AbstractAims and objectives
Substance misuse is a chronic relapsing condition associated with high morbidity and mortality. Treatment attempts to reduce harms associated with drug use and to promote recovery and has developed considerably in the last 30 years. Opioid substitution therapy using methadone (OST-M) is an effective treatment for opioid dependency. Though the effectiveness of OST-M in delivering harm-reduction is well evidenced, evidence demonstrating recovery is limited as is understanding of those factors influencing progress. In this context, national policy makers and stakeholders have repeatedly questioned the value of OST-M as a substance misuse treatment and, at times, have sought to limit its use. Rigorous, long term outcome studies of UK subjects are required to improve clinical outcomes in OST-M subjects and to ensure ongoing availability of evidence-based treatments.
In this context, the study had two main objectives: to demonstrate that standard clinical information systems can deliver rich, valid datasets to support outcome research; to use these data to explore the relationships between a selection of baseline variables (patient characteristics, comorbid conditions, the nature of substance misuse and the treatment received), the clinical process and long term outcomes achieved in a large cohort of OST-M patients in a standard NHS treatment setting.
Methods and materials
Standard clinical information, collected over 7 years, was linked with validated data from a range of databases. A large representative sample (76% of the OST-M treatment population in a region) was described in detail. Follow-up data were retrieved from clinical casenotes (4 years) and linked datasets (4-7 years) and collated to create a database for analysis. Variables for analysis were selected following a review of the published literature. Univariate analyses were undertaken to demonstrate statistically significant associations between baseline and follow-up variables. Significant variables were then entered into multiple regression analyses to develop predictive models for selected outcomes. Any predictive models were then subjected to cross-validation to determine their predictive power in novel datasets.
Many highly significant associations were shown. Significant personal (demographic) factors included: age, gender, having children, having conflict in personal relationships, educational level achieved and being in employment. It was notable that the area lived in (of three districts) was strongly associated with a wide variation in clinical process and outcomes achieved. Whether treated in primary care or specialist services, the medical treatments received, the level of non-NHS support and patient satisfaction showed strong associations with outcome. Baseline illicit drug use was also strongly associated with outcome.
Multiple regression analyses found that despite these highly significant associations, strong predictive models of long terms outcome could not be demonstrated. Where weak models were created - predicting drug use (by self - report); drug use (positive tests); family stability - cross validation showed these had no predictive value in novel datasets.
Standard clinical information, linked with relevant NHS datasets can give rich and comprehensive data suitable for research of large representative samples over long time periods. This study represents one of the largest OST-M populations ever described in the UK with longer follow-up periods than most of the published literature.
In this study strong associations were found between a range of independent and dependent variables over 4-7 years. These findings broadly reflected the evidence base. However, the associated variables could not generate strong useful predictive models of long term outcome. This could reflect issues of study design or data quality.
This type of approach should be further developed in the field of substance misuse research. Issues of data quality would require to be addressed to maximize the value of these datasets. Further research is required to develop better understanding into key factors influencing long term outcomes of treatment in substance misuse.
|Date of Award||2013|
|Supervisor||Keith Matthews (Supervisor) & Douglas Steele (Supervisor)|