Table 1

Data extraction

Author, year, countryPopulationIntervention/comparatorEffectiveness dataMeasure(s) of benefitCost dataReported perspective/discounting/time horizonSensitivity analysisResults
Hurley and Matthews, 2008, Australia22General populationNational Tobacco Campaign, June–November 1997, graphic antismoking television advertisements/no interventionEffectiveness of NTC on prevalence estimated from survey data. Markov model used to estimate future health benefitsLYG, QALYsCost of campaign adjusted to $A10.1 in 2001 $A. Healthcare cost estimated from existing literature, Future savings estimated using Markov model (lung cancer, AMI, stroke, COPD only)Healthcare/3%/lifetimeAssume only half of reduction in smoking prevalence observed attributable to campaignPrevention of 55 000 deaths, gains of 323 000 Lys, 407 000 QALYs, healthcare cost savings $A740.6 m. Campaign remained cost saving in sensitivity analysis
Higashi et al, 2011, Vietnam23General populationHypothetical MMCs implemented over 5 years (TV, radio, newspaper, journal, internet and electronic billboards)/no interventionEffect of campaigns on uptake and cessation estimated from literature. Markov model used to estimate future health benefitsDALYsCosted using WHO CostIt programme, 2006 VND: human resources requirements, media and advocacy, overheads. Healthcare cost savings for IHD, COPD and lung cancerGovernment/3%/lifetimeICER with and without healthcare cost offsetWithout cost offset: VND 78 300 per DALY averted (95% CI 437 000 to 176 300). With cost offset: Campaign dominates
Kotz et al, 2011, UK24General populationNSD, 1 day/year since 1984—national campaign aiming to create supportive environment and highlight available help for people who want to quit. National social marketing campaign/no interventionEffect of NSD estimated from monthly survey data on quit attempts. Previously published model of cost effectiveness for smoking interventions used to estimate permanent rate of cessation and LYGLYGEstimated from NSD charity, report and financial statements—approximately £750 000, price year not statedOrganisational (NSD charity)/3.5%/lifetimeAssume that the true rate of permanent cessation attributable to NSD was only half that observedICER £82.24 per LYG (95% CI 49.7 to 231.6) for 35–44-year-olds. £114.29, £76.19 and £97.45 for age groups <35 years, 45–54 years and 55–64 years, respectively. Campaign remained cost effective in sensitivity analysis
Brown et al, 2014, UK26General populationStoptober—a 1-month national campaign in 2012 which set smokers the goal of being smoke-free for OctoberEffect of Stoptober estimated from monthly survey data on quit attempts. Previously published model of cost effectiveness for smoking interventions used to estimate permanent rate of cessation and LYGLYGKnown costs of campaign obtained from Department of Health (2012 costs)Organisational (Department of Health)/3.5%/lifetimeExamined effects of modelling different adjustments for relapseICER for total population £558 per LYG (95% CI 126 to 989). £414 for 35–44-year-olds, £607, 417 and 566 for <35-year-olds, 45–54-year-olds and 55–64-year-olds, respectively. Campaign remained cost effective in sensitivity analysis
Ratcliffe et al, 1997, Scotland16AdultsCampaign aiming to reduce smoking prevalence via TV, posters and press advertising, a telephone helpline and a booklet containing cessation advice, October 1992–October 1993/no intervention1-year cessation rate assessed by survey of helpline callers. Modelling used to estimate LYGLYGRetrospective analysis of costs of: development and maintenance, mass media advertising, telephone helpline, information booklet, costs. Costs attributable to adults only. Mass media represented two thirds of total cost. Price year not statedOrganisational/6% benefits/lifetimeVariation of campaign costs and number of helpline callersCost per discounted LYG range from £304 to £656 when parameters are varied
Villanti et al, 2012, USA25AdultsEX campaign—television and radio campaign designed to promote smoking cessation, March–September 2008/no interventionSurvey used to estimate probabilities of confirmed awareness and quit attempts among those aware and those unaware of the EX campaign. National survey data used to estimate probability of quit attempts with no intervention. Probability of successful quitting from existing literature. Number of QALYs gained per quit from existing literatureQALYsIntervention costs: Media, public relations, staff salaries. Other societal costs: smoking cessation medication, behavioural therapy. Medical treatment costs saved by quitting smoking assumed to be $0. Price year 2009Societal/3%/lifetimeVariation of model parametersBase case ICER $37 355.
Sensitivity analysis: 95% uncertainty interval $10 779–204 976 per QALY
Fishman et al, 2005, USA1918-year-olds in USAHypothetical 4-year MMC featuring regional and culturally relevant messages, using media outlets likely to reach adolescents/no interventionYears of potential life saved among cohort of 18-year-olds based on existing literatureLYGA range of assumed campaign costs based on existing literature—$0.31/head, $0.97/head, $2.35/head. Tobacco-attributable health costs from existing literature. Price year 2000Societal/3%–7%/lifetimeVarying assumptions of campaign cost and discount rateCost per year potential life saved: $528 for low-cost media campaign with 3% DR, $19 957 for highest cost campaign with 7% DR
Secker-Walker et al, 1997, USA1715–18-year-old students in four US cities4-year TV and radio MMC in addition to school smoking prevention curriculum, 1986–1989. 36 TV ads and 17 radio ads specifically designed to appeal to students at different stages of adolescence/smoking prevention curriculum onlyDifference in smoking prevalence between students in communities receiving intervention and those in comparator communities. Markov model used to estimate LYGSmokers averted, LYGCampaign development and production cost from campaign records, price year 1996. Air-time costs quoted by TV and radio stations. Costs estimated at community level and for whole of USAOrganisational/0%, 3%, 5%/lifetimeDifferent discount rates, different mass media effect sizes, different costs, halving LY lost due to smoking, variations in prevalenceCommunity level: cost per smoker averted $754 (95% CI 531 to 1296), cost per LYG at 3% DR $696 (95% CI 445 to 1269)National level: cost per smoker averted $162, cost per LYG at 3% DR $138 (95% CI 88 to 252)
Campaign remained cost effective in sensitivity analysis
Raikou and McGuire, 2008, UK2013–17-year-olds in the UKHypothetical 5-year MMC/no interventionEffect on smoking prevalence estimated from the existing literature. Markov model used to estimate QALYs gainedQALYs, LYGCampaign costs based on 10× cost of education and communication programmes used to support implementation of smoke-free legislation (price year not stated). Costs of treating smoking-related diseases from existing literature (2006 prices)Public health sector/3.5%/lifetimeVarying assumptions of size of effect of intervention, cost of intervention and background quit rateBase case: £49 per QALY gained £362 per LYG. Campaign remained cost effective in all sensitivity analyses
Holtgrave et al, 2009, USA2112–17-year-olds in USANational youth smoking prevention campaign (truth campaign), February 2000–2002. TV radio, online and print media elements, a campaign tour that followed youth music events/no interventionSmokers averted estimated in previous study. QALYs gained estimated using data from existing literatureQALYsCampaign cost data derived directly from expenditure records. Development and delivery of media elements, campaign tour, evaluation, litigation costs. Price year not stated, collected 2000–2002. Future healthcare costs saved estimated from existing literature (price year 2000)Societal/3%/lifetimeVariation of smokers averted, QALYs gained per averted smoker, treatment costs savedBase case: 178 290 QALYs gained. Cost-saving. Optimistic case: 1 050 000 QALYs, cost saving. Pessimistic case: $4302 per QALY
Stevens et al, 2002, UK18Turkish community in Camden and Islington, London, UK10 min play, poster campaign, leaflets. 1996–1997 (dates not specified)/no interventionBefore and after panel survey used to estimate effect of intervention on quitting. Estimates of 1-year quitters and LYG estimated from literature1 year quitters, LYGActual expenditure from project records—salary costs, other labour costs, non-pay costs, overheads. Price year not stated, collected 1996–1997Local authority/none/lifetimeVarying assumptions of population size, smoking population, quit rate, population smoking trend, continued abstinence, life years saved by quitting. Monte Carlo simulationStudy reports mean cost effectiveness drawn from probability distribution of possible outcomes in sensitivity analysis.
ICER £105 per LYG (95% CI £33 to 391)
ICER 825 per 1-year quitter (95% CI 300 to 3500)
  • $A, Australian dollar; AMI, acute myocardial infarction; COPD, Chronic Obstructive Pulmonary Disease; DALY, disability-adjusted life year; DR, discount rate; ICER, incremental cost effectiveness ratio; IHD, ischaemic heart disease; LY, life year; LYG, life years gained; MMC, mass media campaign; NSD, no smoking day; NTC, National Tobacco Campaign; QALY, quality-adjusted life year; VND, Vietnamese dollar.