 Bloodletting is a treatment for fever. Transorbital lobotomy is a treatment for schizophrenia, being rubbed with a live chicken for the bubonic plague and having holes bored in your head for epilepsy. Humanity is virtually limitless in the creative, often harmful and ineffective solutions it comes up with for problems. But what is most important is which ones we select to continue with. Medicine was largely ineffective until we learned to count, that is to use quantitative data to select which interventions worked and which didn't. So these quantitative data are actually what needs to guide us into the future. We actually see that medicine has come a long way and health has come a long way since those times. But what is absolutely certain is that in a hundred years time people will look back at some of the things that we do now with horror. So how do we avoid the mistakes of the past and how do we ensure in advance that what we do is beneficial? The simple answer to this is that we cannot know it with certainty. What we actually need is systems to test empirically which interventions work and which are safe. And we need to balance that investment across established treatments known to save millions of lives like vaccination and smoking cessation with the hope of research and genetics, those kinds of things. And the fact remains that the gap between our intervention and evaluation and response contains our capacity to harm. Important insights into this come from the incident relating to thalidomide but also from smoking worldwide. Thalidomide was introduced in the late 1950s and hailed as a wonder drug for treatment for nausea in pregnancy. During this time doctors were convinced that medications could not cross the placenta and affect the unborn child. It was licensed in over 46 countries and it was used by hundreds of thousands of women. It was hailed as safe and effective. An estimated 10,000 children were born with the characteristic shortened limbs and abnormalities relating to thalidomide. This drug illustrates the harm that can happen from exposure to these types of drugs. The harm that was observed was very soon after exposure and it was also established fairly quickly that thalidomide was the cause. So the drug was withdrawn for treatment during pregnancy fairly rapidly and a lot of systems for drug regulation were actually created in response to thalidomide. And thalidomide could have been much worse. It was actually never licensed in the United States and that's because the United States had stronger drug regulation systems and also because Dr Francis Kelsey a brilliant drug evaluator actually unconvinced of its safety blocked its approval despite six attempts by the manufacturer. So thalidomide illustrates that gap between intervention, evaluation and response and the harm contained with that in this case these stricken families and affected babies. But understanding harm is not enough. In the case of smoking we've known since the 1950s and 60s that smoking is harmful to health but we still have around a billion smokers worldwide. And up to two thirds of smokers in countries such as Australia, US and UK can expect to die from their habit. We have around 6.3 million deaths a year and that's up to 10% of deaths worldwide are attributable to smoking. We expect around a billion deaths attributable to smoking in the 20th century. We had around 100 million deaths in the 20th century. This is a major implementation gap. It's not just about our understanding. So intervention without evaluation and without action is the most harmful and costly thing known to humanity. When we intervene and we understand after evaluation that an intervention is harmful we must cease that intervention but we also need to use data to look at interventions which are beneficial and to ensure they are scaled up because often the harms are even greater there when you think about clean water and smoking cessation, blood pressure lowering. And we have better data and systems now than ever before. So we're really running out of excuses. And this applies not only to health but into any other area where we act or choose not to act. We have to ensure that we have systems that can evaluate, not just innovate but evaluate and respond. So there are probably harms as stark as the lidomide happening in places where we have systems that are underdeveloped and there are probably interventions as ineffective as being rubbed with a chicken as well. The way to avoid repeating mistakes in the past is to ensure that we have systems that can listen and learn and take action to ensure that innovation is inextricable from evaluation and action as rapidly as possible to ensure that that gap, our capacity to harm is as small as possible. Thank you.