The identification of subtypes among injection drug users: HIV and hepatitis C differences as indicated with the theory of planned behaviour

Abstract

Recent studies have reported differences in terms of high-risk drug using and sexual behaviour between younger and older injection drug users (IDUs), however, no studies have looked at the clustering pattern across IDUs using the theory of planned behaviour (TPB). The TPB is useful in the prediction of behavioural outcomes. This analysis examines the interplay of TPB components and needle sharing among a sample of IDUs from Montreal, Canada. The study includes 109 eligible IDUs, recruited via respondent driven sampling. Hierarchical cluster analysis was used to identify a typology of the respondents based on the responses to TPB measure. The results indicate that, among all of the measures, descriptive and injunctive norms, HIV/hepatitis C (HCV) status, and a measure of direct control were significantly different across the subtypes. Furthermore, it was possible to identify patterns in TPB measures across the subgroups. Specifically, the two TPB measures increased from the entrenched (Cluster 1), to the chronics (Cluster 4), to the initiators (Cluster 3), with the highest reported average within the survivor group (Cluster 2). Correspondingly, the proportion of the group reporting HIV/HCV status decreased in the opposite direction. Consequently drug use and HIV/HCV treatment and prevention programs need to concentrate intervention differently for each of these heterogeneous drug user populations.

• Addiction
• HIV
• injecting

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

Notice of Correction:

Changes have been made to Table 3 since the original online publication date of December 16, 2013.

Notes

¹Respondent-driven sampling is based on the mathematical modeling of the theory of Markov chains (Malekinejad et al., 2011).

²Squared Euclidean distance was selected as it progressively greater weighting for values that are further apart. Although Euclidean distance (and squared Euclidean distance) is thought to be problematic for variables of different scales that was not the case here.

³This method is distinct from all other methods because it uses an analysis of variance approach to evaluate the distances between clusters. In short, this method attempts to minimize the Sum of Squares (SS) of any two (hypothetical) clusters that can be formed at each step.

⁴Although sex was excluded as a demographic measure due to the low number of females represented within the sample (∼11%), of those females the majority (∼67%) we included within the third cluster.