Projects & Research Partnerships

Introduction

Problem gambling is a significant public health issue with far-reaching negative consequences. While effective treatments such as cognitive behavioural therapy (CBT) are available through the NHS, ensuring individuals access and remain engaged in gambling treatment is a persistent challenge. High rates of dropout not only hinder individual recovery but also represent an inefficient use of healthcare resources.

This project will use machine learning – a method where computers learn from data to identify patterns and make predictions without explicit programming – to pinpoint the factors that predict engagement and retention within NHS gambling treatment pathways, paving the way for more effective and personalised interventions.

Background

Engaging and retaining people in gambling treatment is a well-known problem. Research indicates that, on average, around 39% of individuals drop out of multi-session, face-to-face psychological interventions for gambling disorder.

The NHS Northern Gambling Service, a provider of specialist gambling treatment and partner in this project, experiences this challenge directly. Approximately 40% of people referred to the service failed to attend their initial assessment, and of those who completed the initial assessment and were due to start group therapy, around 50% disengaged from the service before completing treatment.

These statistics highlight a critical need to understand why individuals do not engage with or drop out of gambling treatment. Existing research in this area has often used simple analytical methods and may not reflect the current treatment landscape, which now includes more online options.

This project seeks to address these limitations by employing more sophisticated techniques to gain a deeper understanding of engagement in gambling treatment.

Project studies

The project will consist of two studies:

Study 1: Predicting dropout after initial assessment

This study will analyse anonymised records from over 2500 individuals referred to the NHS Northern Gambling Service. Using advanced statistical and machine learning methods, we will explore a range of factors including demographic characteristics, gambling behaviour, the severity of gambling disorder, co-occurring mental health conditions, and motivation for change, to identify predictors of both retention in treatment and dropout at different stages. The goal is to develop a robust understanding of what influences a service user's journey through gambling treatment once they have attended an initial assessment.

Study 2: Predicting dropout before initial assessment

This study will focus on the significant proportion of individuals who are referred to NHS gambling treatment services but do not attend their initial appointment. By applying machine learning-driven content analysis to the information contained within their referral records, we aim to gain novel insights into the reasons behind this early disengagement. This will involve analysing free-text descriptions of their gambling and health issues, alongside demographic and referral source information, to identify key factors associated with not attending the first step of gambling treatment.

Future research

The findings from this project will provide a crucial foundation for future research aimed at developing and testing interventions to improve engagement and retention in gambling treatment.

Our longer-term goal is to identify modifiable risk factors – things we can actually change – that can be targeted through brief interventions or form the basis of personalised treatment pathways. By understanding distinct subgroups of individuals with gambling disorder and their patterns of engagement, we aim to explore whether these groups might benefit from different types or intensities of gambling treatment.

Ultimately, this programme of research seeks to develop evidence-based strategies to match people to the most appropriate gambling treatment, thereby increasing retention and improving outcomes for those seeking help.