This study examined national substance abuse treatment data (TEDS) and two research studies, one of young injection drug users (CIDUS-III) and one of young noninjecting heroin users (NIHU-HIT). A striking finding in our analysis is the substantial decline in young NH-Black heroin injectors entering treatment and enrolling in research studies. As an example of the latter, NH-Blacks were only 12% of Baltimore's CIDUS-III sample despite constituting 64% of the city's population (US Census Bureau, 2000
). In comparison, the ALIVE study begun in the late 1980's in Baltimore recruited a sample that was 95% African-American (Vlahov, et al., 1991
). More recently, the proportion of Whites in Baltimore's second and third Risk Evaluation and Assessment of Community Health (REACH) studies of 15−30 year old IDUs, increased from 31% in 1997−1999 (Fuller, et al., 2001
) to 71% in 2000−2002 (Havens, et al., 2006
) using similar recruiting methods. The latter study recruited young noninjecting drug users simultaneously with IDUs, and the non-IDU sample was 78% Black. In Chicago NH-Blacks comprised over half the heroin snorters in NIHU-HIT but only 2% of the IDUs in CIDUS-III, despite both studies recruiting concurrently at the same locations.
Studies conducted elsewhere also support these findings. Some report that Blacks have declined as a proportion of subjects in studies of IDUs (Friedman, et al., 1999
; Sherman, et al., 2005
; Garfein, et al., 1998
) and are less likely to be newer (<7 years) or recently transitioned (≤2 ye ars) injectors (Des Jarlais, et al., 1999
; Fuller, et al., 2002
). In New York City, Black heroin snorters were less likely to have ever injected or to resume or initiate injecting (Neaigus, et al., 2006
; Neaigus, et al., 2001
). The 1999−2002 National Health and Nutrition Examination Survey found that among persons 20 to 39 years old, lifetime history of injection was more common in NH-Whites than NH-Blacks, while the reverse was true for those 40 to 59 years old (Armstrong, et al., 2006
). An analysis of the 1979−2002 National Household Survey on Drug Abuse found that NH-Blacks had higher lifetime probability of drug injection than NH-Whites in cohorts born before 1955, but lower in cohorts born after 1955 (Armstrong, 2007
). These national surveys were limited, however, in their underrepresentation of populations with high rates of injection drug use, including the homeless. As a point of comparison, half of CIDUS-III participants reported recent homelessness (Garfein, et al., 2007
) and other studies of young IDUs report equivalent or higher levels (Havens, et al., 2004
; Lankenau, et al., 2007
; Evans, et al., 2003
). Whether young NH-Blacks who currently use heroin intranasally can avoid injection for the rest of their heroin use careers is unknown. The event most likely to alter this trajectory is a decline in the quality of heroin to a level that makes snorting it untenable (Strang et al, 1997
; Frank, 2000
A second key finding is that injection trends for NH-Whites moved in the opposite direction compared to NH-Blacks. Consistent with this finding, researchers in Ohio reported increases in heroin use and injecting among White young adults in middle- and upper middle-class families, often in suburban or rural areas (The Ohio Substance Abuse Monitoring Network, 2006
). In South Florida, White noninjecting heroin users transitioned faster to injection than did other racial/ethnic groups (Kelley and Chitwood, 2004
) and a study in San Francisco reported that IDUs younger than 30 years were significantly more likely to be White than were older IDUs (Kral, et al., 2000
Of the 5 PMSAs examined using TEDS, Baltimore and Chicago showed the greatest increases in heroin injection in younger NH-Whites. However, injection was so pervasive in Los Angeles and Seattle that similar increases were impossible. The variation in injection prevalence probably reflects geographic differences in the type of heroin available. Mexican black tar heroin, the dominant form in most Western states, is difficult to use intranasally and is usually injected, while powdered South Asian and South American heroin that is easily snorted dominates the Midwestern and Eastern PMSAs (NDIC, 2006
). The smaller proportion of young Whites among heroin-using IDUs in TEDS for West Coast PMSAs may be due in part to methamphetamine having a greater presence in these areas (SAMHSA, 2007
). In CIDUS-III, over a quarter of participants at the West Coast sites mostly injected methamphetamine/amphetamine, while almost none did so at the other sites.
While these trends clearly warrant further investigation, several explanations have been advanced regarding declines in heroin injection and use among young African-Americans and increases in young Whites. Some hypothesize that African-American communities have experienced particularly severe consequences from drug injection, including high rates of HIV infection, that motivate members to avoid this practice (Johnson, et al., 1998
; Boyle and Brunswick, 1980
Agar and Reisinger hypothesize that heroin epidemics cluster in groups experiencing “open marginality,” a rapid and negative change in the ability to realize socioeconomic expectations (Agar and Reisinger, 2001
). For instance, they suggest that the Civil Rights movement in the 1960s raised African-Americans' expectations beyond actual opportunities for moving into mainstream American life, causing pain and dissatisfaction some treated with heroin. The authors suggest that more recent declines in well-compensated industrial/blue-collar jobs generate gaps between the expectation and actual ability of many young Whites' to achieve a standard of living at least comparable to that of their parents. For young Whites who grew up in neighborhoods where drug injection and HIV were rare or hidden, there appears to be fewer inhibitions about initiating injection. Bourgois and colleagues similarly hypothesized that differences they observed between older African-American and White IDUs in San Francisco, including differences in self-identity and drug use practices, were tied to racial/ethnic variations in how major social structural changes in the 1960s-1970s, including upheavals in urban labor markets, were experienced (Bourgois et al, 2006
). It is worth noting that these explanations avoid ascribing risk status to racial and ethnic groups, a practice that implicitly suggests behaviors, such as injection drug use, are mostly a product of a specific racial/ethnic culture (Schiller, 1992
Our findings have implications for future HIV trends. Infections attributable to sharing contaminated injection equipment should decline markedly among African-Americans, though those who do inject will remain at elevated risk due to the high background prevalence of HIV in African-American communities. While needle sharing is common among young White IDUs (Kral, et al., 2000
; Novelli, et al., 2005
; Bailey, et al., 2007
) and suburban residents (Thorpe, et al., 2001
), it often takes place in low HIV prevalence settings that produce few infections (Garfein, et al., 2007
). To better understand the potential for future infection, research is needed that examines how risk networks of young IDUs evolve over time to see if changes in membership, resources, means of support and other attributes cause a drift toward new network configurations marked by higher HIV prevalence.
Our findings also have implications for other injection-related viral infections. For example, the prevalence of hepatitis C virus (HCV) infection, including the difficult-to-treat genotype 1, is greater among African-Americans than Whites (Armstrong, et al., 2006
; Nainan, et al., 2006
; Conjeevaram, et al., 2006
; Layden-Almer, et al., 2003
). Absent a vaccine, new HCV infections should decline among African-Americans and may increase among Whites. In the near term, an aging cohort that initiated injection in the 1970s will increasingly seek care for conditions related to chronic HCV infection.
These findings also suggest an avenue of research regarding needle exchange programs (NEPs). Given consistent evidence of the effectiveness of NEPs in discouraging needle sharing (GAO, 1993
; Lurie, et al., 1993
; Wodak and Cooney, 2005
; National Academy of Sciences, 2006
; Gibson et al 2001
), opposition to NEPs now largely centers on the contention that these programs enable and signal approval of injection. Reviews of research have found no evidence to indicate NEPs contribute to the initiation of drug injection (GAO, 1993
; Lurie, et al., 1993
; Wodak and Cooney, 2005
; National Academy of Sciences, 2006
), but the 1995 review by the National Research Council and Institute of Medicine suggested revisiting these findings to see if they held up over the next decade (Normand, et al., 1995
). Since that report, NEPs became commonplace. Because it is likely that NH-Blacks were exposed more often to NEPs than were other groups, due to a greater likelihood of living in neighborhoods with relatively high levels of drug injection and HIV infection, future research should assess the relationship between NEPs and the large decline in drug injection among NH-Blacks reported in this and other studies.
Our study has several limitations. First, routine efforts to improve TEDS may account for some historical variation in the data. Second, the extent to which TEDS covers admissions data is affected by differences in State systems of licensure, certification, accreditation, and disbursement of public funds and the data set may underrepresent persons receiving treatment in private programs, the Veterans Administration hospitals, and correctional facilities. As a result, selection biases may exist that influence the racial/ethnic composition of TEDS data. However, assuming that such biases are relatively stable across time, our analyses of trends should not be affected substantially. Further, veterans and criminal justice referrals are present in TEDS, constituting 2% and 10% of all TEDS admissions in 2004, respectively. Third, TEDS does not represent the prevalence of substance use in the general population. Fourth, TEDS is an admission-based system, and therefore multiple admission records for the same individual in a single year may occur. However, the number of previous episodes reported in our TEDS sample did not differ substantially by race/ethnicity and year. In addition, within a single continuous treatment episode, TEDS is able in many cases to distinguish between an initial admission and subsequent transfers to different service types. Fifth, all 3 data sources relied on self-reports, which are subject to inaccurate recall and socially desirable reporting. However, substance abuse treatment data have been shown to be reliable and consistent (Adair, et al., 1996
; Turner and Hubbard, 1995
; Adair, et al., 1995
), and both CIDUS-III and NIHU-HIT used short recall periods and A-CASI interviews that minimize socially desirable reporting (Ghanem, et al., 2005
; Perlis, et al., 2004
). Sixth, the representativeness of participants in CIDUS-III and NIHU-HIT is unknown, though both employed multiple methods to recruit a cross-section of participants. Despite these potential limitations, the consistency of the multiple data sources we examined suggests that findings presented here reflect real changes in the racial/ethnic composition of heroin-using IDUs.
It is worth noting that our data have little to say about the risk practices of IDUs, such as injection frequency and the sharing of injection paraphernalia, that are key determinants in assessing IDUs' risk for bloodborne infections, abscesses and other injection-related morbidities. Here we sought only to examine US trends in injection overall. Future research should examine whether and to what extent demographic changes in injection patterns influence individual level risk practices.
In summary, our findings document important changes in the profiles of young heroin injectors likely to shape future trends in HIV and hepatitis C infection and their treatment. Developing a greater understanding of the reasons young NH-Blacks are refraining from injecting and the extent to which NH-Whites constitute a new and growing cohort of heroin IDUs could advance the prevention and treatment of substance abuse and its associated ills.