PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Am J Public Health. Author manuscript; available in PMC 2011 August 5.
Published in final edited form as:
PMCID: PMC2866590
NIHMSID: NIHMS312591

Race and distance effects on regular syringe exchange program use and injection risks: A geobehavioral analysis

Abstract

Objectives

Geography has become important in understanding HIV risk and developing prevention strategies to reduce HIV transmission. We conduct “geobehavioral” analyses by race to understand how distances among drug users’ residences, drug purchase locations, injection drug use locations and syringe exchange programs are associated with injection behaviors.

Methods

Data are from the HIV Prevention Trial Network 037 site in Philadelphia, PA, USA (2002-2006), a randomized study evaluating the efficacy of a network-oriented HIV prevention intervention for injection drug users. As part of the study’s pre-screening assessment, participants were asked the nearest intersection to their residence, to where they buy and use drugs, and questions about their injection behaviors.

Results

Geographic distances had both independent and interactive effects on injection risk behaviors and syringe exchange program use. Blacks, regardless of distance, were less likely than whites to inject in public places (OR=0.62, CI: 0.43, 0.90), to use needles after someone (OR=0.27, CI: 0.19, 0.38), and to access syringes from exchange program sites (OR=2.08, CI: 1.48, 2.92). Latinos’ injection behaviors were more distance-dependent than other races.

Conclusions

Distances among injectors’ homes, drug purchase and injecting sites and prevention resources affect safe injection practices differentially by race group.

Keywords: harm reduction, HIV, risk-taking, spatial analysis

Geography has become increasingly important in understanding the epidemiology of HIV, as cities across the country observe geographic clustering of incident and prevalent cases in some neighborhoods but not in others. Injection-related HIV risk, in particular, has been demonstrated to be susceptible to place effects. For example, at the micro-level, research has shown that injection drug users (IDUs) who inject in public locations, such as parks or vacant properties, are more likely to engage in riskier practices and more likely to acquire HIV and HCV1-4. Limited access to sterile prevention supplies, such as unused syringes, bleach or clean rinse water, in public places are a factor in increased risk in these places1. At a macro-level, local area effects of neighborhoods also affect injection practices. Results from a study examining neighborhood differences in syringe access, use, and discard revealed that IDUs living in more economically advantaged neighborhoods were less likely to inject in public places, more likely to have a single source for syringes, and to more appropriately discard syringes5. Additionally, neighborhood characteristics such as perceived social disorder, heightened police presence, and socioeconomic disadvantage have been shown to increase injection-related risks6-9.

Geography has also gained attention in light of persistent racial disparities in HIV. Recent epidemiological evidence shows that HIV prevalence and incidence rates are significantly higher for Blacks than for Whites despite blacks engaging in less HIV risk behavior10-12. Blacks tend to live and carry out daily functions, including selection of drug and sex partners, in neighborhoods with higher background HIV prevalence, and thus are more likely to be embedded in networks that have higher HIV prevalence13,14. Along with having higher HIV prevalence, black neighborhoods also tend to have other characteristics that increase injection risk, including poverty, abandoned properties, and racial targeting by police15-17. While neighborhood studies have broadened our understanding of area effects on injection drug use, they represent only one means of examining geographical context on risk behavior.

In this paper, we examine geographic distances between places of relevance to IDUs as an example of what we term “geobehavioral” analyses, to connote analyses using geographic data to understand behavior and relationships, and their spatial localization18. Whereas place represents lived, meaningful environments, geography also encompasses studies of space, or the relative position between locations19-21. The approach we employ here is distinct from the study of behaviors where analyses fix individuals within a place using a single point location—typically, one’s own residential address. For drug users, where they buy and use drugs are at least equally relevant locations of interest for understanding risk. Our primary interest is in determining whether the distances between these important places (home, where drugs are purchased, where drugs are used, and where prevention services are located) have influence on risk behaviors. Geographic distance is a dimension of accessibility, and represents the ease with which individuals can reach needed services22. Thus, understanding the effects of distance also has implications for placement and utilization of prevention programs such as syringe exchange programs (SEPs). We examine 1) how distances between IDUs’ residences, drug use and purchase locations are related to place of injection (public or private), 2) distances between SEPs and IDUs’ residences, drug use and purchase locations and their effects on syringe source and receptive syringe sharing, and 3) interactions between distance, race, and injection risk to illuminate potential explanations for existing racial disparities in HIV.

MATERIALS AND METHODS

Study Design and Recruitment

Data are from the prescreening database of the HIV Prevention Trials Network 037 (HPTN 037) conducted in Philadelphia, PA, USA. HPTN 037 is a phase III randomized study evaluating the efficacy of a network-oriented peer educator intervention for HIV prevention among IDUs and their network members. Prospective index participants were identified through a community-based recruitment model that employed ethnography and outreach in zip codes with a high prevalence of individuals living with HIV/AIDS as reported by the Philadelphia Department of Public Health. Outreach workers disseminated information about the study, provided verbal and written study descriptions to prospective participants who were encouraged to consider being prescreened to determine eligibility.

Prescreening interviews were conducted on mobile assessment units that were parked in close proximity to risk pockets within the target zip code areas23. Those who appeared eligible at prescreening were invited to complete additional screening to determine study eligibility. Details about the study intervention are reported elsewhere24. The study was approved by the Office of Human Research at the University of Pennsylvania School of Medicine and on-going ethical guidance was provided by the HPTN Ethics Working Group.

Measures

Prescreened participants (N=3,078) were asked the address of their current residence; and individuals who reported injecting drugs within the previous 6 months (N=2,599) were asked to provide the nearest intersection to where they typically purchased drugs and where they last injected. Intersections were defined as the place where the nearest two cross streets meet. Latitudes and longitudes were determined for intersections using ArcView v.3.2 geographic software25, and Manhattan distances (miles) were calculated between each pair of points (2-point distances) for which data were available19. Manhattan distances are calculated along axes of right angles using the shortest paths available by street in order to get from one point to another. For 2-point distances, we calculated the distances between injectors’ home and drug purchase (buy) locations, between their home and injection (use) locations, and between their buy and use locations. Similar calculations were performed for distances between each of these locations and the location of Philadelphia’s SEP sites. We also calculated 3-point and 4-point distances that combine the 2-point distances in either “geographic paths” or “average distances” among points. For example, for geographic paths, we calculated the sum of the distances from injectors’ homes to drug buy locations and from drug buy to use locations (3-point path distance) or from home to SEP to buy to use locations (4-point path distance). So, paths measured routes injectors might take with regard to procuring and using drugs. For average distances, we calculated the mean distance among 3 or 4 points as a more general distance measure.

For race, participants were asked their race and whether they considered themselves to be Latino/Hispanic. Respondents were coded as White if they indicated White race and African American if they indicated African American race, both irrespective of the response on ethnicity. Less than 5% who reported White or African American race also reported Latino/Hispanic origin. Latino/as predominantly classified themselves as “Other” race.

Primary outcomes were place of last injection, usual source for obtaining syringes, and any receptive sharing of syringes and other injection equipment in the past 3 months. Place of last injection was recoded into 3 nominal categories: own/family/friend’s residence (reference category), shooting gallery (1), public place (2) which included cars, abandoned houses, or “other public places.” Syringe source was recoded into 2 categories: SEP (reference) and non-SEP. Receptive sharing of syringes, cookers, cottons, or rinse water (past 3 months) was recoded into 3 ordinal categories: never/almost never (reference), less than twelve times (about once/week) (1), twelve times or more (more than once/week) (2).

Analysis

We analyzed data from participants who reported injecting drugs in the past 6 months (N=2,599). Exploratory data analysis included inspection of variables for missingness, frequency distributions, and geo-mapping of all point locations. Of the address locations that were collected from injectors, we were able to successfully geocode 1,790 home addresses, 2,360 drug purchase locations, and 2,256 injection drug use locations. Therefore, the final sample sizes for analyses varied according to completeness of each individual point location, geocoding success, and completeness after combining point locations to create the distance measures.

The primary outcomes—place of last injection, syringe source, and receptive sharing—were discrete variables, while the primary independent variable, distance, was continuous. Preliminary analyses used ANOVA to test for differences between the discrete levels of each of the outcomes by distance (miles) and racial differences by distance. Next, multiple regression analysis was performed to model the effects of distance and race, and their interaction, on each outcome. Place of last injection was a nominal outcome, so multinomial regression was performed with private residence serving as the reference level to which shooting gallery and public place were compared. We used logistic regression for the dichotomous outcome, syringe source (SEP vs. non-SEP). SEP use was the reference level. After testing the proportional odds assumption, we performed ordinal regression for the outcome variable measuring frequency of receptive sharing, with “never/almost never” serving as the reference level. In all models, we adjusted for age, gender, education, and partner status and included interaction terms for distance by race. We initially adjusted for HIV status, but it consistently had no effect, so we removed it from all models. Self reported HIV prevalence in the prescreening sample was around 2%. Among those who were behaviorally eligible for the intervention and subsequently tested (N=829), HIV prevalence was 8.2% overall—5% for whites, 13% for blacks, and 7% who self identified as Latino. Additionally, depending on the outcome being examined, we included other pertinent covariates, such as with whom one injected, or unstable housing, as defined as living in a shelter, group home, rental room, or homeless. Stata/SE v.8 was used for all statistical analysis26.

RESULTS

Of the 2,599 individuals reporting injecting drugs in the past 6 months, 41% were White, 45% were Black, and 14% were Latino. The mean age was 39 years and ranged from 18 to 75 years. Seventy-five percent of the sample was male and 68% reported having a high school diploma or equivalent. Just over half (54%) reported having a primary sex partner. For outcome variables, 54% reported injecting in a private residence at their last injection, while 34% reported injecting in public venues and 12% attended a shooting gallery. Thirty-seven percent of the sample regularly used SEPs as their usual source for syringes. Thirty-four percent reported receptive syringe sharing and 44% had used water, cotton, or cookers after someone else.

Figure 1 shows mean distances (miles) between location pairs for the total sample and by race. When place of residence is included in the location pair, we see that distances traveled for Whites to buy (F2,1635=47.89, p<0.000) and use (F2,1562=24.69, p<0.000) drugs, and access SEPs (F2,1775=81.61, p<0.000), is on average twice the distance traveled for Blacks and Latinos. Thus, Blacks and Latinos live closer to where they buy and use drugs than do Whites. The distance between where drugs are bought and used is a relatively short distance for all race groups. Finally, SEPs in Philadelphia appear to be very appropriately placed. They are less than 1 mile away from drug purchasing locations and an average of 1.6 miles from injection drug use locations.

Figure 1
Two-point Manhattan Distances in Miles (Kilometers) by Race, Philadelphia, PA, USA, 2002-2006. (* P < 0.05)

Race, Distance and Place of Last Injection

We fit models examining the effects of two distances on place of last injection—the path from home to buy location to use location, “pathhbu,” and the average distance among these same locations. Table 1 shows results from the two models with pathhbu as the primary predictor of injecting in a shooting gallery, public venue, or private residence (reference). No significant associations were observed between path distance and shooting gallery use, between race and shooting gallery use, nor between a distance by race interaction term and shooting gallery use. Blacks, however, were significantly less likely than Whites to inject in public places as path distance increased (OR=0.94, CI: 0.90-0.98). We fit a similar model using average distance, rather than path, and with a few exceptions, observed similar results. The effects of average distances were in the same direction, but greater in magnitude than those of pathhbu.

Table 1
Multinomial Regression Results: Main and Interaction Effects of Path Distance (Miles) and Race on Place of Last Injection, Philadelphia, PA, USA, 2002-2006 (N=1,443)

Race, Distance and Regular Syringe Source

Table 2 shows the associations of distances between the nearest SEP site and home, buy, and use locations, race and regular source of syringes. All models test the main and multiplicative effects of race and distance on the regular use of non-SEP sources. The three models differ with respect to the distance measure used. Model I uses the distance between the nearest SEP site and injectors’ homes (SEP-home); Model II uses the distance between the nearest SEP site and drug buy location (SEP-buy); and Model III uses the distance between the nearest SEP site and use location (SEP-use). Each model adjusts for age, gender, education, and partner status. The most common non-SEP sources used were “works sellers” (34%), friends (11%), diabetics (11%), and [drug] dealers and other users (8%). In Model I, because there were no significant interaction effects, we interpret the main effects of race and distance as independent effects. We found that for each mile of increased distance between SEP and IDUs homes increased, there was a 6% increased likelihood of using non-SEP sources for syringe access. In this same model, results show that both Blacks (OR=1.65, CI: 1.22-2.25) and Latinos (OR=1.57, CI: 1.03-2.39) were significantly more likely than Whites to access syringes from sources other than SEPs. In Models 2 and 3, we found evidence for significant interactions with distance for Latinos, but not Blacks. As distances increased, Latinos were more likely to use non-SEP sources to access syringes; distance did not affect where Blacks acquired syringes. In both Models 2 and 3, we find that the greater the distance between SEP and buy (OR=6.70, CI: 2.32, 19.4) and use locations (OR=5.35, CI: 2.53, 11.3), Latinos were much more likely than Whites to access syringes from non-SEP sources.

Table 2
Logistic Regression Results: Main and Interaction Effects of Path Distance (Miles) and Race on Regular Use of non -SEP Sources for Syringes, Philadelphia, PA, USA, 2002 -2006

Race, Distance and Receptive Sharing of Injection Equipment

Results on the effects of distance on receptive sharing are displayed in Table 3. For these models, we used the average distance among all four injection-related sites--SEP, home, buy, and use locations. We also fit models using path distances that yielded similar results. We examined the effect of average distance on 1) receptive syringe sharing and 2) using water, cookers, and cotton after someone. Similar to our results for regular source of syringes, we observed significant interaction effects for Latinos, but not for Blacks. Main effects for Blacks indicate that they were the least likely group to use needles (OR=0.27, CI: 0.19-0.38) or other injection equipment (OR=0.37, CI: 0.27-0.52) after someone else, an effect not moderated by distance. Latinos’ use of injection equipment (needles, water, cooker, cotton), however, was moderated by distance, showing that their odds of using a needle or other works after someone else increased by 21% and 24%, respectively, with each mile increase in average distance among the four locations. Results also showed increased odds of receptive syringe sharing after a partner, relative, or friend (OR=1.94, CI: 1.51-2.50) and other non-kin, non-friend users (OR=1.59, CI: 1.10-2.31). With regard to the sharing of rinse water, cooker, and cotton, individuals were more than twice as likely to use these after a partner, relative or friend (OR=2.18, CI: 1.73-2.73), but not with “other” users. Regular use of non-SEP sources of syringes increased the odds of receptive syringe sharing by 60%, but had no effect on using water, cooker, and cotton after someone.

Table 3
Ordinal Regression Results: Main and Interaction Effects of Average Distance (Miles) and Race on Receptive Syringe Sharing, Philadelphia, PA , USA , 2002-2006

DISCUSSION

Distances which IDUs traverse in procuring and using drugs are related to injection risk behaviors, but the relationship varies by race. Injection-related behaviors among Blacks tended to be less distance-dependent than among Whites or Latinos. Independent of distance, Blacks were less likely to inject in public places or to inject after someone else. This is consistent with studies reporting that Blacks engage in less risk behavior than Whites 10-12. On the other hand, our results also showed that, regardless of distance, Blacks were significantly more likely to access syringes from non-SEP sites, including works sellers, drug dealers, and other users. This is disturbing, but consistent with other findings from Philadelphia and other major US cities7,8,27. In a paper assessing the impact of Operation Safe Streets, a police intervention conducted in Philadelphia from 2002 to 2003, researchers found that the intervention had an unintended effect of significantly reducing Blacks’ use of SEPs up to 9-months post-implementation7. Despite the decriminalized approach of the Safe Streets intervention, Blacks’ use of SEPs declined at a rate of more than twice that of Whites7. Our results show that Blacks, who live significantly closer to SEPs than Whites, were up to two times more likely than Whites to access syringes from non-SEP sites. This suggests that geographic accessibility is not a factor in SEP utilization for blacks. Other reasons such as racial profiling, mistrust of the police, secondary exchange, and stigma may be more prominent. The finding deserves some public health attention in light of SEPs established effectiveness in decreasing injection risk, preventing HIV, and serving as a point-of-entry into other services needed by IDUs, primarily drug treatment28-34.

Despite Blacks and Latinos having similar average distances among the four locations, significant interactions with distance were observed for Latinos only. Latinos with greater distances between SEPs and buy and use locations were significantly less likely than Whites to use SEPs as their regular syringe source. Further, unlike Blacks, compared to whites, Latinos were more likely to engage in receptive sharing as average distance among the four locations increased. This may be due to the fact that the Latino community in Philadelphia is much smaller and more concentrated than the Black community. Our data do not address why Latinos appear to be more distance-dependent than Blacks and Whites. Additional research is warranted to understand the experience of Latino injectors.

Other results showed the odds of receptive sharing with partners, relatives, and friends were higher than the odds of using after other, less well-known users. Whereas not using after near-strangers is a helpful precaution, it is important to not use needles and other works after anyone—including family and friends—particularly for Blacks who have high rates of HIV in their social networks and neighborhoods. Refusing to engage in receptive sharing with friends or family may present challenges, as it could raise suspicion of potential infection of or by the person refusing. At the same time, injecting in isolation from others is discouraged because it has been shown to increase the risk of overdose35-36. Interventions that address issues around stigma, open communication, and supporting each other to protect everyone’s health are needed, particularly among users who are socially close such as family and close friends. Additionally, stronger efforts should be taken to ensure that every IDU has consistent and convenient access to his/her own sterile injection equipment, so that the need to use needles and other injection equipment after others is completely eliminated. SEPs contribute to this mission; and, although politically unpopular, the data also highlight the potential value of safe injection facilities.

In this study, we analyzed novel data using point locations of where IDUs purchase and inject their drugs to help illuminate geographic predictors of risk that residential address locations cannot do alone. Analyzing distances offers a more nuanced approach to understanding how geography influences HIV risk behaviors, incidence and prevalence. Because HIV prevalence in our sample was insufficient for analysis, we were limited to examining behavior as an endpoint. It is noteworthy that despite our results indicating less risky injection behavior among Blacks, chi-square results of race by HIV serostatus among study enrollees indicated significantly higher HIV prevalence among Blacks as compared to Whites and Latinos. This is consistent with other studies and highlights the limitations of behaviors in explaining racial disparities in HIV; however, behavior remains a necessary factor in transmission.

The data used were non-random and cross-sectional; therefore, caution should be taken in making causal interpretations. We did not have complete data for all prescreened injectors, but instead reported on available data. Responses were voluntary and answering survey questions, particularly related to specific address locations of one’s residence and illegal activities, may have presented concerns for some respondents. Respondents with missing location information were significantly more likely to inject in public places and shooting galleries and to not use SEPs as their regular source of syringes, but were not more likely to engage in receptive sharing (data not shown). Because the initial sample was not a probability sample, the missingness likely has little effect on generalizability. We feel our results can be generalized to similar diverse populations of injectors who are street-recruited from high risk areas in urban settings. The consistency of our results with IDU studies in other cities supports this claim1-4,8,27-29.

Notwithstanding the limitations, the use of novel data to report on geobehavioral risk for HIV among a large sample of injectors is a major strength of our study and makes a contribution to understanding racial disparities in HIV. Because of the greater HIV prevalence in minority networks and neighborhoods, the least bit of risky behavior could result in transmission. Where people are located in relation to the risks and resources which surround them is an important aspect of understanding effects of the environment on health and behavior, and also developing interventions. Geographic clustering of HIV in minority urban neighborhoods presents unique opportunities for place-based interventions since, to some degree, the epidemic is geographically contained. Individual preventive measures, such as drug treatment, needle hygiene and regular HIV testing and care continue to be important.

Acknowledgments

This work was supported by the National Institutes of Health grants to the Penn HIV Prevention Trials Unit (U01 AI048014) and the Penn Center for AIDS Research (P30 AI45008). The authors would also like to thank all study staff, and participants who provided the data.

Human Participant Protection

This study was approved by the Office of Human Research at the University of Pennsylvania School of Medicine and on-going ethical guidance was provided by the HPTN Ethics Working Group.

Contributor Information

Chyvette T. Williams, University of Illinois at Chicago School of Public Health.

David S. Metzger, University of Pennsylvania School of Medicine.

REFERENCES

1. Latkin C, Mandell W, Vlahov D, et al. My place, your place, and no place: Behavior settings as a risk factor for HIV-related injection practices of drug users in Baltimore, Maryland. Am J Community Psychol. 1994;22(3):415–30. [PubMed]
2. Celentano DD, Vlahov D, Cohn S, et al. Risk factors for shooting gallery use and cessation among intravenous drug users. Am J Public Health. 1991;81:1291–5. [PubMed]
3. Klein H, Levy J. Shooting gallery users and HIV risk. J Drug Issues. 2003;33:751–67.
4. Thorpe LE, Ouellet LJ, Levy J, et al. Hepatitis C virus infection: prevalence, risk factors and prevention of opportunities among young injection drug users in Chicago, 1997-99. J Infect Dis. 2000;182:1588–94. [PubMed]
5. Buchanan D, Shaw S, Teng W, Hiser P, Singer M. Neighborhood differences in patterns of syringe access, use, and discard among injection drug users: Implications for HIV outreach and prevention education. Journal of Urban Health. 2003;80:3. [PMC free article] [PubMed]
6. Latkin C, Williams CT, Wang J, Curry A. Neighborhood disorder as a determinant of drug injection behaviors: A structural equation modeling approach. Health Psychology. 2005;24:96–100. [PubMed]
7. Davis CS, Burris S, Kraut-Becher J, et al. Intensive street-level police intervention on syringe exchange program use in Philadelphia, Pa. Am J Public Health. 2005;95(2):233–236. [PubMed]
8. Cooper H, Moore L, Gruskin S, Krieger N. The impact of a police drug crackdown on drug injectors’ ability to practice harm reduction: a qualitative study. Soc.Sci.Med. 2005;61(3):673–684. [PubMed]
9. Rhodes T, Singer M, Bourgois P, Friedman SR, Strathdee SA. The social structural production of HIV risk among injection drug users. Soc Sci Med. 2005;61(5):1026–44. [PubMed]
10. Hallfors D, Iritani B, Miller W, et al. Sexual and drug behavior patterns and HIV and STD racial disparities: The need for new directions. American Journal of Public Health. 2007;97(1):125–132. [PubMed]
11. Ellen J, Aral SO, Madger LS. Do differences in sexual behavior account for the racial/ethnic differences in adolescents’ self-reported history of a sexually transmitted disease? Sexually Transmitted Diseases. 1998;25:125–9. [PubMed]
12. Broz D, Ouellet LJ. Racial and ethnic changes in heroin injection in the United States: Implications for the HIV/AIDS epidemic. Drug and Alcohol Dependence. 2008;94(1-3):221–233. [PMC free article] [PubMed]
13. Laumann EO, Youm Y. Racial/ethnic group differences in the prevalence of sexually transmitted diseases in the United States: A network explanation. Sexually Transmitted Diseases. 1999;26(5):250–261. [PubMed]
14. Adimora AA, Schoenbach VJ. Social Context, Sexual Networks, and Racial Disparities in Rates of Sexually Transmitted Infections. The Journal of Infectious Diseases. 2005;191:S115–S122. [PubMed]
15. Williams CT, Latkin CA. Neighborhood socioeconomic status, personal network attributes, and use of heroin and cocaine. American Journal of Preventive Medicine. 2007;32(6 Suppl):S203–10. [PMC free article] [PubMed]
16. Burris S, Blankenship KM, Donoghoe M, et al. Addressing the “risk environment” for injection drug users: The mysterious case of the missing cop. The Milbank Quarterly. 2004;82(1):125–156. [PubMed]
17. Maas B, Fairbairn N, Kerr T, et al. Neighborhood and HIV infection among IDU: Place of residence independently predicts HIV infection among a cohort of injection drug users. Health & Place. 2007;13(2):432–439. [PubMed]
18. Camara G, Monteiro AM, Fucks SD, Carvalho MS. Spatial analysis and GIS: A primer. [Accessed June 12, 2009]. http://www.dpi.inpe.br/gilberto/tutorials/spatial_analysis/spatial_analysis_primer.pdf.
19. Agnew J. The United States in the World Economy. Cambridge University Press; Cambridge: 1987.
20. Kearns RA, Joseph AE. Space in its place: developing the link in medical geography. Social Science & Medicine. 1993;37(6):711–717. [PubMed]
21. McLafferty S. Placing substance abuse: geographical perspectives on substance use and addiction. In: Thomas YF, Richardson D, Cheung I, editors. Geography and Drug Addiction. Springer; New York: 2008.
22. Penchansky R, Thomas JW. The concept of access: definition and relationship to consumer satisfaction. Medical Care. 1981;19(2):127–140. [PubMed]
23. Perkins WE, Metzger DS. AIDS in Philadelphia: Emerging from the Shadow of Crack. In: Bowser BP, Quimby E, Singer M, editors. When Communities Assess Their AIDS Epidemics: Results of Rapid Assessment of HIV/AIDS in Eleven Cities. Lexington Books; New York: 2007.
24. Latkin CA, Donnell D, Metzger D, et al. The efficacy of a network intervention to reduce HIV risk behaviors among drug users and risk partners in Chiang Mai, Thailand and Philadelphia, USA. Social Science & Medicine. 2009;68(4):740–8. [PMC free article] [PubMed]
25. Environmental Systems Research Institute, Inc ArcView 3.2a. ESRI; Redlands, CA: 1992 - 2000.
26. StataCorp Stata Statistical Software: Release 8. StataCorp LP; College Station, TX: 2003.
27. Bluthenthal RN, Kral AH, Lorvick J, Watters JK. Impact of law enforcement on syringe exchange programs: a look at Oakland and San Francisco. Med Anthropol. 1997;18(1):61–83. [PubMed]
28. Watters JK, Estilo MJ, Clark GL, et al. Syringe and needle exchange as HIV/AIDS prevention for injection drug users. JAMA. 1994;271(2):115–120. [PubMed]
29. Vlahov D, Junge B. The role of needle exchange programs in HIV prevention. Public Health Rep. 1998;113(Suppl 1):75–80. [PMC free article] [PubMed]
30. Strathdee SA, Celentano DD, Shah N, et al. Needle-exchange attendance and health care utilization promote entry into detoxification. Journal of Urban Health. 1999;76(4):448–460. [PMC free article] [PubMed]
31. Vertefeuille J, Marx MA, Tun W, et al. Decline in self-reported high-risk injection-related behaviors among HIV-seropositive participants in the Baltimore needle exchange program. AIDS and Behavior. 2000;4(4):381–388.
32. Bluthenthal RN, Kral AH, Gee L, et al. The effect of syringe exchange use on high-risk injection drug users: a cohort study. AIDS. 2000;14(5):605–611. [PubMed]
33. Latkin CA, Davey MA, Hua W. Needle exchange program utilization and entry into drug user treatment: is there a long-term connection in Baltimore, Maryland. Subst Use Misuse. 2006;41(14):1991–2001. [PubMed]
34. Huo D, Ouellet LJ. Needle exchange and injection-related risk behaviors in Chicago: a longitudinal study. J Acq Immune Defic Synd. 2007;45(1):108–114. [PubMed]
35. Davidson P, McLean RL, Kral AH, et al. Fatal heroin-related overdose in San Francisco, 1997-2000: a case for targeted intervention. Journal of Urban Health. 2003;80(2):261–73. [PMC free article] [PubMed]
36. Tobin K, Hua W, Costenbader E, et al. The association between change in social network characteristics and non-fatal overdose: Results from the SHIELD study in Baltimore, MD, USA. Drug Alcohol Depend. 2007;87(1):63–68. [PubMed]