Cost-effectiveness of changes in alcohol taxation in Denmark: a modelling study

Table 4 Sensitivity analyses: cost-effectiveness of alcohol taxation under alternative assumptions (for the Danish population aged 16+ in 2009, population: 4.5 million)

Intervention		DALYs averted ^a			Cost offsets (€ million)			ICER ^b(€/DALY)
Intervention		Mean	CI95% low	CI95% high	Mean	CI95% low	CI95% high	Mean ^c	CI95% low	CI95% high
Price elasticity^d	20% increase	43,873	35,130	52,556	-263	-332	-201	Dominant	Dominant	Dominant
	100% increase	204,291	167,407	243,095	-1,246	-1,584	-973	Dominant	Dominant	Dominant
	10% decrease	-23,209	-27,760	-18,645	137	107	173	Dominated	Dominated	Dominated
Intervention cost^e	20% increase	19,995	16,080	24,007	-119	-148	-92	Dominant	Dominant	Dominant
	100% increase	95,639	76,949	113,174	-575	-718	-440	Dominant	Dominant	Dominant
	10% decrease	-10,085	-12,107	-8,077	60	46	75	Dominated	Dominated	Dominated
Taxation pass-through rate of 1.66^f	20% increase	33,282	26,760	40,024	-197	-247	-152	Dominant	Dominant	Dominant
	100% increase	157,148	129,415	186,267	-944	-1,182	-733	Dominant	Dominant	Dominant
	10% decrease	-17,010	-20,368	-13,793	100	77	126	Dominated	Dominated	Dominated
Taxation pass-through rate of 2.57^f	20% increase	50,580	40,628	60,523	-301	-379	-229	Dominant	Dominant	Dominant
	100% increase	239,017	191,170	291,742	-1,518	-2,194	-1,104	Dominant	Dominant	Dominant
	10% decrease	-26,348	-31,770	-21,220	154	119	194	Dominated	Dominated	Dominated

^aDALY = disability-adjusted life year. ^bICER = incremental cost-effectiveness ratio. ^cCalculated as ‘ratio of means’ [46]. ^dIn the first sensitivity analysis the effects of higher price elasticities of -0.35, -0.7 and -0.7 respectively for beer, wine and spirits are tested. ^eIn the second sensitivity analysis the effects of assuming the same intervention cost for taxation as for other legislative interventions (yearly cost of €375,000 plus €270,000 the first year) are tested. ^fIn the third sensitivity analysis the effects of the assumed taxation pass-through rate of 1 is tested by applying two alternative pass-through rates from Young and Bielinska–Kwapisz (a rate of 1.66) and Kenkel (a rate of 2.57) [49, 50].

ISSN: 1478-7547