A comparison of drug use networks across three cities

This study presented drug use network data collected in three cities in the United States and Puerto Rico. Respondents in the study were out-of- treatment intravenous drug injectors at risk for HIV infection. All respondents in the study had injected at least once during the previous 30 days. There were significant differences among the respondents related to the site where the data were collected. The racial/ethnic composition of the samples varied according to the city in which respondents were interviewed. Not surprisingly, all respondents in the Rio Piedras sample reported that they were Hispanic. About two-thirds of the respondents in the Houston and Dayton/Columbus samples were African American. In addition, respondents in the Puerto Rico sample were more likely to have been younger and to have injected heroin and cocaine mixed together. Respondents from the Houston site were more likely to have been older, considered themselves homeless, and to have injected cocaine. Given these significant differences in the samples, the number, heterogeneity, strength, and frequency of network linkages were investigated, controlling for site. The number of drug use linkages reported by study participants did not vary significantly by site. Respondents at the three sites reported having slightly over two drug use linkages or, put another way, a drug use network size of slightly over three people. However, there were significant differences in network characteristics among the sites in regard to heterogeneity, strength of linkages, and frequency of linkages. Respondents in Puerto Rico were more likely to be involved in networks that were relatively homogeneous, with relatively weak ties and frequent drug use interactions. Networks in Rio Piedras were much more likely to be all Hispanic, all male, use only injectable drugs, to have used drugs with each other a short period of time, and to use together multiple times daily. Injectors in the Dayton/Columbus and Houston samples, on the other hand, were more likely to be involved in heterogeneous networks, with comparatively stronger ties and less frequent drug use interactions. Networks in Houston and Dayton/Columbus were much more likely to be of mixed gender or race/ethnicity, to use both injectable drugs and crack cocaine, to have used drugs together a comparatively long time, and to use together once a week or less. Although these differences in network characteristics did emerge, there was no significant difference between the sites in a respondent reporting that he or she received or gave a needle to another member of the network. There was a significant difference, however, in reported sexual relationships. Respondents in Dayton/Columbus and Houston were more likely to have reported a sexual relationship with one or more network members. Some network characteristics were found to be significantly correlated. Networks with mixed genders were related mixed-age groups, with the respondent giving needles to or receiving needles from one or more network contacts and the respondent reporting having sex with one or more network contacts. However, networks with mixed genders were negatively related to the size of the network. In addition, network size was negatively related to more than one age group in a network. Networks with members of more than one racial/ethnic group were significantly related to the respondent reporting that he or she gave one or more network contacts a needle and had a sexual relationship with one or more contacts. Networks in which the respondent reported giving a needle were highly correlated with the respondent receiving a needle. Receiving a needle from a network member was correlated with reported sex with a network member.