Determinants of Property Crime Among Illicit Opiate Users Outside of Treatment Across Canada

Abstract

Criminal activities account for a major proportion of the social costs related to illicit drug use. This article examines the factors contributing to property crime activity among a community sample of 653 untreated regular illicit opiate users in 5 Canadian cities (OPICAN study). Multivariate analyses showed the frequency of heroin, cocaine, and crack use, gender, housing status, and past criminal justice involvement as predictors of property crime. Furthermore, crack use had a significantly different impact on property crime depending on housing status and city. These findings underline the need for targeted intervention efforts toward a reduced crime burden.

The authors acknowledge funding support from the Canadian Institutes of Health Research (CIHR) and thank the study participants, staff, and other OPICAN investigators who made this study possible. Dr. Manzoni acknowledges post-doctoral funding from the Swiss National Science Foundation.

Notes

¹An interaction effect means that the influence of one variable on another depends on the value of a third variable. However, interaction effects do not suggest causality (for which longitudinal data would be required) nor reciprocal influences of these relations.

¹Transformation with natural log: x = ln(x + 1).

²R: reference category. n = 653.

²In our models we did not include the age of our participants, who on average were 35 years old. Because this average was much above the typical crime peak in late adolescence or early adulthood (Greenberg 1985), there was no reason to include age. However, we tested the models also including age, but it was found not to be significantly related with property crime.

Percentages do not sum up to 100% because several income sources could be indicated. n = 653.

(a): logged frequencies. n = 653. Model-χ² = 82.960 (df = 21; p < .001); Hosmer and Lemeshow Test: χ² = 7.533 (df = 8, p = 0.480); Nagelkerke R² = 0.201.

³Overall, our model fits the data fairly well. Nagelkerke's R², an analog to the R² in OLS regression, has a value of .201, which means that property crime is moderately explained by the included variables.

⁴The odds ratios of these variables reflect changes per unit of the natural log scale. For example, an increase in heroin use from once a day (ln of 30 times = 3.40) to twice a day (ln 60 times = 4.09) equals an increase in the odds of involvement in property crimes of 22% (the ln-difference of .69 equals an odds increase of 22%, given the 32% increase of one unit heroin use, as indicated by the respective odds ratio). The same increase of crack and cocaine use result in increased property crime odds of 66% and 9%, respectively.

⁵However, respective interactions tested in the sub-samples of Edmonton, Montreal, and Toronto were not significant (p > .05). Quebec City had too few crack users to be examined.

⁶We found in Vancouver that, compared to permanently housed drug users, non-permanently housed users were significantly more engaged in sex work and used less prescription opioids, but were not significantly different regarding the use of other substances or the rate of drug dealing and property crime. Hence, we assumed that a larger engagement in sex work of non-permanently housed users is related to the different impact of crack use on property crime in Vancouver. This assumption was supported by two additional results. First, crack use was significantly and positively correlated with sex work (r = 0.29; p < .01). Secondly, a cross-tabulation of property crime and sex work separated by groups of high and low crack use (split by the median) showed for the high-use crack group that a significantly higher proportion of those not involved in sex work committed a property crime (10 out of 43; 19%) compared to those involved in sex work (1 out of 40; 2%). No significant association between engagement in sex work and property crime was found for the low-use crack group.

⁷For our analysis we considered the prevalence rate of these activities. Substitutional associations may exist if the actual funds generated from drug dealing or sex work were considered.