Development of a comparative cost-effectiveness appraisal instrument for local alcohol policies: a modelling study

Abstract

BackgroundAlcohol-related ill health is estimated to cost the UK National Health Service £3 billion per year, and there were over 1 million alcohol-related admissions to hospital in England (primary or secondary diagnosis) in 2010–11. Although liver disease mortality (for which the principle cause is alcohol consumption) has been falling in other northern European countries, rates in the UK continue to rise. Correspondingly, alcohol is now prominent on the UK policy agenda, and a range of new control and prevention measures have been proposed, including new powers for local authorities, as part of a focus on localism in English public health. Although many local alcohol policy interventions are supported by published evidence, there has been little exploration locally of the contextual factors that determine the optimum mix of policies for a particular area. As part of the National Institute for Health Research School for Public Health Research, we aim to localise the Sheffield alcohol policy model (SAPM) to enable appraisals of policy interventions (particularly pricing, licensing, and protocols for screening and brief intervention policies) that are specific to local population characteristics, alcohol consumption patterns, rates of harm, and existing policy frameworks. MethodsA key challenge will be obtaining estimates of key model indices from local authorities. Baseline alcohol consumption for local authorities will be estimated by pooling multiple national datasets (eg, the General Lifestyle Survey) to create sufficiently large samples for local authorities (or clusters of authorities) and by identifying local consumption surveys. Estimates of alcohol consumption derived from different survey years will need to be rebased to the latest available year before pooling with age-period cohort models that infer projected trends in alcohol consumption. Data might also need to be adjusted to account for differences in survey method. Estimates will be cross-validated with the local alcohol profiles for England (LAPE) produced by the North West Public Health Observatory. Data-pooling methods will also be used to explore whether local estimates of alcohol prices and spending, grouped by age, sex, and ethnic origin, can be derived. For alcohol availability, spatial analyses of full postcode data for all English alcohol outlets in 2003, 2007, 2010, and 2013 will be used to characterise local authorities by availability measures. After cross-validation with established local datasets on health harms and other societal effects by regression techniques, the predictive capacity of the locally adapted SAPM (SAPM-LA) to estimate localised harms corresponding to different local consumptions will be tested. When validated, SAPM-LA should help to facilitate the comparative appraisal and optimisation of a wide range of policies aimed at local reduction of alcohol harms from licensing and promotion restrictions through to targeted brief intervention strategies. Any reduction in harms will therefore be modelled on the projected reductions in consumption resulting from licensing and planning approaches to restricting alcohol supply, as well as more overtly behavioural approaches such as screening and brief interventions. Although LAPE identifies local authorities with high rates of alcohol harms (eg, Bradford and Rotherham in the Yorkshire and Humber region), most have little additional detail on accompanying local consumption patterns. A few local authorities with which we already have strong working links have undertaken dedicated surveys locally, which will provide a comparison for our baseline estimates. InterpretationThe work outlined here is a methods-only presentation and is intended to stimulate discussion and interest with front-line public health practitioners within local authorities to enable discussion of the types of intervention mixes that will be the most useful, feasible, and practicable. The locally adapted model could then be used to build the evidence case for implementation, as has been most persuasive at both national and devolved government levels. FundingMulticentre National Institute for Health Research School of Public Health Research.