International Journal of Drug Policy - 2014

Volume 25 Issue 3 May 2014

Issue link: http://digitalreprints.elsevier.com/i/364061

Contents of this Issue

Navigation

Page 21 of 153

522 A.N. Martinez et al. / International Journal of Drug Policy 25 (2014) 516–524 Table 2 Logistic regression models of individual level characteristics and activity space variables independently associated with three health outcomes (N = 989). Adjusted odds ratios (95% CI) Adjusted odds ratios (95% CI) Adjusted odds ratios (95% CI) Syringe sharing a Overdose b HIV serostatus c Activity space variables Distance ns ns 0.84 (0.74, 0.95) Individual-level variables Race White ns ns ns Black 0.49 (0.35, 0.70) 0.35 (0.19, 0.63) ns Latino ns ns ns Other ns ns ns Under 30 years ns 3.0 (1.3, 6.7) ns Considers self homeless 1.5 (1.04, 2.5) ns 0.55 (0.35, 0.85) Traded sex for cash or drugs in past 6 months 1.8 (1.2, 2.6) – 1.7 (0.99, 2.9) Illegal source of income in past 30 days ns ns ns Smoking crack in past 6 months 1.7 (1.2, 2.5) ns – Heroin injection in past 6 months – 14.3 (1.97, 104.6) – Arrested in past 6 months 1.3 (0.95, 1.88) ns ns Residential transience 1.11 (1.01, 1.22) – – Census tract-level variables Concentrated poverty 0.67 (0.50, 0.97) ns ns a Variables not included in the multivariate models because of their non-significance include: gender, methamphetamine injection or heroin injection in past 6 months, SEP accessibility, years of injection drug use, received government assistance in past 30 days. b Variables not included in the multivariate models because of their non-significance include: gender, methamphetamine injection in past 6 months, smoking crack in the past 30 days, residential transience, SEP accessibility, years of injection drug use, received government assistance in past 30 days. c Variables not included in the multivariate models because of their non-significance include: gender, methamphetamine or heroin injection in past 6 months, smoking crack in the past 30 days, residential transience, SEP accessibility, years of injection drug use, received government assistance in past 30 days. 9% reported at least one non-fatal overdose in the past 12 months. All activity space locations reported by participants were within the City and County of San Francisco. Mean activity space distance for all participants was 1.5 miles with a standard deviation of 2.6 miles. The median distance was 0.24 miles with an interquartile range of 2.1 miles. Two percent of the sample had an activity space distance of 10 miles or more, whereas 36% of the sample had an activity space distance of 0 miles, meaning they reported the same geographic intersection for all three questions (sleeping, hanging out, using drugs). The maximum distance reported by any one participant was 16.6 miles, which can be explained by movement between San Francisco and Treasure Island, an island that is part of the City and County of San Francisco and is accessible via bus and car. Residential transience was high, with a mean and standard deviation of 2.7 (4.4) different sleeping locations reported by participants over a 6 month period. Partici- pants who slept in more than 3 different locations over a 6 month period also had higher mean activity spaces. Participants who most often slept in concentrated poverty tracts had lower mean activity space distances than participants in low poverty tracts (Table 1). Activity space routes, with a buffer of 50 feet, intersected with at least one SEP for 96 participants (9.7%). The mean activity space distance for these 96 participants is 2.3 miles, which is significantly higher than the mean distance for participants with routes that did not intersect an SEP (1.4 miles, p = 0.002). Frequency of SEP use was not higher among IDUs with activity space routes that inter- sected an SEP. Participants that used an SEP at least once a week in a 6 month period of time also had the smallest activity space distance compared to participants with less frequent usage. Thematic maps All activity space routes are depicted in Fig. 2. The longest activ- ity space routes were extended all over San Francisco. Fig. 3 shows the 96 routes that intersected with at least one of the 17 SEP loca- tions operating in San Francisco between 2004 and 2005. Statistical analysis Table 2 shows that activity space distance had an inverse relationship with HIV positivity, with a unit increase in distance decreasing the odds of being HIV positive by 19%. The odds of syringe sharing increased by 11% for every one-unit increase in residential transience, or the number of different locations where participants slept in the previous 6 months. Residing in a tract with concentrated poverty decreased the odds of syringe sharing by 20%. SEP accessibility in an activity space was not associated with any of the three health-related outcomes. Conclusion Our study focuses on activity spaces among street-based IDUs living in San Francisco, a densely populated urban city with one of the highest median individual incomes in the United States (Cen- sus 2012). Among IDUs, activity space distance is only associated with HIV seropositivity. We found that residential transience and sleeping in a concentrated poverty Census tracts is associated with the outcome of syringe sharing but in different directions, with transience increasing the odds of syringe sharing and sleeping in poor tract decreasing the odds of syringe sharing. SEP accessibility was not identified as significant for any of the three health related outcomes. Place-based measures, including activity space distance, SEP accessibility, residential transience, or concentrated poverty Census tract, do not help explain the outcome of non-fatal overdose in the past 12 months among IDUs. Our finding that activity space distance is not associated with syringe sharing differs from what has been previously reported in the literature (Hahn et al., 2008), and brings up interesting ques- tions about the unique social geography and policy climate of San Francisco as well as methodological differences in studying place. Syringe sharing has been shown to be higher among geographically mobile IDUs in Tijuana (Brouwer, Lozada, et al., 2012). For IDUs in San Francisco, activity spaces are likely shaped by topography, policies that criminalize drug use, and the distribution of social, medical, and prevention services (Galea & Vlahov, 2002; Iguchi

Articles in this issue

view archives of International Journal of Drug Policy - 2014 - Volume 25 Issue 3 May 2014