Projects

Introduction

While many high-risk drinkers express a desire to reduce their alcohol consumption, achieving meaningful reductions remains challenging. This project addresses this challenge by investigating how contextual factors influence efforts to reduce alcohol consumption, and by testing policies designed to modify those contexts.

Methodology

This project will integrate two complementary computational techniques never before combined in public health research:

• Drift-diffusion modelling (DDM): This computational technique sheds light on the cognitive processes underlying an individual's momentary choices.
• Agent-based modelling (ABM): This technique examines how individual-level behaviours and interactions result in population-level trends.

By combining these models, we can simulate how environmental changes influence individual decision-making in real-time, and how these repeated individual decisions lead to long-term patterns of behaviour across a population.

The model will translate ideas from behavioural economics into computer code, and will be grounded in real-world data. We will begin with a large empirical study where nationally representative participants from England repeatedly make decisions about alcohol and other alternatives while they are in different contexts. Fitting the DDM to this behavioural data enables estimation of how these decisions unfold in real time.

These estimates will feed into the model, which is then calibrated against nearly two decades of drinking trend data from the Health Survey for England (2005–2024). Other large datasets, such as Understanding Society and Kantar, will help us understand the wider context – for example, whether people have opportunities to take part in alcohol-free activities and hobbies, and the different situations where drinking tends to happen.

The role of context and policy

The project will look at how different contexts shape behaviour. Examples of these contexts include:

• Location: For example, whether a person is at home, in a pub, or in a park.
• Company: For example, whether they are alone, with a partner, or with friends.

Using simulation techniques, the model will be used to appraise novel alcohol policies that alter these contexts. Proposed policies include increasing availability of, and access to, alcohol-free alternatives and reducing exposure to alcohol advertising.

A key part of the project will involve collaborating with stakeholders from the NHS, government, and people with lived and living experience of addiction to determine the most relevant policies to appraise.

Policy implications and health inequalities

The findings will provide a powerful tool for policymakers by demonstrating how different policy interventions might influence alcohol reduction attempts at both the individual level and across the population.

Another core strength of the project is that, by using demographic data, the modelled population can be stratified by factors such as socioeconomic status to investigate potential health inequalities, ensuring that policies are designed to be effective and equitable.

While this project focuses on alcohol, the methodological tools developed will be adaptable to other public health challenges, including diet, vaping and gambling.

Work Packages

The project is split into four Work Packages (WP):

• WP1: Understand, with stakeholder input, what are the critical contextual factors that drive real-time alcohol consumption decisions, and how can we leverage this knowledge to design targeted, evidence-based policy interventions.
• WP2: Synthesise multiple datasets to statistically inform model relationships and develop a synthetic population that can be used in constructing the ABM.
• WP3: Construct a new ABM (containing DDM processes) and calibrate and validate this model to explain alcohol consumption trends in England (years 2005–2024).
• WP4: Appraise how hypothetical alcohol policies affect real-time decision-making and long-term patterns of alcohol consumption across different subgroups in England.