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 118 of 153

C.E. Sterk et al. / International Journal of Drug Policy 25 (2014) 616–623 619 multinomial distribution with a cumlogit link. The coefficients for perceived neighbourhood disorder were multiplied by its standard deviation prior to exponentiation in order to interpret associations with a meaningful increase in this variable. Missing data was min- imal (<1%) and case wise deletion was applied. Results Sample Descriptive statistics are displayed in Table 1. On average, respondents used crack cocaine 48.75 (sd = 32.17) days out of the past 90 days. The mean age of first use was 27.48 (sd = 8.01), mean- ing that most respondents began using in the early 1990s, the time period when crack cocaine began to dominate the local drug market. The average age of the respondent was 46.75 (sd = 8.39), and 43% of the sample was female. Respondents, on average, had completed at least 12 years of school, and the mean monthly income was $687.70 (sd = 813.48, median = 500). Nearly one half (49%) reported having a partner. Forty-seven percent of the sample reported having stable housing and less than one half (45%) had been in the neighbourhood for at least five years. Involvement in drug distribution was reported by 21% of the respondents. Approximately one-third (32%) indicated having traded sex in the last year. More than half (55%) of the sample reported using crack cocaine only in their own home, the home of friends, or the home of relatives in the past 90 days. In terms of people with whom they used crack cocaine, use with people with whom they held social or biological ties was most common (47%). Least common was using alone (14%). Crude associations An older age of initiation of crack cocaine use (IRR: 0.99, p < .01) and being older (IRR: 0.99, p < .01) were associated with fewer days of crack cocaine use (see Table 1). Conversely, higher income (IRR: 1.01, p < .01) was associated with an increase in the number of days of use. An increase of a standard deviation (5.60) of perceived neigh- bourhood disorder was associated with an 8% increase in the number of days of use (IRR: 1.08, p < .01). Crack cocaine use-related practices were associated with increases in the number of days of crack cocaine use with involvement in drug distribution showing the stronger association (IRR: 1.49, p < .001) compared to having traded sex for drugs or money (IRR: 1.38, p < .001). Using crack cocaine only at home, the home of relatives or the home of friends was associated with an decrease in number of days of use (IRR: 0.70, p < .001) compared to ever using in a (semi-)public place. Using only alone (IRR: 0.55, p < .001) or alone and with per- sonal connections only (IRR: 0.76, p < .001) were associated with decrease in number of days of use compared to ever using with acquaintances. Multivariable analysis results The results of the multivariable analysis are displayed in Table 2. Model 1 shows that perceived neighbourhood disorder (IRR: 1.06, p < .05) was a significant predictor of number of days of crack cocaine use controlling for the significant effects of age at first use (IRR: 0.99, p < .05) and income (IRR: 1.01, p < .01). Model 2 included the variables describing crack cocaine use-related practices and the social context of crack use. Involvement in drug distribution (IRR: 1.26, p < .01) and having traded sex (IRR: 1.16, p < .05) were both associated with increases in days of use. Using only at home, the home of a relative, or the home of friends was associated with a decrease in days of use (IRR: 0.82, p < .01) compared to ever using in (semi-)public places. Using alone only was associated with a decrease in days of use (IRR: 0.74, p < .05) compared to ever using with acquaintances. Using alone and with personal connections only did not differ significantly in days of using compared to ever using with acquaintances. Perceived neighbourhood disorder was no longer associated with frequency of crack cocaine use after con- trolling for the meso-system variables. Mediation results The initial regressions necessary to test for mediation assessed the associations between perceived neighbourhood disorder and the crack cocaine use-related practices as well as the associa- tions between perceived neighbourhood disorder and the social context of crack cocaine use. Among the practices of involve- ment in drug distribution and having traded sex, only having traded sex was significantly predicted by perceived neighbourhood disorder, controlling for socio-demographics (OR: 1.05, p < .05). Among the variables describing social context of use, setting and people with who use occurred, only the latter was significantly predicted by perceived neighbourhood disorder, controlling for socio-demographics (OR: 0.96, p < .01). When having traded sex was included with perceived neighbourhood disorder in a model predicting frequency of crack use, perceived neighbourhood disor- der was no longer significantly associated with the outcome while having traded sex was strongly and positively associated (IRR: 1.32, p < .001). When people with who use occurred was included with perceived neighbourhood disorder, perceived neighbourhood dis- order was no longer significantly associated with the outcome while using only alone was associated with a decrease in fre- quency of crack use (IRR: 0.59, p < .001) compared to ever using with acquaintances and using alone and with personal connections only was also associated with a decrease in frequency of crack use (IRR: 0.80, p < .01) compared to ever using with acquaintances (Table 3). Discussion Homogenizing crack cocaine users masks the heterogeneity of their use experiences (Daniulaityte, Carlson, & Siegal, 2007). Understanding how the neighbourhood environment or place and the specific context of the use experience relate to frequency of use can enhance knowledge of domains that can be addressed in efforts to reduce the role that crack cocaine might play in the lives of individuals situated in particular places and contexts. Per- ceived neighbourhood disorder was associated with an increase in frequency of use when accounting for socio-demographic charac- teristics, providing support for research that has found a direct link between neighbourhood disorder and illicit drug use (Boardman et al., 2001; Gibbons et al., 2004; Sunder et al., 2007). However, including crack cocaine use-related practices and the social context of crack cocaine use reduced the effect of perceived neighbourhood disorder to non-significance with having exchanged sex and peo- ple with who use occurred emerging as mediators of the effect of perceived neighbourhood disorder on frequency of crack cocaine use. Consistent with previous literature, greater perceived neigh- bourhood disorder was positively associated with having traded sex (Latkin et al., 2007). In contrast, greater perceived neighbour- hood disorder was negatively associated with using only alone or using with personal connections versus using with people who were not as close to the respondent, indicating that place may inhibit people from trying to mitigate the harmful associ- ated consequences of crack cocaine use that may occur in such neighbourhoods when using with others, such as violence (Ribeiro et al., 2010; Sterk-Elifson & Elifson, 1993). The inclusion of these

Articles in this issue

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