The potential for tobacco control to reduce PBS costs for smoking-related cardiovascular disease

Interventions to achieve cessation of tobacco smoking have consistently been shown to be effective and cost-effective.1-3 There is a compelling body of evidence from Australia and other countries that expenditure on tobacco control programs, particularly anti-smoking advertising, is correlated with reduced smoking rates.1 In turn, reductions in tobacco use have been clearly shown to reduce healthcare costs. Applied Economics has recently estimated that, since 1975, $2 in Australian healthcare expenditure has been saved for every $1 spent on tobacco control campaigns.4 Wider benefits to the community of reduced tobacco use were estimated to have totalled $8.602 billion, so that the net present value (NPV) of tobacco control programs was estimated to be $8.427 billion, with a benefit to cost ratio of 50 : 1.

About 20% of the Australian population smoke tobacco, and smoking rates exceed 30% in 20–29-year-old men, and 25% in 20–29-year-old women.5 Despite the substantial evidence that effective tobacco control programs would result in long-term cost savings, such interventions have not been fully or optimally implemented in Australia. For example, the very successful 1997 federal government-funded mass media advertising campaign has not continued, and the federal government now funds only a couple of weeks of anti-smoking advertising each year.1 Most smokers cannot recall their GP advising them to quit smoking,6 and, although bupropion is subsidised under the Pharmaceutical Benefits Scheme (PBS), nicotine-replacement therapy is not. This contrasts with New Zealand, for example, where a government subsidy for nicotine-replacement therapy, linked to calls to the Quitline, was introduced in 2000.1

One reason for suboptimal investment in smoking-cessation interventions is the paucity of data on the cost savings achievable within specific health budgets. To address this issue, we estimated the potential effect of a smoking-cessation intervention on PBS expenditure using PBS subsidy data and the government’s Intergenerational Report (IGR) model.7

For each drug, the proportion of prescriptions for specific CVDs was estimated from BEACH (Bettering the Evaluation and Care of Health) survey data, in which GPs report the indication for which they are prescribing each drug over the study period.8 A Supplementary Analysis of Nominated Data from BEACH on prescribing of lipid-lowering agents9 and a customised BEACH report on prescribing indications for the specified drugs over the two financial years 2000–02 were used. The former survey covered 1977 GPs and 2661 patient encounters, and found that 41% of patients who were prescribed lipid-lowering therapy had prior ischaemic heart disease. The latter survey covered the same number of GPs and 197 700 patient encounters.

For each CVD, the proportion of costs attributable to tobacco smoking was assumed to be equal to the product of the aetiological fraction for ever smoking tobacco10 and the total cost of subsidies for that disease. Although smoking has not been found to be a risk factor for hypertension, the National Heart Foundation of Australia guidelines recommend considering smoking in management decisions for patients with high blood pressure.11 We assumed that patients who were smokers would be started on antihypertensive medication at a lower blood pressure than non-smokers, based on absolute risk. The proportion of antihypertensive drug costs attributable to smoking was estimated by the proportion of individuals in the Ausdiab study12 who were receiving treatment for hypertension and had either controlled or mild hypertension and were smokers. This proportion was 7%.

PBS subsidies were then estimated for CVD and smoking-related CVD, by age and sex, by therapeutic class and by individual CVD.

The government spent $1.29 billion on drugs for CVD in 2001–02, 30.3% of total PBS subsidies. Subsidies for smoking-related CVD were $126 million, 2.96% of total PBS subsidies and 9.77% of CVD drug costs. Males under 65 years of age had the highest costs for smoking-related CVD. Costs for men and women aged 65 years and older were about the same (Box 1).

Ischaemic heart disease accounted for just over 60% of the costs of smoking-related CVD, and hypertension accounted for about a quarter (Box 2). The three most expensive drugs were all statins, comprising 48% of subsidies (Box 3). Note that these figures relate only to statin-prescribing subsidies for patients with prior ischaemic heart disease attributable to smoking.

The projected growth in total PBS subsidies and subsidies for drugs to treat CVD over the 40-year period of the IGR are shown in Box 4. The annual cost of the PBS is predicted to increase almost 14-fold, from $4.26 billion to $58.5 billion. The cost of drugs for CVD is predicted to increase from $1.29 billion to $17.7 billion per year by 2041–42.

Box 5 shows cost projections for PBS subsidies for drugs to treat smoking-related CVD. Assuming smoking rates do not change, PBS subsidies for smoking-related CVD are predicted to increase to $1.73 billion per year. With a 5% reduction in smoking prevalence, cumulative PBS subsidies would decrease by 17%, with a $4.5 billion reduction in costs for smoking-related CVD over the period, from $25.9 billion to $21.4 billion. When future costs are discounted at 5% per year, a $1.14 billion saving in cumulative PBS subsidies over the period is predicted, a 15.4% reduction, and annual costs in 2041–42 would be $40 million lower. With a 3% per year discount rate, the cumulative saving would be $1.9 billion.

The obvious question is how can this 5% reduction in smoking prevalence be achieved, and at what cost? After tax increases and bans on tobacco advertising, the most cost-effective smoking-cessation intervention is anti-smoking advertising using mass media, particularly television.1,2 Smoking prevalence in Australia had reached a plateau in the 4 years before the 1997 National Tobacco Campaign (NTC). However, it dropped by 1.7% during the initial 7-month phase of the campaign, which launched the now well-known “tumour” and “artery” advertisements. The NTC cost the federal government $7.1 million and the states contributed around $1.85 million. International experience shows that intensive, sustained anti-smoking advertising can produce reductions in smoking prevalence of 5% or more.14

Based on the cost and outcomes of the NTC, we estimate that an enhanced tobacco control program, encompassing a mass media campaign and improved population-level cessation services, would cost $10 million in the first year, $15 million in the second, and $20 million in the third year, and would produce the target 5% reduction in smoking prevalence. We assumed that more intensive (and expensive) advertising over the campaign period would be necessary to achieve further reductions as smoking prevalence declines. If $5 million per year was required to sustain this reduction in smoking prevalence, such a campaign would have an NPV of $1.039 billion, and an internal rate of return (IRR) of 33%, at a 5% per annum discount rate. The payback period would be less than 8 years. At a 3% per annum discount rate, the NPV would be $1.77 billion. Such a campaign would be an excellent investment.

Clearly, our estimates of the costs of this intervention are somewhat arbitrary, but any campaign with a discounted cost of less than $1.14 billion over the 40 years would be cost-saving. Furthermore, our calculations consider only the pharmaceutical costs to treat CVD. The IRR would be far greater and the payback period shorter if other cost offsets were included, for example reduced hospital and medical costs for CVD15 and healthcare costs for myriad other diseases caused or worsened by smoking. The effectiveness of the campaign would also affect its efficiency. We assumed that a 5% reduction in smoking prevalence would occur, based on Australian and international experience. This might vary, but our estimate seems plausible.

Our cost estimation and projection method has limitations. Our disease-costing method is consistent with that of the Australian Institute of Health and Welfare,10,16 but we made the additional assumption that some costs of treating hypertension are attributable to smoking, given the National Heart Foundation of Australia’s guidelines for management of hypertension.11 Further support for this assumption comes from the recently published HOPE study of secondary prevention, which treated people on the basis of absolute risk and reported improvement in cardiovascular outcomes.17

Our cost and cost-saving projections are predicated on the government’s own forward IGR report estimates of PBS costs, which were based on demographic, GDP and spending-growth assumptions.7 The PBS projections were underpinned by a 5.64% per year non-demographic growth rate, which corresponds to the observed real compound growth rate since 1983–84, corrected for the changing age composition of the population. Many assumptions in the IGR could be, and have been, questioned, and it should be noted that its projections related to total PBS subsidies rather than CVD drug costs. Nevertheless, the IGR projections have been used extensively by government as a basis for policy change, such as the proposal to increase patient copayments by almost 30%,18 and the prediction that PBS costs would increase from 0.6% to 3.4% of GDP has raised the question of the PBS’s sustainability. Although there are clearly uncertainties inherent in cost projections spanning 40 years, our model still predicts substantial savings after only 20 years — a 13.5% reduction in cumulative smoking-related CVD drug costs associated with a 5% reduction in smoking prevalence.

Our study demonstrates that anti-smoking interventions have the potential to help control cost pressures on the PBS. Although the projected savings might not be fully realised (as a consequence of changing CVD epidemiology or health system changes, for example), and savings would accrue to the PBS rather than the tobacco control budget, this should not deter implementation of anti-smoking interventions. Smoking-reduction strategies improve health outcomes as well as containing treatment costs borne by government, and therefore offer benefits not achieved through increased patient copayments.

4 Projected growth in Pharmaceutical Benefits Scheme subsidies

5 Smoking-attributable Pharmaceutical Benefits Scheme costs for drugs for cardiovascular disease