: either the evolution of smoking crack to injecting it, or in substituting crack for powder cocaine. The results of this research found that crack cocaine injection was more prevalent among IDUs who are white, better educated and more likely to be employed and who have a previous history of injecting powder cocaine. Since smoking crack has been found to more prevalent in the African American population in the US, these findings would seem to support the latter substitution hypothesis. In addition, since we found no differences in rates of speedball injection by race in this sample, we speculate that the innovation in drug-using practices may be diffusing more rapidly in the white, higher socio-economic status social network of IDUs at the time of this research. Our results appear to support the hypothesis that white IDUs with a history of speedball injection and powder cocaine injection may have learned how to substitute crack for powder cocaine before other user groups, which has important implications for outreach and education programs. Recent findings from the study by Santibanez et al. (2005) of large local variation across study sites and higher prevalence in the Lower East Side of New York City, a primarily white population, lend further support to this interpretation. Finally, the apparent lack of association between crack cocaine injection and HIV sero-positivity deserves comment, given the high rates of risky behaviors found in this study. We suggest two possibilities. The first potential explanation is that these null findings may be due the relative recency of the practice of crack injection and/or the failure to get tested since initiating higher risk behaviors. Second, there may be a potential protective effect due to the advantages of race, education and social privilege. Although crack cocaine injection was found to be associated with elevated rates of high risk behaviors, the crack injectors in this sample may have a lower exposure rate due to the lower HIV prevalence within networks of white injectors. Future studies with larger sample sizes are needed to test this hypothesis. One additional factor that may mediate the transmission of HIV is the use of acid to dissolve crack cocaine. Acidification is also a feature of heroin injection in western Europe, where much of the heroin is weakly soluble until the ph of the solution is substantially lowered (Page and Fraile, 1999; Strang et al., 2001). While exposure to acid is known to inactivate HIV when the exposures are on the order of a half hour or

8 228 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) longer (Martin et al., 1985; Tjotta et al., 1991), little is known about the short exposures typical during the preparation and injection of drug solutions. In conclusion, drug use behaviors are known to be in constant flux, with new practices, drug combinations, consumption equipment, populations of users and locations of use changing over time (Singer, 2000, 2005). Crack injection represents an important change in drug use, with potentially significant public health consequences. Given the widespread injection of cocaine and the multiple routes that seem to lead to the initiation of crack injection, there is an urgent need to develop new programs appropriately tailored to the needs and behavioral propensities of this newly emergent drug user group. Acknowledgement The research described in this paper is supported by the National Institute on Drug Abuse, grant #R01 DA12569, Merrill Singer, Principal Investigator. References Buchanan, D.R., Shaw, S., Teng, W., Hiser, P., Singer, M., 2003a. Neighborhood differences in patterns of syringe access, use, & discard among injection drug users: implications for HIV outreach and prevention education. J. Urban Health 80, Buchanan, D.R., Shaw, S., Ford, A., Singer, M., 2003b. Empirical science meets moral panic: an analysis of the politics of needle exchange. J. Public Health Policy 24, Carlson, R.G., Wang, J., Siegal, H., An ethnographic approach to targeted sampling: problems and solutions in AIDS prevention research among injection drug and crack cocaine users. Hum. Organ 53, Carlson, R., Falck, R., Wang, J., Siegal, H., Rahman, A., HIV needle risk behaviors and drug use: a comparison of crack-smoking and nonsmoking injection drug users in Ohio. J. Psychoactive Drugs 31, Carlson, R., Falck, R., Siegal, H., Crack cocaine injection in the heartland: an ethnographic perspective. Med. Anthropol. 18, Clatts, M., Rees, D.W., Atillasoy, A., Hitting a moving target: the use of ethnographic methods in the development of sampling strategies for the evaluation of aids outreach programs for homeless youth in New York City. In: Lambert, E., Ashery, R., Needle, R. (Eds.), Qualitative Methods in Drug Abuse and HIV Research. NIDA Research Monograph 157. National Institute on Drug Abuse, Rockville, MD, pp Diaz, T., DesJarlais, D., Vlahov, D., Perlis, T., Edwards, V., Friedman, S., Rockwell, R., Hoover, D., Williams, I., Monterroso, E., Factors associated with prevalent hepatitis c: differences among young injection drug users in Lower and Upper Manhattan, New York City. Am. J. Public Health 91, Drug and Alcohol Services Information System (DASIS), March Black Admissions to Substance Abuse Treatment. Office of Applied Studies, Substance Abuse and Mental Health Administration. U.S. Government Printing Office, Washington, DC. Edlin, B., Irwin, I., Faruque, S., McCoy, C., Serrano, C., Inciardi, J., Bowser, B., Schilling, R., Holmberg, S., Intersecting epidemics crack cocaine use and HIV infection among inner-city young adults. Multicenter Crack Cocaine and HIV Infection Study Team. NEJM 331, Erickson, P., Butters, J., McGillicuddy, P., Hallgren, A., Crack and prostitution: gender, myths and experiences. J. Drug Issues 30, Evans, R., Forsyth, C., Gauthier, D., Gendered pathways into and experiences within crack cultures outside the inner city. Deviant. Behav. 23, Fuller, C.M., Vlahov, D., Ompad, D.C., Shah, N., Arria, A., Strathdee, S.A., High-risk behaviors associated with transition from illicit non-injection to injection drug use among adolescent and young adult drug users: a case-control study. J. Drug Alcohol Depend. 66, Golub, A., Johnson, B., The crack epidemic: empirical findings support an hypothesized diffusion of innovation process. Socioecon. Plann. Sci. 30, Hunter, G., Donoghoe, M., Stimson, G., Crack use and injection on the increase among injecting drug users in London. Addiction 90, Iguchi, M.Y., Bux, D.A., Reduced probability of HIV infection among crack cocaine-using injection drug users. Am. J. Public Health 87, Inciardi, J., Surratt, H., Drug use, street crime, and sextrading among cocaine-dependent women: implications for public health and criminal justice policy. J. Psychoactive Drugs 33, Irwin, K.L., Edlin, B.R., Faruque, S., McCoy, H.V., Word, C., Serrano, Y., Inciardi, J., Bowser, B., Holmberg, S.D., Crack cocaine smokers who turn to drug injection: characteristics, factors associated with injection, and implications for HIV transmission. The Multicenter Crack Cocaine and HIV Infection Study Team. J. Drug Alcohol Depend. 42, Johnson, W., Ouellet, L., The injection of crack cocaine among Chicago drug users. Am. J. Public Health 86, 266. Kral, A., Lorvick, J., Edlin, B., Sex- and drug-related risk among populations of younger and older injection drug users in adjacent neighborhoods in San Francisco. J. Acquir. Immune Defic. Syndr. 24, Martin, L.S., McDougal, J.S., Loskoski, S.L., Disinfection and inactivation of the human T lymphotropic cirus type III/lyphadenopathyassociated virus. J. Infec. Dis. 152, National Institute on Drug Abuse, Risk Behavior Assessment Questionnaire. October edition. NIDA Community Research Branch, Rockville, MD. Page, J.B., Fraile, J.S., Lemon juice as a solvent for heroin in Spain. Subst. Use Misuse 34, Rogers, E., The Diffusion of Innovations. Free Press, New York. Santibanez, S., Garfein, R.S., Swartzendruber, A., Kerndt, P.R., Morse, E., Ompad, D., Strathdee, S., Williams, I.T., Friedman, S., Ouellet, L.J., Prevalence and correlates of crack-cocaine injection among young injection drug users in the United States, Drug Alcohol Depend. 77, Singer, M., Studying hidden populations. In: Schensul, J., LeCompte, M., Trotter, R., Cromley, E., Singer, M. (Eds.), Mapping Networks, Spatial Data and Hidden Populations, Book 4, The Ethnographer s Toolkit. Altamira Press, Walnut Creek, CA, pp Singer, M., Drug-use patterns: an ever-whirling wheel of change. Med. Anthropol. 18, Singer, M., Something Dangerous: Changing Illicit Drug Use and Community Health. Waveland Press, Long Grove, IL. Singer, M., Mirhej, G., The understudied supply side: public policy implications of the illicit drug trade in Hartford, CT. Harvard Health Policy Rev. 5, Singer, M., Stopka, T., Siano, C., Springer, K., Barton, G., Khoshnood, K., Gorry de Puga, A., Heimer, R., The social geography of aids and hepatitis risk: Qualitative approaches for assessing local differences

9 D. Buchanan et al. / Drug and Alcohol Dependence 81 (2006) in sterile syringe access among injection drug users. Am. J. Public Health 90, Singer, M., Clair, S., Syndemics and PUBLIC HEALTH: reconceptualizing disease in bi/o-social context. Med. Anthro. Quar. 17, Sterk, C., Elifson, K., German, D., Female crack users and their sexual relationships: the role of sex-for-crack exchanges. J. Sex Res. 37, Sterk, C., Elifson, K., Fluctuating drug markets and hiv risk taking: female drug users and their relationship with drug markets. Med. Anthropol. 18, Strang, J., Keaney, F., Butterworth, G., Noble, A., Best, D., Different forms of heroin and their relationship to cook-up techniques: data on, and explanation of, use of lemon juicwe and other acids. Subst. Use Misuse 36, Substance Abuse and Mental Health Services Administration, Results from the 2001 National Household Survey on Drug Abuse: Volume I. Summary of National Findings. Office of Applied Studies, Rockville, MD, NHSDA Series H-17, DHHS Publication No. SMA Tjotta, E., Hunges, O., Grinde, B., Survival of HIV-1 activity after disinfection, temperature and ph changes, or drying. J Med. Virol. 35, Warner, B., Leukefeld, C., Racial differences in HIV infection and risk behaviors among drug users in a low seroprevalence area. J. Drug Issues 29, Weibel, W., Identifying and gaining access to hidden populations. In: Lambert, E. (Ed.), The Collection and Interpretation of Data from Hidden Populations. NIDA Research Monograph 98. National Institute on Drug Abuse, Rockville, MD, pp