Brazil has been a leader in tobacco control policy over the past two decades, having introduced many of the key policies of the Framework Convention for Tobacco Control. This has proved a great success in that smoking prevalence has halved, both in men and women since the early 1990s. This is reflective of the broader trend in many middle-income countries, which are now faced with a growing burden of non-communicable disease, and are taking bold policy action to address this. So for example, Mexico was one of the first countries to introduce a tax on sugar-sweetened beverages. The findings from evaluations of this policy have persuaded governments around the world, including in Europe and North America, to adopt similar measures. I've had the good fortune to have worked in Brazil since 2016. I know firsthand, the country serves as a rich laboratory for the evaluation and learning for public health policy globally. It has introduced bold policies to reduce poverty, deliver universal health care, and improve nutrition, which many other countries are learning from. It has also been a great place to undertake robust evaluation of tobacco control policies for a variety of reasons. Firstly, the country introduced bold action on tobacco control since the 1990s, as I've discussed. Secondly, the country has devolved government in decision making. Third, it has phased implementation of tobacco control policies. The fourth reason is the country has high quality data compared to many other low and middle-income countries. So let's now look at how smoke-free laws in Brazil were implemented, and think about how this influences the type of study design you can consider. Up until 2008, most Brazilian states were following the federal law implemented in 1996, which had minimal restrictions, and meant that smoking was still widely permitted in indoor public places. Around 2009, some states implemented stricter but still partial restrictions, which typically involved the creation of sealed smoking rooms in restaurants and bars. Other states, including the big states of San Paolo and Rio de Janeiro, introduced comprehensive laws which were fully compliant with the Framework Convention for Tobacco Control and prohibited smoking anywhere in an indoor public place. Finally, a federal law was implemented in 2015, which mandated that all states implement comprehensive bans. Here you can see this is quite different from the India Tobacco Control Policy you've looked at, which essentially involved one time point when the intervention was implemented. Here we have two different policies, partial and comprehensive smoke-free laws, being implemented over multiple time points, which lends itself to a very different study design. The chosen study design was panel regression modelling, which made use of local area mortality data. Panel regression models can be seen in some ways as a hybrid between [inaudible] through time series analysis and difference in difference. Because you have trend data which is a feature of an ITS, we also have localities serving as an intervention and control sites at different points in your study period so this lends itself to a DID approach. Again, this approach was developed within the field of economics, and was not very commonly used within public health research, although this is now changing. Our results are very interesting and important. The model produces three results: the underlying trend in infant and neonatal mortality rates, the step change, which is the initial impact of the policy, and the slope or trend change, which is the extent to which the policy has an ongoing impact. I'll focus on the step change results highlighted here. For infant mortality, partial laws were associated with a 0.5 percent reduction. Whereas comprehensive laws were associated with a higher 0.7 percent reduction. For neonatal mortality rate, partial laws were not associated with a significant reduction, but comprehensive laws were associated with a 0.3 percent reduction. Importantly, these models laid to predict the number of infant deaths that could be averted due to laws. This tells us how many additional lives could have been saved if all states in Brazil adopted comprehensive laws rather than partial laws first. Our findings suggest that nearly 11,000 infant deaths could have been averted if comprehensive laws were implemented. This provides a simple and very powerful message to policymakers in other countries that inaction on tobacco control measures is costing lives. This is especially important because the tobacco industry has lobbied governments very hard to introduce partial smoke-free laws rather than comprehensive ones. Now one of the key reasons that we focused our study on children, is that politicians are far more likely to listen and change policy when they're presented with clear evidence that their inaction is harming child health. We and others are using this data to try and frame tobacco control as a child rights issue. That is that all children have the right to grow up without being exposed to the harmful effects of secondhand smoke.