International Journal of Drug Policy - 2014

Volume 25 Issue 3 May 2014

Issue link:

Contents of this Issue


Page 92 of 153

K.E. Tobin et al. / International Journal of Drug Policy 25 (2014) 591–597 593 validate if this appeared to be the location that they provided. Upon validation, the latitude and longitude of each place was entered into the database and used for subsequent mapping and analysis. Drug/alcohol use locations identified through the place section of the socio-spatial inventory were mapped. Analysis of the mapped drug/alcohol use places focused on examining spatial distribution, specifically, spatial intensity and spatial clustering of these places. Spatial intensity is defined as the expected number of events per unit area. After the places were generated, Indexes were asked 15 relationship/attribute questions, such as the type of place (residen- tial, bar/club, outside) and frequency of time spent at the place. Linking section To link social network members with specific places, for each of the social networks listed in response to the item, "who did you use drugs or alcohol with in the past 90 days?," the Index was asked where they were the last time they used drugs or alcohol. Using the procedures described in the place section, the latitude and lon- gitude coordinates of the drug/alcohol places were obtained and characteristics of the place were assessed. Index characteristics Index participants reported age, highest educational attainment and current employment status. Sexual identity was assessed with the question, "Do you consider yourself to be: heterosexual or straight; bisexual; queer, homosexual, gay, same-gender loving; or not sure/questioning?" They also reported their HIV status as negative, positive, or unknown. Analysis Spatial intensity was estimated throughout Baltimore City based on the mapped point pattern of drug/alcohol use places using the edge-corrected kernel density approach (Kelsall & Diggle, 1995a, 1995b; Waller & Gotway, 2004). The output from this analysis was a map describing the spatial concentration of drug and alcohol use places within Baltimore City (expected numbers of drug and alcohol use places per unit area). Spatial clustering was assessed using K-function analysis. The K-function is defined as the expected number of events within a given distance of an arbitrary event (scaled relative to the over- all intensity of all observed events), with event here denoting a drug and alcohol use place location (Ripley, 1976). Therefore, larger K-function values for given distances suggests a higher degree of spatial clustering (events more spatially compact). For the cur- rent analysis, the difference in the degree of clustering (assessed using difference in K-functions) for the drug/alcohol use places was examined for varying characteristics of the Index person and of the place. For example, we examined whether the degree of spatial clustering for drug/alcohol use places differed by the index person being HIV positive or negative, self-reported as being gay or not, 30 plus years of age or younger, and whether or not the place was identified as a residence or non-residence. Statistical significance was assessed using the random labeling permutation approach (Diggle, 2003). Analyses were performed in R, with contributed packages spatstat, splancs, maptools, and maps (R Development Core Team, 2008). Maps for presentation were generated in ArcGIS (Environmental Systems Research Institute, 2010). Results A total of n = 77 participants completed the socio-spatial inventory. Table 1 presents the characteristics of n = 51 Indexes who provided data on n = 187 social network members with whom they used drugs/alcohol in the past 3 months and n = 187 places/locations of last use. The majority had at least 12 years of education (82%), nearly half were working full or part time, about Table 1 Characteristics of n = 51 Index participants who drank alcohol/used drugs with social network members and provided place location. Variable N (%) Mean age (SD) 36.5 (10.9) Education ≤11 years 9 (18) 12 years/GED 19 (37) ≥some college 23 (45) Employment status Full-time 9 (18) Part-time 12 (24) Not working 18 (35) On disability 12 (24) Sexual identity Gay 26 (51) Bisexual 16 (31) Heterosexual 4 (8) Not sure/questioning/other 5 (10) Self-reported HIV positive status 14 (27) Substances used with social network at most recent use Alcohol only 51(27) Cannabis only 31 (16) Alcohol and cannabis 50 (27) Combinations of alcohol, cannabis, crack, heroin 37 (20) half identified as gay and nearly one-third self-report HIV positive status (31%). Substance use with social networks included alcohol only (27%), alcohol and cannabis (27%), cannabis only (16%), and combinations of alcohol, crack, heroin (20%). Fig. 1 shows a map of the residential locations of the n = 51 Indexes overlaid on a map of poverty levels. This map indicates that residences were located throughout the City in fairly impoverished areas. Fig. 1. Distribution of Index participant residence in Baltimore City.

Articles in this issue

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