Background: The UK has a high prevalence of opioid use; although this population is ageing. The use of opioids is associated with excess mortality and offending and so remains a priority for public health and criminal justice policy. Aims: There are two broad aims for this thesis: firstly, to quantify excess mortality and offending associated with opioid use, and secondly, to assess the effectiveness of drug treatment at reducing these harms.Methods: Cohorts were extracted from the Drug Data Warehouse (DDW); a collection of case-linked drug treatment and criminal justice datasets, linked to mortality and offending records. Excess mortality was quantified by comparing deaths observed in a cohort of opioid users to that expected from the general population, matched on age and gender. The association between opioid use and offending was quantified using a cohort of drug-tested offenders, comparing those who tested positive for opiates and/or cocaine with those who tested negative. The effectiveness of treatment was calculated using two separate measures: the effect of being in treatment, using a time-dependent treatment exposure, and the effect of initiation to treatment. In the latter analysis, the approach was to use observational data to emulate randomised controlled trials, in an attempt to better establish the causal effect of treatment initiation. Confounding bias, when treatment exposure is time-dependent, was discussed, using path diagrams and simulations.Results: The opioid-user cohort extracted from the DDW is the largest assembled to-date (N = 198,247). Controlling for age and gender, opioid use was associated with considerably higher mortality and offending than non-users. Older age was a risk factor for drug related poisoning (DRP) death. The association with offending was considerably higher for females. Compared to periods out of treatment, the risk of a fatal DRP was lower during pharmacological treatment but not during periods where the user received psychological support only. Simulations illustrated that when treatment exposure is time-dependent, analyses may be biased in the presence of confounding by a time-dependent variable. Among opioid-using offenders, there was little support for the hypothesis that initiation to drug treatment was effective at reducing the risk of future acquisitive offending. Conclusion: In England and Wales, there remain considerable excess mortality and offending associated with opioid use. Age and gender have an important influence on these relationships. Treatment (as delivered in England and Wales) appears effective at reducing the risk of a drug-related death, provided there is a pharmacological component, but not at reducing the risk of future offending.