Rationale Understanding the extent of harmful illicit drug use (e.g. opioids, cocaine) is crucial for planning services, evaluating the effects of interventions and ultimately reducing harms. However, available data (population surveys, routine administrative records) are subject to potentially very large biases and long delays. Studies worldwide have demonstrated the potential of ‘wastewater-based epidemiology’: monitoring population-level substance use using concentrations of biomarkers in communal wastewater (Zuccato et al 2008). This could offer near real-time data that does not depend on self-report. Biomarker concentrations can be measured with great precision. However the accuracy of subsequent estimates of drug consumption is unclear, since these rely on many additional assumptions, e.g. about the population size, the excretion profile of the drug and stability of its metabolites. Wastewater data could be more informative for comparisons over time or across locations (Jones et al 2014).
Aims & objectives This project will investigate how measured concentrations of biomarkers in communal wastewater can best contribute to our understanding of population illicit drug use. The particular focus will be on how such data could be used to infer changes in drug use in the population, such that they could be used as an outcome for evaluating interventions. There is a strong need internationally for robust statistical models in this area. The student will primarily work in a Bayesian statistical framework.
Methods The student will: 1. Investigate recently suggested methods for estimating ‘de facto’ population size (number of people using the sewerage system) using biomarkers. Accounting for fluctuating population size could make analyses more responsive to changes in drug consumption. 2. Develop time series and/or control chart methods for detecting sudden changes, e.g. in response to an intervention. 3. Assess agreement or conflict between wastewater-based evidence on drug consumption and more traditional epidemiological measures. 4. Synthesise multiple evidence sources regarding population-level drug consumption together, in a Bayesian ‘multi-parameter evidence synthesis’ model. Wastewater-based data will be provided by our collaborator, Dr Kasprzyk-Hordern, at the Department of Chemistry, University of Bath. There is also potential for collaboration with several other teams internationally, where more detailed data sets (longer term / across more locations) are available.
When applying please select ’PhD in Social and Community Medicine’ from the Faculty of Health Sciences.