Andrew M Tonkin (Andrew Tonkin), Monash University, Melbourne, Australia, et al. Government policies have a huge impact on national health. Tobacco control may illustrate this point most vividly. However, smoking remains a global problem and a major cause of preventable death. Countries with the highest per capita prevalence of smoking include Bangladesh (20.9% of adults), Brazil (16.2% of adults), China (31.4% of adults), Germany (27.2% of adults), India (32.7% of men and 1.4% of women), Indonesia (34.5% of adults), Japan (43.3% of men and 12% of women), the Russian Federation (60.4% of men and 15.5% female), Turkey (34.6%), and the United States (23.2%). The prevalence of smoking among young people varies, with 18.4 percent of youth in the United States still smoking. The power of government intervention is further confirmed by the results of a meta-analysis on the impact of smoking bans on hospitalizations for acute myocardial infarction (AMI), which is included in this issue of the journal. The meta-analysis compared the incidence of AMI in different populations before and after legislation to restrict smoking in public areas in North America and Europe. At 12 months after legislation, the pooled random effects estimate for AMI hospitalization rate was 0.83 [95% confidence interval (CI) = 0.80 ~ 0.87]. This estimate continued to increase as the time was extended to 3 years (the longest follow-up period). Differences in follow-up duration were the most significant reason for inconsistent results across studies. The conclusions of this study are reliable because it realistically simulates individual risk and exposure scenarios. The inability to directly assess group activity effects is a difficulty encountered in evaluating group-level health promotion interventions. Public health is a complex issue influenced by multiple factors, and it is often not possible to measure the effects of a particular program in isolation from other factors. Thus a particular study can only measure effects that can or cannot be generalized in a particular context. Meta-analysis can address this problem to some extent by combining the results of studies done in different settings to assess overall effects. However, the results of the analysis are often limited by the different designs and endpoints of the studies, and the results must still be interpreted with caution. Another approach to assessing population health interventions is to use modeling techniques that combine epidemiological observations and clinical research evidence to extrapolate the possible effects of a health intervention to a population that may differ from the original study. However, epidemiological models are often based on untestable hypotheses and simplified characteristics of interventions and disease processes, and often integrate a large number of imputed parameters, which can lead to false approximations between model results and the data on which the parameters are based. Lightwood and Glantz used a combination of these methods. Their meta-analysis and model, derived parameter estimates from different independent data sources. Thus, the consistency between their results lends credibility to the results of each method. This is important because decision makers need to be informed about these data when weighing costs/benefits and preferred measures between health interventions for different populations. Another feature of epidemiologic models is that they can highlight gaps in the data informing health policy and strategy. Secondhand smoke risk studies typically rely on self-reporting. As discussed by the authors, current methods for assessing individual exposure by measuring cotinine levels (a stable metabolite of nicotine) are not yet widely available. However, all of the scenario models applied by Lightwood and Glant that approximated the results of the meta-analysis were derived from studies that used cotinine levels. No control information is available for comparison to similar modern communities without smoke-free regulations. However, some single study data support the positive impact of the legislation. Longer-term data may detect possible effects of other factors that have an impact on long-term change. The referral tested in the meta-analysis was for hospitalization due to AMI. Coronary plaque rupture and thrombosis were key. A reduction in out-of-hospital deaths has also been demonstrated, and studies with broad endpoints have also shown similar effects for sudden myocardial infarction. The term “secondhand smoke” encompasses the involuntary nature of exposure and includes both the “sidestream” smoke from the shaded combustion of a cigarette and the “mainstream” smoke exhaled by the smoker, the latter being more important. Secondhand smoke is a major cause of preventable death and was addressed in full in a 2006 report by the U.S. Surgeon General. The report provides strong evidence to support the immediate adverse effects of secondhand smoke on the cardiovascular system and, importantly, inferred a causal relationship between secondhand smoke and coronary heart disease (CHD) disability and death rates. The report also notes that evidence of a causal relationship between secondhand smoke and increased risk of subclinical vascular disease and stroke is available, but not sufficient to draw conclusions. Although exposure to secondhand smoke has decreased in recent decades, the report estimates that 60% of US nonsmokers have biological evidence of exposure to secondhand smoke. The National Health and Nutrition Examination Survey examined cotinine levels in nonsmokers ≥20 years of age during 1999-2000 and found that cotinine was detectable in 46% of residents of smoking jurisdictions compared with 13% of residents of jurisdictions without such an ordinance. Secondhand smoke is estimated to increase the risk of incident CHD by 25-30%. Exposure is usually estimated based on the number of cigarettes smoked per day by a spouse or partner. The relative risk was 1.16 for even low-moderate exposure compared with no exposure, and the evidence supporting these associations was derived from cohort and case-control studies with different exposure patterns and differences in control factors over a 20-year follow-up period. The slightly stronger correlation for case-control studies may partially reflect bias associated with retrospective recall and possibly other biases. Although exposures may be misclassified, the strength of the association with referral can be underestimated by not considering background secondhand smoke and by using only nonsmokers whose spouse is a smoker or not. Effect estimates were increased by a factor of 2 when environmental range cotinine levels of exposure were used. In a 20-year prospective study, Whincup et al. demonstrated a dose-dependent association between serum cotinine and CHD events, with a 57% increased risk for individuals in the highest quartile of cotinine levels. The effect of any confounding factors such as poor diet appears to be rather small. With a relative risk of CHD of 80% for active smokers and only 1% of the smoke exposure of nonsmokers who smoked 20 cigarettes per day, the effect of secondhand smoke on the risk of CHD was greater than expected. However, this large effect seems biologically possible and consistent with a nonlinear effect of small doses of tobacco exposure, including important effects on platelet and endothelial function. This effect on platelet and endothelial function, arterial stiffness, oxidative stress, and inflammatory markers is approximately 80% to 100% of the effects associated with active smoking. Other adverse effects of secondhand smoke include effects on matrix metalloproteinases, HDL cholesterol, and mitochondrial energy utilization. The effects of secondhand smoke are not only larger but also quickly apparent. An important early study showed that platelet activation and aggregation, as well as endothelial cell damage, occurred within 20 minutes of secondhand smoke exposure in nonsmokers, whereas there was no further activation of platelets in active smokers. Similarly, breathing secondhand smoke for 30 minutes can lead to similar levels of endothelial cell dysfunction as in active smokers. Recovery of endothelial functional impairment is slow after ending chronic high levels of secondhand smoke exposure. However, platelet aggregation, a key factor in acute coronary syndromes, is rapidly diminished. Secondhand smoke exposure can occur in a variety of settings, particularly at home and in the workplace, but also in restaurants, bars, gaming establishments, and motor vehicles, with varying degrees of exposure. Women are not often active smokers, but are consistently the largest bearers of secondhand smoke load. Secondhand smoke effects may be particularly dangerous for children. Children have narrower airways and faster breathing rates, and inhale 3 to 4 times more air and possibly secondhand smoke relative to body weight than adults. The association of secondhand smoke exposure in children with their subclinical atherosclerosis has not been proven. However, the correlation between secondhand smoke exposure and carotid intima-media thickness has been shown in the Atherosclerosis Risk Community Study, where children take many years to manifest a disease with a long latency period such as atherosclerosis. Disturbingly, between 1988-1991 and 1999-2002, cotinine levels in children decreased less than in adults. During 1999-2002, 59.6% of U.S. children aged 3-11 years had cotinine levels ≥0.05 ng/ml, with a median cotinine concentration of 0.09 ng/ml, compared with 0.035 ng/ml in older adults. non-smokers of lower socioeconomic status are also vulnerable due to exposure to environments with higher rates of active smoking and other environmental conditions. Among indigenous populations with high smoking prevalence, reducing smoking, and thus secondhand smoke, may be the single most important short-term action to improve their life expectancy. As of July 2009, the WHO Framework Convention on Tobacco Control included 166 groups. Secondhand smoke is one of the six most important and effective policies proposed by the framework, along with higher taxes and prices, health warnings, QUIT programs, advertising and sponsorship bans, and careful monitoring of the tobacco epidemic and prevention policies. Where possible, policies should also include cotinine or other biomarker data. The current meta-analysis provides stronger evidence to support legislation to promote smoke-free environments with significant effectiveness. Although the effects are rapid, they continue to increase over time. The public health impact of secondhand smoke is enhanced by high CHD prevalence, thus emphasizing the importance of legislation. The California Environmental Protection Agency estimates that in 2005, 3,400 lung cancer deaths among adult nonsmokers and 430 deaths associated with sudden infant death syndrome were attributed to secondhand smoke in the United States alone, and 46,000 CHD deaths were caused by secondhand smoke. The above regulations can also influence active smoking. A systematic evaluation of 26 studies showed that smoke-free workplaces resulted in a 3.8% reduction in smoking prevalence, a 3.1 cigarette/day reduction in the number of cigarettes smoked by persistent smokers, and a 29% reduction in total cumulative cigarette consumption. A study included in the meta-analysis found a reduction in hospitalizations for acute coronary syndrome among smokers and nonsmokers. In addition, rather than negatively impacting business activity, a smoking ban could increase patronage of food and beverage establishments. Clinicians should advise patients to avoid public places where smoking is allowed and should advise their family members not to smoke in their homes or in cars carrying patients. Health care workers can also be strong advocates, and studies such as the one reported in this issue of the journal can help strengthen government action.