Research article
Cigarette Tax Increase and Media Campaign: Cost of Reducing Smoking-Related Deaths

https://doi.org/10.1016/j.amepre.2005.03.004Get rights and content

Background

Tobacco use results in 500,000 premature deaths annually. Most smokers begin using tobacco before age 21, so the greatest impact on preventing smoking-related mortality is likely to come from campaigns targeting youths. This study estimates the cost-effectiveness of an anti-smoking media campaign and $1 per pack increase in cigarette taxes on the lifetime decrease in smoking-attributable mortality among the cohort of all 18-year-olds in the United States during the year 2000.

Methods

Cost-effectiveness analysis conducted from a societal perspective.

Results

The combined effects of a media campaign and $1 per pack tax increase will result in a societal savings of between $590,000 per life-year saved, at a 3% discount rate and $1.4 million per life year saved, at a 7% discount rate.

Conclusions

A media campaign and $1 per pack cigarette tax increase will reduce overall smoking prevalence, significantly decrease smoking-attributable mortality, and decrease net societal costs.

Introduction

Tobacco use remains the greatest source of preventable morbidity and mortality in the United States, resulting in 500,000 premature deaths annually.1, 2 Annual costs in the United States resulting from medical expenses and lost productivity associated with tobacco-related disease, disability, and death are estimated at more than $150 billion.3, 4, 5, 6

Most smokers begin using tobacco as youths,7 so the most effective way to decrease long-term smoking rates is to reduce the likelihood that adolescents and young adults start smoking. School-based programs have short-term benefits that rapidly decay and have not been effective in reducing long-term smoking prevalence,8, 9, 10, 11, 12, 13, 14, 15 but media campaigns have been shown to reduce smoking prevalence rates among teens and young adults.16, 17 High tobacco price sensitivity among adolescents makes cigarette tax increases another effective way of deterring cigarette use.18, 19, 20

In previous research,21 we evaluated the impact of interventions designed to reduce smoking prevalence among teens and young adults, and found that the most effective way to decrease smoking-attributable mortality is through a media campaign accompanied by a tax increase on cigarettes. The combined effects of a $1 per pack tax and media campaign would reduce smoking-attributable mortality by 168,539 lives, and 2.4 million years of potential life lost by the time the cohort reached age 85.

Tobacco control is an efficient way of improving population health, but public dollars for state and local programs in the United States are increasingly scarce.22 Better evidence of the health and economic value of tobacco control may support continued funding of these efforts. To contribute to this evidence base, this study estimates the cost-effectiveness of a media program and cigarette tax increase to reduce smoking-attributable mortality in the U.S. cohort of 18-year-olds alive in 2000 that builds on prior research.21 Research reviews suggest that this is the first study to provide estimates of the cost associated with decreased smoking-attributable mortality achieved through both a media-based intervention and tax increases from a long-term social perspective.

This study provides estimates of the long-term costs of an antismoking media campaign and $1 per pack tax increase on cigarettes to achieve decreased smoking-attributable years of potential life lost (YPLL) among all 18-year-olds in the United States in the year 2000 from a societal perspective. The cohort has been exposed to a media campaign similar to those used in several U.S. states,23, 24 Canada,24, 25 and the Netherlands26 that is assumed to run for 4 years, and the national effort is composed of a series of local programs tailored to reflect regional and demographic differences. At the end of this 4-year period, a $1-per-pack national tax on cigarettes is imposed.

Although a media campaign and cigarette tax increase will have an impact on decisions for smokers and potential smokers of all ages, this analysis focuses on a specific cohort of 18 year olds to establish the impact of effective prevention strategies on a defined group followed over its lifetimes. To evaluate the cost effectiveness of these interventions, equation 1 is estimated: Costeffectiveness=t=067[Δ̣Costst/ΔSmokingAttributableYPLLt]/(1+r)twhere Δ̣Costs are the change in costs associated with the interventions in year t, and ΔSmokingAttributable YPLL is the decrease in smoking-attributable YPLL due to the media campaign and $1-per-pack tax increase over the 67-year period between ages 18 and 85 for the target cohort. The intervention’s impact on YPLL among cohort members is based on previous research.21 Costs and life-years saved are discounted by a factor “r” reflecting the social rate of time preference, as future costs and benefits are valued less than those that occur in the short run. The sensitivity of the results was tested by varying the discount factor for each value between 3% and 7% as recommended by several leading guides to the economic evaluation of health programs.27, 28

Net costs associated with the media campaign and tax include the development and dissemination of the media program, and changes in healthcare costs among cohort members and their families due to the reduction in smoking among members of the target cohort. Also measured is the change in public-sector revenue resulting from the $1-per-pack increase in cigarette tax to assess the net cost of the interventions. Data analysis occurred in 2004.

A national media campaign requires regional and culturally relevant messages, as well as access to media outlets likely to reach adolescents. Media efforts have a shelf life,29 so it was assumed that the message must be redesigned each year. Therefore, a 4-year campaign requires costs equal to four 1-year campaigns, even though some start-up costs would likely be shared across the years of the campaign.

Several studies report the costs of developing and transmitting antismoking media campaigns that range at the low end from $0.31 to $0.56 per person for initiatives in California,23 Vermont,30 Minnesota,23 and Canada24 to more expensive programs of $2.10 to $2.35 in Arizona,23 and Massachusetts.23 The relatively expensive Arizona campaign was the only one specifically targeted to youth. The cost effectiveness of a media program on reducing smoking-related mortality was estimated using three different per capita costs that represent the lowest ($0.31), median ($0.97), and highest ($2.35) values reported in the literature. The Centers for Disease Control and Prevention (CDC) recommends an antismoking program estimated to cost $3.17 per person,31 but this program is not included because it comprises more than a media program and has never been tested.

Decreased smoking prevalence will affect lifetime health services costs among cohort members. The effect varies over time because healthcare utilization and the impact that smoking has on utilization are a function of age and life expectancy. Therefore, estimates of net direct health-care costs due to reduced smoking prevalence among cohort members over their lifetimes are given by equation 2: t=067[ΔSmokingPrevt×ΔCostt]/(1+r)t where ΔSmokingPrevt is the change in smoking prevalence for cohort members in year t and ΔCostt is the difference in mean costs for cohort members in year t attributable to decreased smoking prevalence.

The scientific literature and data from U.S. federal agencies provide information on mean healthcare costs by age32, 33, 34 and the expected impact that smoking has on direct healthcare costs.3, 4, 5, 6, 35, 36, 37, 38, 39 However, few studies provide information on both of these elements for the entire population from age 18 to 85 using the same data source. Using different data sources for mean healthcare costs and the impact of smoking on health costs for different age groups may risk discontinuities in cost estimates, so these analyses are based on two sources that provide mean costs for the entire population within a single data set for 10-year age groups.32, 33 The sensitivity of the results is evaluated by varying mean costs by age as reported in Table 1.

Smoking increases average healthcare costs from 6.3%6 to 18%,36 but published values are not age specific. Smoking-related morbidity is relatively low among young adults,3 so smoking-attributable healthcare costs likely grow at an increasing rate as smokers age. Because there are no published age-specific estimates of the impact of smoking on direct healthcare costs, the sensitivity of the results were tested by varying the expected increase in age-specific smoking-attributable costs, based on ranges reported in the literature.6, 35, 36, 37, 38, 39

The interventions will reduce neonate complications linked to risk of low-birth-weight babies among women who smoke. Neonate complications attributable to smoking are only relevant for the children of pregnant female smokers, so the cost of neonate complications was estimated for the 22-year-period until the cohort reaches age 40 and fertility rates decrease dramatically. Fertility decreases with age,40 but the cost of smoking-attributable neonatal complications increases among older women.41 Therefore, using published fertility data, the impact of smoking-attributable complications was estimated separately for children born to women between the ages of 18 and 35, and ages 36 and 40 through equation 3: i=12t=039[ΔSmokingPrevWoment×FertRatei,t×NeoNateCosti,t]/(1+r)t where ΔSmokingPrevWoment is the change in smoking prevalence among women in the cohort in year t; FertRatei,t is the expected number of births per year for women in year t; NeoNateCosti,t is the mean difference in neonate costs for each child born to smokers relative to nonsmokers. Cost and fertility values are calculated for two periods (women aged 18 to 35 and 36 to 40) as noted by the subscript i.

Model parameters for cost estimates of neonate complications among children born to cohort members come from several sources. The expected number of children born to women in the cohort is determined by overall fertility rates,43 and the impact that smoking has on fecundity is based on evidence from the scientific literature.44 The costs of neonate complications attributed to smoking are based on the CDC’s Smoking-Attributable Mortality, Morbidity, and Economic Costs (SAMMEC)41 model, which reports that smoking increases direct healthcare costs between $96 and $118 per child in 1996 U.S. dollars, which were inflated to 2000 U.S. dollars using the Medical Care Consumer Price Index (MCPI).44 Because fertility and smoking-attributable neonate complications both vary by age, the cost impact of neonate complications was evaluated for women in two age ranges: 18 to 35 and 36 to 40.

Respiratory diseases experienced by children born to smokers also affect healthcare costs.45, 46, 47, 48 The change in costs associated with decreased childhood illness related to parents’ smoking was estimated through equation 4: t=033i=12[ΔSmokingPrevt×Childreni,t×ΔCostRespComplicationsi,t]/(1+r)t where ΔSmokingPrevt is the change in smoking prevalence within the cohort at time t; Childreni,t are the number of children in households with at least one adult smoker at time t and CostRespComplicationsi,t are healthcare costs attributable to smoking-related childhood respiratory illnesses among children living in a household with an adult smoker at time t. Costs are estimated for 33 years to cover only the time in which smoking has been demonstrated to have an impact on the costs of childhood respiratory illness, from birth to age 5, which includes all children born to cohort members through age 40.49

The number of children born to families with at least one smoker is estimated from overall fertility data described above, adjusted to account for a child being exposed to only a father who smokes. Stoddard and Gray47 report environmental tobacco smoke (ETS) costs among children with parents who smoke ranging from $141 to $205 per year per child in 1995 U.S. dollars, which were inflated to 2000 U.S. dollars using the MCPI.44 However, reliable published values are available only for children aged ≤5, so estimates of ETS costs among children of cohort members include the period until the cohort reaches age 45, based on the assumption that fertility decreases after age 40.

The change in public-sector revenue due to the $1 per pack tax derives from reduced lifetime demand for cigarettes among the target cohort and the change in cigarette consumption among adults resulting from increased cigarette prices. The change in tax revenue resulting from decreased cigarette consumption among members of the target cohort due to the tax is estimated through equation 5: t=066[ΔSmokingPrevt×(PerPackTax×Packs18)]/(1+r)t where ΔSmokingPrevt is the change in smoking prevalence attributable to the interventions in year t; PerPackTax is the average U.S. per-pack tax rate during 2000; Packs18 is the mean annual number of cigarette packs consumed by the cohort; and r is the discount rate.

The change in tax revenue resulting from decreased cigarette consumption among adults is given by equation 6: i=151t=067[ΔPacks/Yeari,t×$1]/(1+r)t where ΔPacks/Yeari,t is the state-specific change in average annual packs of cigarettes consumed by adult smokers in year t; and r is the discount factor. ΔPacks/Yeari,t is determined by applying estimates of the price elasticity of demand for cigarettes among adults to U.S. state-specific prices per pack before and after the $1 per pack tax is imposed. The demand impact of the tax increase differs due to the cigarette tax rate in each state and the District of Columbia.

Estimates of the change in tax revenue due to decreased smoking among cohort members requires data on current (pre-intervention) per-pack taxes by state and the change in annual packs of cigarettes consumed over the cohort’s lifetime. Cigarette tax information was obtained from the CDC49 and the U.S. Economic Research Service,50 and the change in future demand for cigarettes within the cohort was obtained from Rivara et al.21 Data on adult price elasticity of demand for cigarettes comes from Chaloupka and Warner’s18 comprehensive study, which reports values ranging from −0.3 to −0.6; the sensitivity of this model was tested to these alternative estimates of price elasticity.

Two additional adjustments were made to estimates of tax revenue. First, overall demand for cigarettes will decrease due to lower numbers of adult smokers due to deaths and cessation that is unrelated to the interventions being evaluated. Up to 0.75% of current smokers aged ≥18 may die in any year,51 and up to 2% may quit through various means.52 It was also considered that norms around smoking may change over time, so that larger numbers of adults may quit than has been the case historically. Therefore, forecasts assume decreases in the absolute number of adult smokers ranging from 2% to 6% from the baseline number of smokers as of 2000. The second adjustment considers the prospect that inflation will erode the impact of the $1-per-pack tax increase over time. Price elasticity measures the percentage changes in consumption with respect to the relative change in price. A fixed increment to the price of cigarettes is being evaluated, so changes in the price of cigarettes will affect tax revenue. The relative price increase in cigarettes is adjusted based on historical inflation rates for all consumer goods, with inflation ranging from 3% to 6%.44

Section snippets

Results

The source and range for each model parameter are reported in Table 1. Results for the combined model as well as for each model component for discount factors ranging from 3% to 7% are reported in Table 2. Years of potential life saved are discounted values reported in Rivara et al.22 Direct healthcare costs and net tax revenue results are the product of simulations based on varying each of the model parameters reported in Table 1.

Because it is assumed that the media campaign takes place before

Discussion

The economic impact was estimated for two interventions, a media campaign and a $1 tax increase per cigarette pack, to decrease lifetime smoking prevalence among youths and adolescents. The cost per YPLL was estimated among a cohort of U.S. 18 year olds during 2000 who were followed up to age 85. When the change in tax revenue is included in the analysis, decreased years of potential life lost are achieved at a savings that varies from $586,000 to $1.4 million per YPLL. To put these results in

Conclusion

Smoking prevention programs and tobacco-use dependence treatment programs are among the most efficient ways of increasing population health. To date, the United States has not committed the resources commensurate with the returns that would be realized from a substantial investment in tobacco-control efforts. There is recent evidence of an increasing commitment to tobacco control—at least 18 states54 have recently increased their cigarette excise taxes, and 12 states now have per-pack taxes of

References (58)

  • Botvin GJ, Baker E, Dusenbury L, Botvin EM, Diaz T. Long-term follow-up results of a randomized drug abuse prevention...
  • Dent CW, Sussman S, Stacy AW. Project Towards No Drug Abuse: generalizability to a general high school sample. Prev Med...
  • Ellickson PL, Bell RM, McGuigan K. Preventing adolescent drug use: long-term results of a junior high program. Am J...
  • B.R. Flay et al.

    Six-year follow-up of the first Waterloo school smoking prevention trial

    Am J Public Health

    (1989)
  • D.M. Murray et al.

    Four- and five-year follow-up results from four seventh-grade smoking prevention strategies

    J Behav Med

    (1988)
  • D.M. Murray et al.

    Five- and six-year follow-up results from four seventh-grade smoking prevention strategies

    J Behav Med

    (1989)
  • A.V. Peterson et al.

    Hutchinson Smoking Prevention Projectlong-term randomized trial in school-based tobacco use prevention—results on smoking

    J Natl Cancer Inst

    (2000)
  • P.M. Lantz et al.

    Investing in youth tobacco controla review of smoking prevention and control strategies

    Tob Control

    (2000)
  • M. Siegel et al.

    The impact of an antismoking media campaign on progression to established smokingresults of a longitudinal youth study

    Am J Public Health

    (2000)
  • F.J. Chaloupka et al.

    The economics of smoking

  • E.M. Lewit et al.

    Price, public policy, and smoking in young people

    Tob Control

    (1997)
  • J. Nicholl

    Tobacco tax initiatives to prevent tobacco usea study of eight statewide campaigns

    Cancer

    (1998)
  • A broken promise to our childrenthe 1998 tobacco settlement six years later

    (2005)
  • M. Wakefield et al.

    Effectiveness of comprehensive tobacco control programmes in reducing teenage smoking in the USA

    Tob Control

    (2000)
  • C. Pechmann et al.

    An assessment of the United States and Canadian smoking reduction objectives for year 2000

    Am J Public Health

    (1998)
  • C. Pechmann et al.

    Anti-smoking advertising campaigns targeting youthcase studies from USA and Canada

    Tob Control

    (2000)
  • A.N. Mudde et al.

    The reach and effectiveness of a national mass media-led smoking cessation campaign in The Netherlands

    Am J Public Health

    (1999)
  • M. Gold et al.

    Cost-effectiveness in health and medicine

    (1996)
  • M. Drummond et al.

    Methods for the economic evaluation of health programmes

    (1997)
  • Cited by (0)

    Research was funded by the Robert Wood Johnson Foundation.

    No financial conflict of interest was reported by the authors of this paper.

    View full text