Opioid use disorder (OUD) is a debilitating, chronic psychiatric disorder. Opioid substitution treatment (OST) is associated with the suppression of illicit drug use, criminal offending and improved health. Efficacy trials estimate the response to short interventions and follow-ups but may not reflect outcomes from routinely delivered OST. National administrative databases circumvent this limitation. This thesis addresses the impact of publicly funded treatment for OUD using the English National Drug Treatment Monitoring System (NDTMS).
The thesis has three aims: to identify sub-populations of OUD patients at treatment admission and estimate their likelihood of completing treatment successfully over the next five years; to identify longitudinal sub-populations among those continuously enrolled in OST for five years based on heroin use and estimate their likelihood of completing treatment successfully; to identify longitudinal sub-populations based on alcohol and other drug use and estimate whether membership modifies the likelihood of completing treatment successfully.
Design and method
National, five-year, prospective, observational cohort studies of community specialist treatment for OUD by data linkage with NDTMS and fatal drug-related poisoning data from the Office for National Statistics. The population for Study 1 was adults who entered treatment between 1 April 2008 and 31 March 2009 (n=54,347). The cohort for Study 2 and Study 3 was a subset of Study 1 who were continuously enrolled in treatment for five years (n=7,719). An objective summative outcome measure was used, comprised of Successful Completion of treatment (clinician-verified) with No Representation (SCNR) to treatment in the following six months. Analysis was by Latent Class Analysis, Multilevel Latent Class Growth Analysis, Multinomial Logistic Regression and Multilevel Logistic Regression.
One fifth of discharged patients in Study 1 achieved remission from OUD. Of the four sub-populations of heroin users identified from their co-presenting substance use disorder at admission, heroin and crack cocaine users were less likely to recover. Those continuously in treatment for over two years were more likely to recover. Five longitudinal sub-populations (trajectories) of heroin users were identified for Study 2. Patients whose heroin use tended towards abstinence were more likely to recover in Year 6 and Year 7. Adjunctive psychosocial intervention during OST increased the likelihood of remission. Study 3 identified between three and five trajectory classes for crack cocaine, alcohol, cannabis and unspecified drug use. The continued high-level heroin use trajectory class during OST was associated with high-level crack cocaine and alcohol use and increasing unspecified drug use. Increasing use of crack cocaine was associated with a decreased likelihood that long-term OST was completed successfully.
This thesis presents the first national long-term evaluations of the impact of publicly funded treatment for OUD in England. As relapse requiring treatment is relatively common in the first six months following treatment completion, these studies highlight the importance of incorporating re-presentation to treatment in the assessment of remission. Sub-population analysis is an important facet in evaluating patients’ progress through treatment, produces novel insights into the effectiveness of routinely delivered treatment and offers critical policy and clinical implications.
|Date of Award
|1 Dec 2019
|John Marsden (Supervisor) & John Strang (Supervisor)