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 making indoor workplaces smoke-free environments.

Primary prevention.

Cost-effectiveness analysis.

The study population comprised people working in environments where smoking was allowed. Indoor workers were defined by the 1999 Current Population Survey.

The setting was the community. The economic study was carried out in the USA.

The effectiveness evidence and resource use information were taken from studies published between 1996 and 2002. The price year was 2000.

The clinical data included in the model were:

the percentage of workers covered by smoke-free workplace policies;

smoking prevalence;

the effect of smoke-free workplace policies on quitting;

the initial and steady-state relative risks of MI (both sexes) and stroke;

the initial relative risk of MI among passive smokers; and

the incidence of MI (both sexes) and stroke.

Initially, the authors estimated the number of US indoor workers not covered by smoke-free workplace policies. Then, using an exponential decline model, they calculated the effects on smoking behaviour, and on the relative risks of acute myocardial infarction (MI) and stroke, of making all workplaces smoke free. These 1-year and steady-state results provided a cohort for a Monte Carlo simulation with 10,000 iterations that was used to estimate the distribution of the reduction in cardiovascular events by varying individual parameters simultaneously. This modelled the impact of the policy over a 7-year period during which the relative risks of MI and stroke for ex-smokers and ex-passive smokers stabilised.

The clinical data were derived from a wide range of sources. The percentage of workers covered by smoke-free workplace policies was derived from the 1999 Current Population Survey. Smoking prevalence was derived from the 2000 Census. The initial relative risk of MI among passive smokers was derived from a published meta-analysis. All the remaining parameters were derived from a published study (Lightwood et al. 1997, see 'Other Publications of Related Interest' below for bibliographic details).

The authors did not state whether a systematic review of the literature was undertaken to identify the primary studies.

The measures of benefit, calculated as outputs of the model, were the number of deaths avoided and the number of MIs avoided.

The aim of the costing analysis was to calculate the costs saved and two cost elements were considered. The health care costs of treating MIs and strokes were included, whereas fatal cases of MI and stroke were assumed not to incur health care costs. Few details of the costing were reported in this paper as the authors referred to a published study on the same topic by one of the authors (Lightwood et al. 1997). Since the health technology was a behaviour change it was assumed that there were no direct costs for implementing the policy. These prices were reflated using the medical cost component of the Consumer Price Index, and these values then combined with the number of cases averted (which was derived from the model). The price year was 2000 and the costs were discounted at a rate of 3% per annum.

The costs were treated in a deterministic manner.

No indirect costs were included in this analysis.

US dollars ($).

Probabilistic sensitivity analysis was used to assess simultaneous variability in the parameter estimates. Normal distributions were used for all parameters except for relative risks, which used a log-normal distribution.

In the 7-year period it was estimated that 9,960 MIs, 1,250 strokes and 3,610 deaths would be prevented.

The total health care costs averted were $60.41 million over 1 year and $279.33 million over 7 years.

There was no synthesis of the costs and benefits as the policy was cost-saving and provided health benefits. The probabilistic sensitivity analysis indicated that the total health care saving of making all indoor workplaces smoke free varied between $40.95 million and $81.34 million in the first year.

Making all workplaces smoke free would result in considerable health and economic benefits.

The comparator, to have no smoke-free regulation, was justified as it was the current situation in the study setting. You should consider how the authors' setting compares with your own environment before applying the results of this study.

The prevalence and clinical data inputs to the model were derived from published literature, of which one study was by one of the authors of the present paper. There was no indication that a systematic review of the literature had been conducted.

Several health outcomes were derived from the model. These were all clinical outcomes such as deaths and MIs avoided.

The paper did not explicitly report the economic perspective adopted in the analysis, so it is not clear whether all the relevant cost categories were included. However, the authors did state that the true cost of cardiovascular disease was underestimated as the reduction in productivity loss was not calculated. Also, they did not evaluate the cost-savings arising from a reduction in strokes for passive smokers. Few details of the methods used to identify the costs were reported in this paper as they had been described in a previous study by one of the authors. The unit costs of treating nonfatal MIs and strokes and the number of cases were reported. The paper assumed that fatal MIs and strokes did not incur health care costs. However, it is unlikely that nonfatal MIs and strokes would not incur any health care costs. This means that this study is likely to have underestimated the cost-savings of making all workplaces smoke free. A probabilistic sensitivity analysis examined the impact of variability in the cost parameters on the cost estimates. The price year was reported, which will facilitate future reflation exercises. The costs were appropriately discounted.

The authors presented their analysis in a comprehensive manner and their conclusions reflected the scope of the analysis. The study was specific to the USA and the authors acknowledged that whilst sub-analyses by occupation would have been useful, there were insufficient recent data for this. They also acknowledged that their study did not consider different smoking prevalence rates in different social classes or the fact that not all smokers are regular smokers. The authors also noted that there have been increases in smoke-free workplaces since 1999 (the date of their estimate of the proportion of smoke-free workplaces). This means that any measures to implement universal smoke-free policies might not result in savings as large as those estimated in this study.

The authors did not make any direct recommendations for further research or changes to practice.

Supported by the National Cancer Institute and by a Health Resources and Services Administration Training grant.

Lightwood J, Glantz S. Short term economic and health benefits of smoking cessation: myocardial infarction and stroke. Circulation 1997;96:1089-96.