LAUSR

Background Inequalities in the impact of public health policies can be introduced at many stages of implementation, from need to effectiveness, and the net effect of any policy on health inequalities can be a combination of many, potentially smaller, inequalities. The extent to which these different inequalities can drive both the population level impact and the distributional impacts of policies is not well understood. In this study we use an existing health economic model of alcohol brief interventions (BIs) to explore how socioeconomic inequalities in different model inputs can affect conclusions about value for money and health inequality and contrast these with similar results for a smoking cessation model. Methods BI policies were assessed using the Sheffield Alcohol Policy Model. Previous analysis has shown a national BI programme for alcohol to be both cost-effective and inequality-improving. We examined whether these conclusions changed under three scenarios: i) individually excluding socioeconomic gradients in each model input, ii) raising levels of uptake to those in the ‘best’ group, iii) using different baseline populations. Impacts on total population health and health inequality were assessed using incremental population Net Health Benefit (NHB) and incremental ‘Equally Distributed Equivalent’ (EDE) health respectively. Results are compared with those from similar analyses undertaken using a smoking cessation model. Results A national BI programme improved both health (+43,016 QALYs) and EDE (+50,792 QALYs), reducing health inequalities. Excluding gradients in model inputs had generally small effects on NHB (+0% to +10.4%) but a larger effect on EDE (-7.9% to +15.7%), although not enough to change the conclusion that the policy is inequality reducing. Increasing delivery to the ‘best’ level would increase EDE to a greater extent than NHB (+51.6% and +43.5% respectively), further reducing inequalities. Conclusion Unlike smoking cessation programmes, BIs are likely to be both cost-effective and reduce inequalities. Considering potential inequalities across all stages of intervention delivery is important when considering the impact of policies on health inequalities, even if it may not substantially affect decisions based solely on cost-effectiveness. The relative importance of socioeconomic gradients in different stages is likely to vary between risk factors and settings.