This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

The study examined a tobacco control programme. Since the study focused on the impact of smoking cessation, quitters were compared with smokers.

Primary prevention.

Cost-effectiveness analysis and cost-utility analysis.

The analysis considered the general population of individuals aged between 14 and 84 years. Specifically, 14 age-groups from 15 - 19 to 80 - 84 years were considered.

The setting was primary care. The economic study was carried out in Australia.

The effectiveness data were derived from studies published between 1995 and 2006. The economic data came from studies published between 1992 and 2006. The price year was 2001.

The clinical data used in the decision model were the probabilities of smoking-related diseases in smokers and quitters (AMI, stroke, lung cancer, COPD), smoking prevalence data, and mortality rates associated with smoking-related diseases.

The QBM focused on the effect of quitting on an individual smoker. The outcomes of the model were presented separately for male and female smokers and quitters, for various ages of quitting smoking, and with different durations of follow-up. The QBM is a Markov model. A simplified structure of the model was reported. The cycle length and main health states were described. The outputs of the model were four diseases associated with smoking, deaths, life expectancy and quality-adjusted life expectancy (QALE), and direct health care costs. The smoking-related diseases considered were acute myocardial infarction (AMI), stroke, lung cancer and chronic obstructive pulmonary disease (COPD). The model was analysed separately for smokers and quitters. In this study, the results were presented for a 10-year time horizon.

The probabilities of smoking-related diseases were derived from the general Australian population, such as the National Hospital Morbidity Database, the Australian Institute of Health and Welfare, the North East Melbourne Stroke Incidence Study (NEMESIS) that was carried out in a region of Melbourne in 1996/97, and the National Cancer Statistics Clearing House. Data on COPD were based on a personal communication from the University of Queensland. Smoking prevalence data came from the Australian Bureau of Statistics 2001 National Health Survey. Other population-based studies conducted in Australia were also used to derive mortality rates due to smoking-related diseases. Published non-linear and exponential models were used to estimate the relative risk of smoking-related diseases for quitters compared with smokers. Data to populate these models were derived from large case-control and cohort studies.

The authors did not state whether a systematic review of the literature was undertaken to identify the primary studies, which may therefore have been identified selectively. However, incidence probabilities for smoking-related diseases were taken from large Australian databases or population-based studies to reflect the study context. The exponential models used to estimate the declining risk for smoking-related disease due to smoking cessation were described in detail and justified.

The summary benefit measures used were the reductions in the four diseases associated with smoking (AMI, stroke, lung cancer and COPD), reduction in the number of deaths, the LYs saved and the QALYs gained. All measures were estimated using the modelling approach. The LYs and QALYs were discounted using different rates (3%, 5%, and also 0%). The utility weights for stroke were derived from a published meta-analysis, whilst those for other conditions were derived from several published studies. Most of the data came from the CEA Registry, which is usually considered to be a standard source of utility estimates.

The analysis of the costs appears to have been performed from the perspective of the health care system. It included the costs associated with the four smoking-related diseases (AMI, stroke, COPD and lung cancer). The cost items included hospitalisations, revascularisation procedures, drugs, visits to health professionals, and terminal care. The unit costs and the resource quantities were not presented separately. The costs and quantities of resource use were initially sought in the MEDLINE database and on the Internet. However, for AMI and COPD, no suitable Australian data were found. Thus, AMI costs were derived from Australian Refined Diagnosis Related Groups, while Canadian estimates were used for COPD. Stroke costs were mainly derived from the NEMESIS trial, a prospective Australian study. The costs of lung cancer were obtained from an Australian study published in 1992. Discounting was relevant, owing to the long time horizon of the analysis, and the costs were discounted at rates of 3% and 5% annually. Undiscounted results were also presented. The price year was 2001.

Statistical analyses of the costs or quantities were not performed.

Productivity costs were not included.

Australian dollars (AUD).

A univariate sensitivity analysis was performed to assess the robustness of the estimated costs and benefits to variations (+/- 10%) in all model inputs. A multivariate sensitivity analysis was also performed. In this analysis, AMI and stroke incidences were decreased by 10% to reflect population trends, while AMI and stroke survival, costs and utilities were increased by 10% to reflect the availability of more costly treatments, especially for AMI, that improve survival and quality of life.

In a hypothetical cohort of 1,000 quitters chosen at random from the (Australian) population of smokers aged between 15 and 74 years at the time of quitting and followed for 10 years:

the LYs saved for quitters in comparison with smokers were 47 (57 for men and 35 for women),

the QALYs gained were 75 (85 for men and 62 for women),

the cases of AMI avoided were 11 (14 for men and 6 for women),

the cases of COPD avoided were 19 (18 for men and 20 for women),

the cases of lung cancer avoided were 3 (3 for both men and women),

the cases of stroke avoided were 8 (8 for men and 7 for women),

the cases of any of the above four diseases avoided was 40 (43 for men and 37 for women),

the total deaths avoided were 18 (21 for men and 14 for women),

the AMI-related deaths avoided were 4 (6 for men and 2 for women),

the COPD-related deaths avoided were 1 (2 for men and 1 for women),

the lung cancer-related deaths avoided were 3 (2 for men and 2 for women),

the stroke-related deaths avoided were 3 (3 for men and 4 for women), and

the deaths related to any of the above four diseases avoided were 11 (12 for men and 9 for women).

When considering the different age groups, QALYs gained for quitters with respect to smokers (over a 10-year time period) ranged from:

0 in the age group 15 - 19 years to 0.5 in the age group 70 - 74 years for men; and

0 in the age group 15 - 19 years to 0.4 in the age group 70 - 74 years for women (discount rate 3%).

In a hypothetical cohort of 1,000 quitters chosen at random from the (Australian) population of smokers aged between 15 and 74 years at the time of quitting and followed for 10 years, the total health care costs avoided were AUD 373,000 (AUD 408,000 for men and AUD 328 for women).

The costs and benefits were not synthesised as quitters had lower costs and better outcomes than smokers (dominant).

The results of the sensitivity analysis corroborated the base-case findings. QALYs were most sensitive to a change in the utility of COPD, while costs were most sensitive to a change in the rate at which a quitter's risk of lung cancer or COPD returned to that of a smoker. However, only negligible changes in the expected costs and benefits were observed. A similar conclusion was reached in the multivariate sensitivity analysis.

The Quit Benefits Model (QBM) highlighted the clinical and economic benefits of a tobacco control programme showing the impact of the intervention on individual patients from the perspective of the health care system.

The tobacco control programme was compared with no intervention, although the analysis did not explicitly consider these two strategies. The analysis focused on the change in the burden of disease due to an effective tobacco control programme. You should decide whether they are valid comparators in your own setting.

The clinical data appear to have been identified selectively. The authors did not report the methods and conduct of a systematic review of the literature. Most of the sources used to derive clinical data were described. Whenever possible, Australian sources were used. The authors stated that, when uncertain estimates were found, the most conservative ones were chosen. For example, only the first of the four smoking-related diseases that occurred in an individual was considered, while in the real world individuals with AMI might develop lung cancer. Moreover, the model did not consider the impact of quitting on other individuals as a result of the effect of environmental tobacco smoke.

Several benefit measures were used in the analysis, some being specific to the disease considered in the study and others being comparable with the benefits of other health care interventions (i.e. QALYs and LYs). The sources of the utility weights were given, together with the values. The results were presented for different discount rates and for no discounting.

The analysis of the costs was restricted to the perspective of the health care system. Only the costs of the smoking-related diseases were considered. The sources of data were reported for all categories of costs. However, a detailed breakdown of the cost items was not given and some costs were presented as macro-categories. This may have implications for the generalisability of the study beyond the study setting. The costs were treated deterministically but variations in costs were considered in the sensitivity analysis. The price year was reported, which will facilitate reflation exercises to other time periods.

The authors did not compare their findings with those from other studies. They also did not explicitly address the issue of the generalisability of the study results to other settings. However, the sensitivity analysis enhances, to some extent, the external validity of the study. The authors highlighted the wide applicability of the QBM to other settings. Also, several sub-group analyses were conducted (stratification by age and gender), which is a positive feature of the study. The authors acknowledged some limitations of their analysis.

The authors stated that the "QBM can answer many of the questions posed by Australian policy-makers and health-care funders and will be a useful tool to evaluate tobacco control programs".

Supported by the Victorian Health Promotion Foundation, and the Cancer Council Victoria.

Because readers are likely to encounter and assess individual publications, NHS EED abstracts reflect the original publication as it is written, as a stand-alone paper. Where NHS EED abstractors are able to identify positively that a publication is significantly linked to or informed by other publications, these will be referenced in the text of the abstract and their bibliographic details recorded here for information.

Johansson PM, Tillgren PE, Guldbrandsson KA, Lindholm LA. A model for cost effectiveness analyses of smoking cessation interventions applied to a Quit-and-Win contest for mothers of small children. Scand J Public Health 2005;33:343-52.

Hurley SF. The short-term impact of smoking cessation on myocardial infarction and stroke hospitalisations and costs in Australia. Med J Australia 2005;183:13-7.

Orme ME, Hogue SL, Kennedy LM, et al. Development of the health and economic consequences of smoking interactive model. Tob Control 2001;10:55-61.

Hurley SF, Scollo MM, Younie SJ, et al. The potential for tobacco control to reduce PBS costs for smoking- related cardiovascular disease. Med J Australia 2004;181:252-5.