Decision modelling

The review of effectiveness and cost-effectiveness studies identified data on the short-term behavioural effects of the interventions. Decision models were constructed to extrapolate short-term behavioural effects into lifetime health gains, and to value these gains in terms of public sector costs avoided and improvements in participant well-being. These long-term benefits were discounted at 3.5% in accordance with H.M. Treasury guidance.

It was intended that the estimate of the public sector cost implications of the interventions include both healthcare and social care costs. Reviews of the evidence, however, indicated that there was insufficient data to model the impact of short-term behavioural changes on longer-term social care costs. Consequently, the modelling only included healthcare costs.

The need to standardise estimates of long-term well-being impacts in order to compare interventions meant that the models measured quality of life impacts using QALYs. Research on alternative measures is insufficiently developed to provide the data necessary to estimate quality of life changes across the interventions.

Separate decision models were constructed for each of the seventeen interventions included in the analysis. The figure below illustrates the structure of the decision models constructed, using the example of a smoking cessation intervention. The review provided data on the probability of quitting smoking with (p=A) and without (p=B) the intervention, and the cost of the intervention (a). Supplementary searches were required to identify epidemiological data to extrapolate smoking cessation outcomes into changes in long term health states (e.g. lung cancer), and economic studies of the value of these health states (both in terms of healthcare costs and quality of life).

Example of a decision model for a smoking cessation intervention

The most appropriate effect data was selected for inclusion in each model. The effect data was selected to prioritise the following: UK-based estimates; estimates produced using high quality research designs; estimates for which corresponding cost data were also available; and estimates for which corresponding epidemiological data was available with which to model long-term benefits.

You can access the analysis from each of the models built for the interventions from the Results page

 Back to previous page

All content is copyrighted. Please do not duplicate or redistribute without our permission.

23 November 2017 10:58 Health England Leading Prioritisation vaspiraHub 2017.7760