This section briefly summarises key economic terms and methods. References to texts that describe the methodologies in more detail are provided. Models used to attribute costs to tobacco use are described and the methods used to predict the health benefits of quitting are detailed.
In economics, an item’s cost is its opportunity cost, defined as the value of the foregone benefits because the resource is not available for its best alternative use. Costing is sometimes a two-step process: the quantity of resources used is measured, then the resource is valued according to its opportunity cost. For practical reasons market prices are mostly used as an estimate of the opportunity cost.1
The perspective, or viewpoint, of the economic study is relevant when estimating costs or reviewing economic analyses. For example, patients’ travel costs to attend health care consultations should be included in an analysis from society’s viewpoint (the societal perspective) but would not be included in an analysis from a department of health’s viewpoint.
The concept of transfer payments is also relevant in economic analysis. Transfer payments are money transfers which do not reflect resource consumption. Taxes are an example. They represent a gain to the government and a cost to the tax payer, but they are neither a cost nor a gain to society; money is simply transferred. Workers compensation payments are another example of a transfer payment. Transfer payments are typically not measured in economic studies.1
Costs are categorised in different ways depending on the purpose of the analysis and the analyst. Collins and Lapsley, whose work on the costs of smoking in Australia is summarised in Section 188.8.131.52,2 categorise costs as private costs or social costs using definitions published by Markandya and Pearce.3 Private costs are defined as costs knowingly and freely borne by the consumer or producer. Costs not knowingly and freely borne are referred to as social costs. So, the cost of any activity is the sum of the private and social costs. Social costs are also referred to as negative externalities, and are defined as costs incurred by a party who did not agree to the action causing the cost. Collins and Lapsley’s report estimates the social costs of smoking rather than the private costs.2 They argue that smokers have become addicted to tobacco without full knowledge of the consequences and thus Collins and Lapsley regard the costs of most, if not all, the consequences of smoking as social costs. So, for example, their analysis classifies the costs of medical care and absenteeism from work due to smoking-associated illnesses as social costs.
In economic evaluation of health care programs, such as smoking cessation pharmacotherapies, costs have, until recently, been classified as either direct (costs incurred in the health sector), indirect (costs incurred by patients or their families) or intangible (costs that are difficult to measure, such as pain and suffering). These terms are no longer favoured because of potential ambiguities.1 However, Collins and Lapsley do categorise social costs as either tangible and intangible. The tangible category includes productivity costs, health care costs and the cost of purchasing tobacco. The intangible category includes the psychological impact (on others) of a premature death and the loss of enjoyment of life, although they do not estimate a value for the latter component.
The following cost classification is used in situations where costs vary with the quantity of a resource produced or consumed:
Total cost: the cost of producing a particular quantity of output
Fixed cost: costs which do not vary with the quantity of output
Variable cost: costs which vary with the level of output
Cost function: total cost as a function of quantity
Average cost: the average cost per unit of output
Marginal cost: the extra cost of producing one extra unit of output
17.1.2 Discounting of future costs and benefits
Economic analysis typically requires comparisons of alternative choices of action where costs and consequences do not all occur in the present. For example, when deciding whether or not to invest in a mass media anti-smoking campaign, the cost of the campaign might be incurred over the next five years, but the benefits, in terms of, say, reduced health care costs for smoking-associated illnesses, occur over the next fifty years or so. It is always advantageous to receive a benefit earlier and incur a cost later. Economists refer to this as time preference and in economic analysis future costs and benefits are discounted to present value to reflect time preference.1
When future costs are discounted to present value an annual discount rate must be chosen. Analyses that have used different discount rates will not be comparable. Until the early 1980s, a 5% per annum discount rate was widely used. Since then, a 3% per annum discount rate has become more common as some economist have argued that it more accurately reflects the cost of capital.1, 4
17.1.3 QALYs and DALYs
Two health outcome measures frequently used in economic analysis are the quality-adjusted-life-year (QALY) and the disability-adjusted-life-year (DALY). The QALY and the DALY are similar in concept. They both capture the adverse health impact of an activity such as smoking on mortality (quantity losses) and morbidity (quality losses). The number of QALYs lost by a smoker, for example, is calculated by adding the number of years of life lost due to premature death from illnesses caused by smoking to the number of years of life in which the quality of life is adversely affected by smoking-associated illnesses adjusted for the utility of those years. Utility in this context refers to the preference a person may have for a health outcome; the more preferable an outcome, the higher the utility value. In economic evaluation the utility of a health outcome is typically measured on a scale from zero to one, with ‘zero’ representing the least preferable health outcome—death—and ‘one’ representing the most desirable outcome—perfect health.1
Similarly, the DALY describes the amount of time lost due to both premature death and disability. It is defined as the sum of the number of years of life lost due to premature death and the number of equivalent ‘healthy’ years lost due to disability. The number of years of life lost due to disability is calculated by multiplying the duration of the disability by the disability severity weight.5
17.1.4 Studies of the cost of smoking
184.108.40.206 Aetiological fractions
Quantifying the total costs of smoking in Australia (or other countries) is not straightforward because many of the consequences of smoking (such as stroke and other tobacco-associated illnesses) can also be caused by other risk factors. For example, stroke is caused by hypertension (elevated blood pressure) as well as by smoking, so it would not be accurate to attribute the cost of all hospital admissions for stroke to smoking. A measure known as the aetiological fraction is therefore used. The aetiological fraction (also known as the population attributable risk) is defined as the proportion of cases in the population that can be explained by the risk factor.6 The aetiological fraction is a function of the prevalence of the risk factor (the smoking rate in the case of tobacco) and the risk of the disease in people who smoke relative to those who do not. A comprehensive set of aetiological fractions for morbidity and mortality caused by tobacco has been calculated for Australia by English et al.7, 8 building on the work of Holman et al.9, 10
English et al. estimated that in Australia in 1995, 44% of stroke in men and 39% of stroke in women was caused by cigarette smoking. Lung cancer has a much higher aetiological fraction than stroke. In 1995, 84 per cent of lung cancer in Australian men and 77% in Australian women was caused by cigarette smoking.8 Ridolfo and Stevenson11 and Begg et al.5 subsequently extended and updated English and colleagues’ work.
It should be noted that the aetiological fractions calculated for Australia are based on Australian smoking prevalence data, but the disease risk estimates come from overseas studies. Sitas and colleagues in 2009 called for inclusion of smoking data on death notification forms in order to develop more accurate Australian estimates for the risk of smoking-associated illnesses.1 13
Analysts estimating the economic costs of smoking for other countries generally assume that the aetiological fraction for disease is also the aetiological fraction for costs. In other words, 44% of the cost of hospital admissions for stroke in men and 39% of the cost of stroke admissions for women would be attributed to smoking.
220.127.116.11 Time frame of analysis
Studies of the cost of smoking typically either estimate the cost of smoking in a specified year, or estimate the cost over the lifetime of a smoker or smokers. The first method, which is used by Collins and Lapsley in their reports for the Australian government,2 is referred to as the demographic approach2 or the prevalence approach.5 The consequences of past and present smoking by the population in the year of the study are estimated. The lifetime cost approach estimates a future stream of costs due to smoking, discounted to present value using an accepted discount rate. When morbidity and mortality consequential to smoking are valued in monetary terms (in cost-benefit analyses for example) this method is referred to as the human capital approach.
17.1.5 Economic evaluation
Economic evaluation is defined as ‘the comparative analysis of alternative courses of action in terms of both their costs and consequences’. Economic analysis deals with both costs (inputs) and consequences (outputs).1
18.104.22.168 Purpose of economic evaluation
Economic evaluation of a health program is undertaken to determine if it is worth doing, compared with other options for the same funds. Economic evaluation of tobacco control has been conducted to:
- Determine if tobacco control programs per se have been, or will be, a good investment compared with other public health interventions.14
- Compare the costs, effectiveness and efficiency (net cost per unit of health outcome) of different approaches to tobacco control and make choices about which approach to adopt. For example, cost-effectiveness analyses have compared smoking cessation pharmacotherapies (nicotine replacement therapy, varenicline and bupropion) in terms of their marginal cost per additional quitter.
- Estimate the net cost or net benefit of a tobacco control program as part of a process to decide whether the program should be implemented. For example, a cost-benefit analysis of proposed new health warnings on tobacco products in Australia was conducted in 2003.15
22.214.171.124 Economic evaluation methods
Different economic evaluation techniques are used depending on the purpose of the evaluation and the data available. Methods for the economic evaluation of health programs are described in detail in two key texts: Drummond et al.1 and Gold et al.4 A brief summary of the methods is available in an article by Weinstein.16
There are three main types of economic evaluation: cost-effectiveness analysis (CEA), cost-utility analysis (CUA) and cost-benefit analysis (CBA).1
Cost-effectiveness analyses compare the cost per unit of effect for two or more programs. The unit of effect is the same for both programs but either the magnitude of the effect, or the cost of achieving a unit of effect, or both, differ between the two programs. For example, nicotine replacement therapy and varenicline both increase the chance that a smoker will quit, but the cost per course typically differs. The two treatments can be compared in a CEA.
Cost-utility analyses are a particular type of CEA where the unit of effect involves quality of life. In most CUAs the unit of effect is the QALY. For example, an effective anti-smoking advertising campaign will increase the quit rate, and therefore will improve both life expectancy (because the risk of death from smoking-associated illness is decreased) and quality of life (because the incidence of smoking-associated illness is decreased). A combined measure of these two effects is the number of QALYs gained. The economic efficiency of an anti-smoking campaign can therefore potentially be compared with that of other health programs—a screening program for colon cancer, for example—in terms of their costs per QALY.
In cost-benefit analyses, both the costs and consequences of a program are expressed in monetary units. The results can be summarised either as a ratio of costs to benefits, or as a sum, representing the net benefit. The net benefit can be negative, i.e. a loss. CBA requires assignment of a monetary valuation to health outcomes. Such a value is difficult to estimate and some analysts feel uncomfortable making the estimation. The advantage of CBA, however, is that it provides information on the absolute benefit of a program. If a program is associated with a net benefit, when all costs and consequences are considered and valued, then the budget should be increased to fund the program.
In contrast to CBA, CEA and CUA are best suited to situations where there is a fixed budget. If programs are funded on the basis of their cost-effectiveness—i.e. the most efficient program, with the lowest cost per life-year or QALY, is funded first, followed by the next most efficient, and so on until the budget is spent—then health outcomes will be maximised. In practice, the cost-effectiveness ratio or cost-utility ratio of a program is often compared to some threshold below which health care programs are typically funded, and above which they are not. For example, in Australia, a decision to subsidise a medicine under the Pharmaceutical Benefits Scheme (PBS) takes into account the medicine’s ‘cost-effectiveness’. Although there is no ‘official’ threshold for the PBS, an analysis of Pharmaceutical Benefits Advisory Committee (PBAC) decision-making between 1991 and 1996 indicated that the PBAC was likely to recommend a drug for listing if its additional cost per life-year gained was less than $42,000, but was unlikely to recommend listing if this ratio was greater than $76,000 (1998–99 dollars).17 A more recent study estimated the expected QALYs gained from additional health expenditure in Australia, with results suggesting opportunity costs of one QALY for every additional AUD28,033 of government health expenditure.18 An analysis of decisions by the National Institute for Health and Clinical Excellence (NICE) in the United Kingdom indicated that it tends to recommend technology when the cost per QALY gained is in the range US$30,600 to US$45,900.19
The fact that economic evaluations are used in this manner to inform allocation of resources highlights the need for high quality analyses and consistent methodology. The latter is particularly important because a program’s cost-effectiveness or cost-benefit profile depends on a large number of factors that are specific to the analysis. These include: what the program is being compared to (the comparator), the program’s cost relative to the comparator, its effectiveness relative to the comparator, the perspective of the analysis (for example, only costs incurred by government might be included in one analysis but costs incurred by patients as well as government might be included in another) and the time frame of the analysis. The PBAC has guidelines for drug subsidy submissions that specify how cost-effectiveness analyses should be conducted,20 but these guidelines are not necessarily followed in evaluations conducted for other purposes.
126.96.36.199 Models to assess the health benefits of reduced smoking prevalence
All economic evaluation is underpinned by evidence (or an assumption) that the program being evaluated is effective. In the case of smoking cessation, the intervention must increase the rate of quitting. Other tobacco control interventions would be judged effective if they reduced the rate of uptake of smoking, reduced exposure to environmental tobacco smoke or reduced the harmfulness of tobacco products. Economic evaluations that assess the benefit of tobacco control interventions in terms of health outcomes, such as deaths avoided or years-of-life saved, use models to extrapolate from the number of fewer people smoking, or exposed to the toxins in tobacco smoke, to the health benefits of never taking up smoking, of quitting or of avoiding exposure to tobacco smoke.
A model to estimate the benefits of quitting for Australian smokers was developed by Hurley and Matthews for the Cancer Council Victoria21 and has been used to analyse the cost-effectiveness of the Australian National Tobacco Campaign22 and the cost-effectiveness of smoking cessation programs to prevent age-related macular degeneration.23 Known as the Quit Benefits Model (QBM), it assesses the consequences of quitting in terms of cases avoided of the four most common smoking-related diseases: acute myocardial infarction (AMI), stroke, lung cancer and chronic obstructive pulmonary disease (COPD). The model also assesses deaths avoided, QALYs saved and health care costs saved. Quitting outcomes can be predicted for males and females in 14 five-year age groups from 15–19 to 80–84 years. The model uses 2001 as the reference year.
The QBM is underpinned by exponential models of the decline over time after quitting in the risk of AMI, stroke, lung cancer, COPD and death. These exponential models are based on data from large international studies of smokers who quit, and predict that the risk of AMI and stroke, relative to smokers, is approximately halved two years after quitting. The risk of lung cancer and COPD declines more gradually. Ten to 19 years after quitting, the risk is reduced by 60% relative to smokers for men, and 80% for women.
Using the QBM, Hurley and Matthews predict that the average saving per 1000 random quitters in the first ten years following quitting would be $373,000 in health care costs associated with AMI, COPD, lung cancer and stroke. 21 Overall, 40 of these quitters would avoid a diagnosis of the four smoking-associated diseases in the first ten years following quitting, with an estimated saving of 47 life-years and 75 QALYs.
Models similar to the QBM have been developed for the populations of the United States,24 the United Kingdom and various European countries.25, 26 An American simulation model, Simsmoke, uses a different approach. It predicts the effect of tobacco control policies on smoking patterns, based on the impact of past tobacco control programs on quitting.27 The consequences in terms of smoking attributable deaths are then projected, but the impacts on smoking-associated morbidity and heath care costs are not modelled.
Section 188.8.131.52 highlighted the potential for differences in economic evaluation methodology to explain some of the apparent differences in cost-effectiveness between programs. A tobacco control program’s cost-effectiveness or cost-benefit profile will also depend on the particular model used to predict the health and economic benefits of quitting. Models vary in terms of: the number of smoking-associated illnesses considered; whether morbidity, as well as mortality, is predicted; the period over which benefits are modelled; whether predicted health care cost savings from smoking-associated illnesses avoided are offset; and the discount rate applied to future costs and benefits. These possible variations need to be borne in mind when comparing cost-effectiveness or cost-benefit ratios.
1. Drummond MF, Sculpher MJ, Torrance GW, O'Brien BJ, and Stoddart GL, Methods for the economic evaluation of health care programmes. Third edOxford: Oxford Medical Publications; 2005.
2. Collins D and Lapsley H. The costs of tobacco, alcohol and illicit drug abuse to Australian society in 2004/5. P3-2625. Canberra: Department of Health and Ageing, 2008. Available from: https://nadk.flinders.edu.au/files/3013/8551/1279/Collins__Lapsley_Report.pdf.
3. Markandya A and Pearce DW. The social costs of tobacco smoking. British Journal of Addiction, 1989; 84(10):1139-50. Available from: https://www.ncbi.nlm.nih.gov/pubmed/2819272
4. Cost-effectiveness in health and medicine. ed. Gold MR, et al. New York: Oxford University Press; 1996.
5. Begg S, Vos T, Barker B, Stevenson C, Stanley L, et al. The burden of disease and injury in Australia 2003. Phe 82. Canberra: AIHW, 2007. Available from: https://www.aihw.gov.au/getmedia/f81b92b3-18a2-4669-aad3-653aa3a9f0f2/bodaiia03.pdf.aspx?inline=true.
6. Breslow NE and Day NE, Statistical methods in cancer research. Volume 1 - the analysis of case-control studies. Lyon: International Agency for Research on Cancer; 1980.
7. English DR, Holman CD, Milne E, and et al., The quantification of drug caused morbidity and mortality in Australia 1995: Part 1. Canberra: Commonwealth Department of Human Services and Health; 1995.
8. English DR, Holman CD, Milne E, and et al., The quantification of drug caused morbidity and mortality in Australia 1995: Part 2. Canberra: Commonwealth Department of Human Services and Health; 1995.
9. Holman CDJ, Armstrong BK, Arias LN, and Martin CA. The quantification of drug caused morbidity and mortality in Australia 1995: Part 1. Canberra: Department of Community Services and Health 1990.
10. Holman CDJ, Armstrong BK, Arias LN, and Martin CA. The quantification of drug caused morbidity and mortality in Australia 1995: Part 2. Canberra: Department of Community Services and Health 1990.
11. Ridolfo B and Stevenson C. Quantification of drug-caused mortality and morbidity in Australia, 1998. Canberra, Australia: Australian Institute of Health and Welfare (Drug Statistics Series no 7), 2001. Available from: https://www.aihw.gov.au/getmedia/7e677c0d-e6c1-4ec8-a78f-62982758f61f/qdcmma98.pdf.aspx?inline=true.
12. Australian Government. Taking preventative action: Government's response to Australia: The healthiest country by 2020. 2010. Available from: https://apo.org.au/sites/default/files/resource-files/2010/05/apo-nid21989-1365646.pdf
13. Sitas F, O'Connell DL, Jamrozik K, and Lopez AD. Smoking questions on the Australian death notification form: Adopting international best practice? Medical Journal of Australia, 2009; 191(3):166-8. Available from: https://www.ncbi.nlm.nih.gov/pubmed/19645648
14. Applied Economics. Returns on investment in public health: An epidemiological and economic analysis Department of Health and Ageing, 2003. Available from: http://www.appliedeconomics.com.au/pubs/reports/health/index.htm#TopOfPage.
15. Applied Economics. Cost-benefit analysis of proposed new health warning on tobacco products. Commonwealth Department of Health and Ageing, 2003.
16. Weinstein MC. Economic assessments of medical practices and technologies. Medical Decision Making, 1981; 1(4):309-30. Available from: https://www.ncbi.nlm.nih.gov/pubmed/6820459
17. George B, Harris A, and Mitchell A. Cost effectiveness and consistency of decision-making. Evidence from pharmaceutical reimbursement in Australia 1991-1996 Pharmacoeconomics, 2001; 19(11):1103-9.
18. Edney LC, Haji Ali Afzali H, Cheng TC, and Karnon J. Estimating the reference incremental cost-effectiveness ratio for the Australian health system. Pharmacoeconomics, 2018; 36(2):239-52. Available from: https://www.ncbi.nlm.nih.gov/pubmed/29273843
19. Pearson SD and Rawlins MD. Quality, innovation, and value for money: Nice and the British national health service. JAMA, 2005; 294(20):2618-22. Available from: https://www.ncbi.nlm.nih.gov/pubmed/16304076
20. Pharmaceutical Benefits Advisory Committee. Guidelines for preparing submissions to the pharmaceutical benefits advisory committee (version 4.2). . Department of Health and Ageing, Commonwealth of Australia, 2007.
21. Hurley SF and Matthews JP. The quit benefits model: A markov model for assessing the health benefits and health care cost savings of quitting smoking. Cost Eff Resour Alloc, 2007; 5:2. Available from: https://www.ncbi.nlm.nih.gov/pubmed/17241477
22. Hurley SF and Matthews JP. Cost-effectiveness of the Australian national tobacco campaign. Tobacco Control, 2008; 17(6):379-84. Available from: https://www.ncbi.nlm.nih.gov/pubmed/18719075
23. Hurley SF, Matthews JP, and Guymer RH. Cost-effectiveness of smoking cessation to prevent age-related macular degeneration. Cost Eff Resour Alloc, 2008; 6(1):18. Available from: https://www.ncbi.nlm.nih.gov/pubmed/18783631
24. Tengs TO, Osgood ND, and Chen LL. The cost-effectiveness of intensive national school-based anti-tobacco education: Results from the tobacco policy model. Preventive Medicine, 2001; 33(6):558-70. Available from: https://www.ncbi.nlm.nih.gov/pubmed/11716651
25. Orme ME, Hogue SL, Kennedy LM, Paine AC, and Godfrey C. Development of the health and economic consequences of smoking interactive model. Tobacco Control, 2001; 10(1):55-61. Available from: https://www.ncbi.nlm.nih.gov/pubmed/11226362
26. Johansson P, A model for economic evaluations of smoking cessation interventions - technical report. Stockholm: Samhallsmedicin; 2004.
27. Levy DT, Bauer JE, and Lee HR. Simulation modeling and tobacco control: Creating more robust public health policies. American Journal of Public Health, 2006; 96(3):494-8. Available from: https://www.ncbi.nlm.nih.gov/pubmed/16449585