Co-occurring patterns of gambling, alcohol and tobacco use should inform gambling cumulative impact assessments in the UK

Background: Gambling is increasingly recognised as a public health issue, often co-occurring with alcohol and tobacco use. In response, the UK government plans to introduce Cumulative Impact Assessments (CIAs) for gambling, allowing local authorities to identify areas where venue density may contribute to increased gambling harm. However, CIA models based solely on gambling prevalence may be less effective, as they ignore co-occurring risk factors that can amplify harm. This paper shows that prevalence alone is not a reliable indicator of gambling risks and that considering co-occurring behaviours in specific socio-economic contexts is essential for informing effective CIA design and gambling harm reduction policymaking. Methods: We use data from 50,741 households in the UK Living Costs and Food Survey (LCFS) to examine how gambling co-occurs with alcohol and tobacco use across 26 socio-economic groups defined using census-based characteristics. These groups provide fine-grained spatial detail, allowing councils to identify which populations appear locally and apply results for targeted interventions. We calculate odds ratios to quantify behavioural co-occurrence in each group alongside prevalence, mapping patterns of unhealthy behaviours and assessing where multiple risks are concentrated relative to overall prevalence. Results: We find that gambling associations with alcohol and tobacco use differ across socio-economic groups, with some subpopulations, such as ethnic communities and constrained flat dwellers, displaying closely linked unhealthy behaviours. In these groups, the odds of gambling alongside alcohol or tobacco use can be up to three times higher despite relatively low overall gambling prevalence. Considering co-occurring behaviours alongside prevalence identifies high-risk populations that would be overlooked if either measure were considered alone, as prevalence and co-occurrence capture different aspects of risk. Conclusions: Focusing solely on gambling prevalence would underestimate concentrated harms and overlook communities with strong co-occurring behaviours. Policymakers should incorporate associations with alcohol and tobacco use into CIAs for gambling to better target populations at greatest risk. These insights inform more effective, evidence-based interventions and frame gambling as a public health concern, while emphasising the need for integrated, socio-economically tailored strategies addressing multiple risk factors simultaneously.