Using theories of practice to develop event-level alcohol epidemiology and policy analysis: Studying context, consumption and harm

Background
Alcohol epidemiology and policy analysis typically treat drinking as a single behaviour, rather than considering the many different ways that people drink. This thesis applies a novel quantitative occasion- and practice-based approach to studying the varied relationships between contexts of drinking occasions (such as timing or drink type), consumption and alcohol-related harm.

Methods
Firstly, a mapping review explores the dominant methodological approaches of, and research gaps in, the existing event-level literature on drinking contexts and alcohol consumption/acute harm. Then, a systematic review synthesises the findings of studies linking contexts directly to acute harms. The third study identifies combinations of contexts that are associated with heavy consumption in adults’ drinking occasions using decision-tree modelling. The fourth study uses time series methods to test hypothesised effects of the UK Licensing Act 2003 on the contextual characteristics of drinking occasions.

Results
There is a large and heterogeneous literature on drinking contexts, but this is largely conducted in the United States with young adult participants, which limits generalisability of the findings. Furthermore, few papers consider a broad set of contexts.

Drinking contexts are directly linked to acute harms, particularly drinking at the weekend, in licensed premises, and alongside illicit drug use. Contexts are also strong predictors of consumption - both individually and in combination - particularly long occasions, drinking spirits as doubles and drinking wine.

The Licensing Act 2003 had only small effects on the timing of drinking occasions, which may explain the surprising lack of substantial impacts on alcohol harms in previous evaluations.

Conclusions and recommendations
This thesis has used a practice-based approach to identify risky drinking contexts that are strongly associated with alcohol consumption for future research and prevention efforts. It has also highlighted the importance of considering combinations of contexts, and direct effects on acute harms.