Incorporating external effects in economic evaluation: The case of smoking

Abstract

The aim of this thesis is to explore methods to incorporate external effects on decision making of public health programmes in a UK setting, using smoking cessation as an example. The National Institute for Health and Clinical Excellence (NICE) methodological guidance for evaluating public health programmes is missing the incorporation of external effects. Therefore there is a need for considering their incorporation in such evaluations and to assess what are the appropiate methods to do so. Smoking cessation is an example where epidemiological evidence of external effects exists but has not generally been incorporated into economic evaluation.

This thesis therefore focused in measuring the impact, in terms of costs and QALYs lost, of the incorporation of passive smoking, smoking during pregnancy and transmission of smoking behavior into economic evaluation of smoking cessation programmes previously developed to inform policy. A static Markov model is used to incorporate passive smoking and smoking during pregnancy, whereas transmisison of smoking behaviour is incorporated through a dynamic model.

The findings show that some external effects can be incorporated without a system dynamic model, when this does occur, a static Markov model may be used to account for external effects in economic evaluation. Sometimes, to incorporate external effects, the model needs a change of population. Because smoking cessation interventions are generally highly cost-effective, the incorporation of external effects does not appear to change policy decisions, but there is a clear impact on the magnitude of the ICER. Passive smoking and smoking during pregnancy have higher impact in terms of costs and QALYs lost than transmission of smoking behaviour. Our discussion considers the validity of the methods used; how much the decision making process would be affected considering or not external effects on economic evaluation of smoking cessation interventions; and other valuation approaches for external effects, such as contingent valuation.