 integrated environmental and health impact assessment. Francesco is an environmental epidemiologist working on the health effects, health pollution on the cardiorespiraries. It has been my boss for the last 20 years, so it's really a pleasure, so please, Francesco. So, good morning. Thanks to Carla for introducing me. And it's a pleasure to be here. Not only because it's a very nice place, nice view. I'm sorry, we're just going to wait for two minutes. We've just got to organise it before we could write the links. Sorry. No, it's not there. We've got no link. Yeah, but you've got the mass symbol just under it. Can we use the links, which was before? I'm sorry for that. The mass symbol, this one. Yeah, it would be fantastic. I didn't... I didn't think... It's stored by date. Weird. It's good. Okay. Good morning again. And I was saying, it's a pleasure for me to be... I was saying, it's a pleasure for me to be here not only because of the nice view of the sea, nice surroundings, but also because it's a good opportunity to share with people some scientific issue, discussion on some scientific issues, especially on health impact assessment and the way to go. So my presentation, I realise I have a lot of time, so this would be good because we can have a sort of interaction and question and answer and maybe discussion. So in my presentation, this is the outline, so I will have some historical notes on where health impact assessment is coming from, what are the current debates and use. I have an example on waste. I have examples on air pollution and specifically some health impact assessment that was conducted in Rome sometimes ago. So using this example, we might discuss about the application and we might discuss of the problems and limitations and advantages. One of the main issue I will cover in this presentation is the following. What's the difference between an epidemiological investigation and health impact assessment? And this is critical because especially people coming from epidemiology, they think that health impact assessment is an epidemiological investigation. It is not an epidemiological investigation and we will see this during the course of the presentation. So I think someone has already, probably Carla yesterday, has already provided information regarding this slide and this definition. This is a WHO definition in 1999 and it's actually coming from this quite important Gothenburg consensus paper where the WHO convened a meeting of a conference with several institutions on health impact assessment to define that and to have some guidelines on how to conduct health impact assessment. It's important to notice it's a combination of procedures. So it's not a discipline. It's a combination of procedures, different procedures. And if you Google health impact assessment, you have maybe 20, 25 different recommendations on how to do that. So it's not very clear and there is not one specific procedure. It's a combination of different approaches, methods and tools, but which a policy, a program or a project may be judged as to its potential health effects on a population. And this is also important and the distribution of the health effects within the population. Why do you think this WHO decided to have this last sentence? This is not only to evaluate the impact of the new program and projects or intervention on the old population but also on the distribution of the effects on the population. What's the meaning of this? Do you have an answer for this? Can you imagine? Have you heard about this problem of inequality? So the question is, is this impact equal to all the population or is this effect affecting more the poor people and people with more disadvantages or more vulnerable people? So in our impact assessment we want to know whether this procedure is affecting the population but also if there are differences within the population. So if some specific subgroups suffer or gain most from this intervention. So this is quite important because this concept was not present before. And that was 1999 and then in the World Health Organization in 2016, so only two years ago had a new guideline on how to perform and how to do health impact assessment. And from that guideline I decided to take these graphs which I think are relevant and the first graph from WHO is to define the policy question with respect to the health effects related to that specific exposure. So it's very important to define what is the policy questions and why this is important and sometimes the politicians they don't know what they want to do. So for instance, a policy question can be should I build a new plant for a new refinery plant in a specific area? What could be the health effects and the health impact of this new refinery plant? This is very simple policy question. You have a plant and you want to build politicians and stakeholders want to build a refinery plant. But you can specify better this policy question. I want to build a refinery plant which is able to process such amount of crude oil. So you want to specify the sites of the refinery plant and the sites of the refinery plant may be different. You may want to build a big one or you may want to build a small one. So it's a different policy question and you may decide to use one specific process for a refinery or you may decide to have another specific process one which is more protecting the environment and the other which is not protecting the environment. So it's important to say I want to build a new refinery plant which sites is the following with specifically this way of processing crude oil. So the policy question should be asked to the policy makers and to the proponent of the plant. And this is not trivial because we can not respond to the policy question if the policy question is not well specified. So we have to plan the health impact assessment and then start to respond to this question who is at risk define the population with a graphical scope of that specific health impact assessment. So for instance continuing on this refinery plant what do we want to do? To estimate the impact of this new refinery plant in an area of let's decide five kilometers from the plant. This is a good compromise because we think that the emissions and the concentration of pollutants will be affected only within the five kilometers radius from the plant. So we want to address the effect on the population living within five kilometers. That's the possibility. But you have somebody coming and saying okay you want to build the refinery plant that's a good work to be done but what about production of greenhouse gases? They don't stop at the five kilometers. So your question could be should I do the health impact assessment only for the local population or should I do the health impact assessment also for a much larger population affected from the greenhouse gases and potentially climate change. So it's important to describe the spatial resolution and what kind of pollutants I want to address. So for instance what kind of pollutants are coming out from a refinery plant? Yes, you are all experts of this after case the presentation we all know what kind of pollution. So question. Is ozone coming from a refinery plant? Yes or no? No, okay. Question. Is ozone concentration affected from the refinery plant? Yes. You are good people, very smart. So ozone is not a primary pollutant but it's a secondary pollutant. Okay. Question. Very simple. Is volatile organic compounds coming out from a refinery plant? Yes, of course. So you have all sorts of possible volatile organic compounds including benzene including toluene, xylene all this sort of things. You have NO2 NOx coming out you have some small amount of particles coming from refinery plants. According to the process you have many. So you have to decide what kind of pollutants should I decide to investigate and you have to make a choice because you can decide to address the most important primary pollutants which are the volatile organic compounds but also you may think that also contamination from particles are important and you may think that the formation of ozone compounds are important. So all this should be specified quite in advance. And what are the health effects? And of course you have to study a little bit. If you want to do an health impact assessment you cannot rely on your a priori knowledge. You have to evaluate what is the current literature on the potential health effects. So for instance you are thinking of volatile organic compounds. I don't know how many here have some medical backgrounds. How many of you are physicians or biologists? Okay. Quite a few. So what kind of health outcomes can be related to volatile organic compounds? Benzene for instance. Cancer. So hemopoietic cancer including leukemia is one. What kind of other pollutants? No. Isabel is very say again. Respiratory system. And what kind of effects of a respiratory system? Asthma has been suggested as being related to volatile organic compounds. Okay. So if you are looking at volatile organic compounds are you considering cardiovascular diseases in your health impact assessment? Probably not. Because there is no link up to now. Up to now. You never know. On benzene and cardiovascular disease. You are not evaluating a ischemic heart disease or myocardial infarction. But suppose a guy comes and says okay, this volatile organic compound. But what about particles? You want to assess also particles. So what kind of health effects you should evaluate? You have several others. So it's very important to decide and make some priority list because you cannot do everything, of course. So you have to make a sort of priority of what kind of health effects you are going to identify. So now, now WHO says you have done all the preparatory work. Then you choose a model or a tool. We will see there are several tools that are available for conducting this health impact assessment. Now, to conduct the health impact assessment you have to have here as an input to the model population data. Massimo has already covered this part. You have to have population data, air quality data and baseline health data. Without this data you cannot conduct the health impact assessment. So, you want to know what's the spatial distribution of the population living nearby? Suppose you want to address the issue of the cancer effect of benzene. So the guy providing, people providing the population data come to you and say okay, we don't have the age distribution of the population living nearby. We only have the total population count. So we know that in a five kilometers radio there are 20,000 people living, but we don't have the age distribution. Are you concerned about this? And why you are concerned? Of course you are concerned because benzene can cause leukemia but you are concerned of childhood leukemia. So you want to have population data divided by age because you want to estimate the excess of leukemia cases in the old population and in the children population. And why you want to address this in a different way? Do you have an answer for this? You want to estimate the number of leukemia cases in the nearby population and you want to specify the number of childhood leukemia and the number of adult leukemia. Why do you think this is important? Maybe you know maybe you don't know. I'll tell you. Is anyone having an idea? Please. Who was talking? It's important to repeat because we have people listen from the webinar. I mean that leukemia concerned with adult results from long-term eczema to benzene but childhood leukemia may be concerned to short-term eczema. It's a very nice response also because clarify the time lag between the exposure and the disease in the case of leukemia is different in children and in adults. This was quite important experiment in science which was the atomic bomb in Nagasaki and Hiroshima. I call this the scientific experiment natural experiment. The observation of the population showed that the lag time the latency period between the explosion and the childhood leukemia was much shorter than the time lag between the exposure and the adult leukemia. The time lag is different. If you have a new exposure you will see childhood leukemia sooner than adult leukemia which will come later. You want to address this in a different way. Just going back you want population data if it is possible by age by gender maybe. You want to have this wonderful way of predicting the exposure of the population the case, the org elegant and very interesting presentation. You want population exposure and of course you want to have health data. You want to know what is the background level of leukemia in the population. What is the natural background without that exposure what kind of incidence rates do we have. You have to have this data you see this graph is wrong because it says the same thing twice. Difference between I think it's wrong because it says you have to provide input for the tool of health data and it's just saying the same thing here. So WHO is also wrong. Fantastic. Okay. On the right side my right side you have just the data of the population. So maybe the other one is more general and the other if you want really to focus on your case study I don't know, I just have this interpretation. Yeah, okay, yes. But it's not very well described I mean if in three of us can I reach a conclusion? I mean I agree with you the fact is that what you need for your HIA is having three kinds of data and on the right you have data that are baseline so a reference in some way but obviously it could have been said different. I guess there is also a long test to explain this table if we can go. Okay, so by the way you need health data. You need health data for the population either recorded or estimated. This is something I have to tell this story. I just went to Nigeria a few months ago and the idea is to do health input assessment for different sources of pollutants in Lagos city. Lagos is a city of 18 millions people very high air pollution levels coming from different sources. Yeah, you can imagine traffic you know old vehicles there is a huge hybrid there, there is burning of waste there are electricity generators, diesel so lots of different sources so the minister of environment there is doing the source so very nice part from the monitoring side. And we decide the government is deciding new intervention especially to reduce the electricity generators which are thousands and thousands polluting. So the question was we should do health input assessment and see what is the result of this intervention and so then we were asking what kind of health data do we have here to use for the health input assessment and this is probably true for other countries in the world so we were expecting to have mortality data and we got the response. Of course we have the mortality data but only for people dying in the hospital and we said what's the proportion of people dying in the hospital so maybe 40% and what about the others okay, you can die at home of course and what about the death certificate or we don't use that and why we don't use we only use if people dying have a legal like I don't know in English property succession you live your house then you need the death certificate for this transaction but if you don't have anything for the cultural people you should be buried in 24 hours so for most of the population you don't have the medical death certificate and this is widespread so we realize that Nigeria have no mortality data and say how this is possible we read on the WHO report we have the mortality data from all over the world including Nigeria so what is providing this data we went to WHO and they told us you didn't read carefully the report why because Nigeria is not reporting and how can you provide this data it's just imputing the data from other countries so basically it's an estimate of mortality in nearby countries where they report the data so in that specific case of Laidos we don't have local data we don't have national data we have the WHO estimates there and we don't know how far are those estimates from the real data many there are countries in Africa who are reporting so I'm not an expert but this is a problem in the world WHO score the different countries and there are completely no reporting and there are countries reporting some data and of course there are countries like France very good so there is a process of estimating the mortality data so going back to the question of Masimo Staffordja had before we think that the health data are available worldwide this is not true we have to rely on some specific places in some estimation but just going back to this Nigeria example so we said ok we don't have the mortality data but do you have any data on incidents you can imagine in Africa in several places incidents of lower tract respiratory conditions is very frequent especially in children pneumonia in children do you have data on pneumonia in the population yes because we regularly conduct a sort of health interview survey so in some years not every year but some year we go to the families on samples on a sample of the population and we ask in the last year how many episodes of respiratory infection your child had and what kind of respiratory infection so we have estimates of pneumonia in children so this is the paradox so you don't have the mortality data but you have quite good morbidity data so just to finish this so you have this ingredient for making your cake and then you run with population estimate of exposure and your exposure estimate and then you use the concentration response function this is something that we will see I will present tomorrow it's basically the dose response relating the pollutant that we are considering to the health effects so basically if I want to estimate what is the potential impact of this new refinery plan on child leukemia in the area what kind of health function should I take do I have the health function from the local data? Probably not so I should go in the scientific literature and see whether there is any systematic review and probably Carla will speak about systematic review and meta-analysis providing with a number what is this number this number is just telling you what is the percentage increase of child leukemia in relation to exposure to benzene so my problem is that I have to borrow this number from somewhere and of course if I borrow this number I have a lot of problems because I'm not taking this number from the local population I'm taking this number from the scientific literature ok and we will speak about this so I will estimate the impact and then after estimating the impact I have a response to the policy question and then when I provide a response I have another question before going to the policy maker and my question would be what about the uncertainty on my estimation ok so and this is something Andrea Ranze will speak about so what is the uncertainty in what I'm doing and where are the sources of uncertainties we have sources of uncertainties here estimation data maybe they are coming from a census so this means they have very good quality maybe they are coming from estimation so they are less certain the air quality data maybe they are coming from some monitoring so we have actual monitors but if the refinery plant is not existing we don't have monitors so we have to borrow the data from somewhere so there are uncertainties related to this aspect and of course suppose the example I was doing on using the incidence data of plemonia in children from Nigeria this incidence data are based on random sample of the population so it's a sample it's not the old population so they come with uncertainty and let's go here have this concentration response function are we sure about this concentration response function no because it comes with uncertainty so all these uncertainties will come up and I have to tell to my policy makers I'm estimating three more cases per year just giving a number three more cases per year of leukemia in the nearby population but these three new cases they come this estimation comes with uncertainty and this uncertainty is such and such so maybe from 0 to 6 or maybe from 0 to 50 to the policy makers that I am not sure how much I am sure about my prediction ok so let's see the current applications of this health input assessment we have several applications in the last 2025 years the first one which is an easy one is the estimation of the burden of disease and this is a nice example that we will see is the comparative risk assessment that the global burden of disease does so it's comparing different risk factors and tell to the people what is the ranking of the risk factors in the world so the global burden of disease comes with a ranking where you have on the top tobacco smoking low physical activity alcohol consumption all the big risk factors so you can do a comparative risk assessment you can do a scenario comparison and the scenario comparison is very much the sample I was doing before what's the effect of a scenario where we have a refinery plant in that area in a situation where we don't have the refinery plant and of course I can do we can do a cost benefit analysis so to do an evaluation what are the actual health costs on the population but what are the benefits of this intervention so not only cost health cost maybe due to interventions but also benefits and I could do a balance between the two okay this is the example of the global of disease I think Carla you already have shown this maybe why I'm showing this because this is the first one was published in the Lancet in 2012 and this is related to 2010 was ranking the various risk factors in the world responsible for the total mortality and you see high blood pressure is up there but it is also in this graph also ambient air pollution was predicted as one of the most important risk factors in the world so why this is comparative risk assessment because I'm comparing you know various factors this is a comparison of various factors this is an example coming from France I think where the authors here in this publication they compare the actual legislation the current legislation regarding the emission versus the legislation with maximum feasible reduction of emission and you can see here they did this comparison in Paris in Paris suburbs and in a rural area of nearby Paris and you see that that's the effect of the current legislation and this is the effect decreasing the impact of the maximum feasible reduction so this is a scenario comparison and this is for the European Union when the European Union approved in 2013 the so called air pollution package which was a package to reduce the emission of pollutants in Europe they did the cost benefit analysis of that specific policy and this was published in 2015 just examples of application so just reviewing the critical steps in what I was presenting so the critical steps are estimation of exposure something which I didn't address but this is quite important it's called the counterfactual value this is of course not important when you are comparing to different scenarios so if you are comparing a scenario where you have a new plant versus a known plant you don't have a counterfactual because the counterfactual is that you are not building the new plant but suppose you want to estimate the burden of disease this is a question for you so suppose your government asks you to estimate what's the burden of air pollution in your specific country and your specific country comes with an average of pm2.5 let's say of 25 micrograms cubic meter of pm2.5 and you have all the ingredients you have the concentration response function so you can easily estimate the impact of air pollution in your country but your question would be where I start to count the effect of air pollution which level should I start from zero because someone can tell you we don't accept any air pollution level so zero is the maximum amount I would tolerate this green Taliban so it's a very extremist point of view so zero so you have someone responding to your question we should start from zero but you have someone saying you cannot start from zero because you have some natural background so dust exist in nature so you cannot arrive to zero concentration you should arrive to some natural background and so you have people saying the natural background depending on the area could be 2.5 micrograms could be meter maybe 5 maybe in that specific area maybe 7 so you have to decide which is the level you start counting the impact and then you have another guy saying okay you are speaking about the natural background but we have WHO saying that in air quality guidelines 10 micrograms could be meter is the safe value why don't we start at 10 micrograms so you have several options and these options are called the counterfactual so it's the situation where you don't have the exposure you want to count as an effect so you have to choose you can choose zero you can choose 2.5 you can choose 5 you can choose 10 but each has a different impact on your impact assessment okay so it's a big difference of starting at zero or starting at 10 because if you assume that air pollution as an effect for each 10 micrograms could be meter 2.5 you have an increase in mortality of let's say 6% the country has 25 one thing is estimating the difference between 25 and 10 which is 15 and another thing is estimating for all the range of the pollutants which means from 0 to 25 you see it's a 6% difference which is a lot in terms of mortality estimate so is this clear this counterfactual so my question to you if you have to do the health impact assessment the burden of disease of air pollution in your country which counterfactual would you choose 0, 2.5, 5 or 10 this is not trivial so let's open this discussion what 4.2 just 4.2 why 4.2 you have a scientific explanation what do you think should we start from 0 or should we start from 10 if you want to compare the burden of disease compared to other countries maybe 10 is accepted all around the world so it can produce data that can be comparable in the other part of the world if you want to calculate the maximal possible effects I would say use 0 as worst case scenario so ok that's very good very good response I think is any objection to this response done ok ok so this is a review we already have seen we need the exposure response function and I will explain in a better way tomorrow morning we have to report and here is the question what should I report so some people are saying I should report number of excess death or premature death other people are saying I should report cases of disease some people are reporting years of life lost others are reporting disability adjusted life fear of change in life expectancy so which one do you like in the last 10 years we had this very long discussion about this options and when we are reporting premature death sometime people forgot to say premature death and just say air pollution is causing 4.2 millions death worldwide actually it's causing 4.2 million premature death worldwide why because people are saying we are dying anyway so what we are speaking about we are speaking about anticipating the date of death so it's a message which is sometime confusing people so that's why people are using this years of life lost but if I tell you air pollution is responsible 10 billions years of life lost can you translate this information into something you understand no even worse if I say air pollution is causing 80 billions disability adjusted life so the main problem that we have in this communication of the impact is that these numbers are difficult to get sometime people is using change in life expectancy so it's saying okay air pollution is causing in this country one year shift in survival which means that people are living one as an average one year left or maybe six months less what do you think do you like this information you know time you are you are surviving less time what do you think it's better than number of death or it's worse what kind of information do you like yeah sure to see that you are all scientists or scientists living here and that's a question for you but for us that's a huge question we can't as you said 80 billions years lost something we don't use it because we don't know what it means we can because it makes big 80 billions it must be important you know that's a question we got as a journalist so if you scientists could just have a good answer we'd be really happy so as a matter of fact we don't have a clear response to this question what is the best form of risk communication so nowadays there are two most used numbers is the number of premature death because worldwide it's easy to communicate so the problem is leading especially politicians always say we should die sooner or later so don't bother me that's the usual response from the policy maker and from the policy maker if you say 3 million so you say 4 million it's the same it's a big number so we have problems in communicating here we have problems also in communicating this change in life air pollution is deprivating people of one year life maybe it's good but you have another tendency the main message that we get is that people are living longer so we have the message because of the better life condition because of much more resources because of the medical technologies we live longer and actually the aging of the population is one of the issue so if you say you know this is causing you to live one year less people say I'm aging in any way so why should I bother I would be old enough instead of dying at 96 people die 95 it's basically the same so all this communication reports are causing problems in communications and we are waiting for journalists to give us the answer wait yes say again I asked if you could put it in economic terms because policy makers seem to be more interested in economic terms because they are in environmental issues or in health issues so if you could explain the decrease in productivity things like that yes but then my question is how much to evaluate a life now I was thinking about how much do you evaluate a life because I usually use the years of life or sometimes better the days of life lost for one year each year because for the years of life lost you for example in Italy you consider the average life expectancy so if you use the days of life lost each year is probably you can compare also between different nations that has different life expectancy this was about the years of life lost but usually the value that I found in some report that is now quite old is like the the cost of the productivity like the GDP that you give to a life for the 2000 for Europe it was 50,000 euros for each life lost to stress what Cara was saying two years ago we computed estimated the cost of one year of air pollution in the case only of respiratory diseases three billions so which is 30% of the deficit of the health security in France only for the respiratory what is evident because we know that the inhalation is bad if you remember yesterday I showed the work of Francesca and I that people made with the American Terrasso Society showing that the effect of air pollution are beyond the longer they are many so and we use the tangible cost we went to the social security trying with the through a tributary risk saying 3% is due to pollution asthma cost that so this is clearly an underestimation but I went to the senate to the assembly so I think this is really an argument good for the journalist and also for the policy maker sorry here I find it very interesting that we have a discussion here saying journalists say please tell us how we should speak about that scientists say please journalists tell us how we should bring this to the news and to the people and I think that HIA are very good case study I think for what I think should become the relation between science and society because you've showed that at the beginning of the process there is a political question that should be asked and tailored in the right way and then the science comes in and make some assessment and then science brings the communication back okay and so I think all the process should be a discussion between politics or stakeholders and I think politics is not the only stakeholders that should be in and then we should find a way tailor the question find the right solution monitor the solution and maybe choose the words we use and the way we communicate together so that everybody understands well and that finally we're credible, silent and legitimate I agree and this is really a scientific process it's not the epidemiological one but it's a scientific problem we need the protocol we need the policy question we need high a priori high potency we need the statistical plan everything need to be standardized at the very beginning because what we would like to have is a result that can be reply somewhere else so it's a tool easy to implement I mean in order to have the data of course but it's quite easy we will see this afternoon an example a practical example because if you have this standardization of the methods of the methodology at the beginning we can do the same in several countries and compare okay so I think we had several good inputs to this discussion so we had the economical evaluation as a possibility we had also the suggestion of days of life lost within a year to make comparisons across various risk factors in various areas I think we had a very good suggestion that health input assessment is something which is very much related to the policy questions and the way we respond to the policy questions so it's a communication problem. There's also other suggestions instead of reporting the crude number to report the number as a percentage of in case that the percentage of total mortality which is of course a way to produce good information. As a matter of fact there is a very nice website which is called breath life it's a doublet show website for communication on the risk of a pollution and breath life and it's reporting the number as 1 in 10 death is due to your pollution or 3 out of 4 so it's a kind of percentage but is a very simple way of expressing the numbers. Okay so just to finish this slide so of course evaluate the uncertainties involved in the process is part of the risk communication we have to report in some way what is our level of confidence on the work we have done. So let's continue and I continue when I say rewind let's go back to the historical development and I think Carla has already done this why I rewind starting from this because I start back from epidemiology this is a very useful graph from epidemiology in epidemiology you have very nice effect estimates one is the relative risk of course and the other one is the attributable risk so it's you know year one course one of epidemiology relative risk attributable risk and so what we estimate is the incidence of disease in unexposed group versus the incidence of disease in exposed group we make the difference and we say this is the attributable risk it's very easy from the epidemiological perspective and as a matter of fact I really recommend you to go to this paper which I think was I forgot this was in epidemiology in maybe this year is an epidemiology in 2007 and is an overview of methods for calculating the risk of disease and I like very much this formula here I always present so what is the attributable risk of attributable fraction among the exposed group and if you go to this part which is the relative risk minus one divided the relative risk so if I ask you suppose you have a relative risk of 1.1 so 10% increase so what is the attributable fraction in the exposed group so it's very easy to do 1.1 minus one divided 1.1 which is basically almost 10% or maybe 9% so it's very easy to say if you have a relative risk of 1.1 so even as very small relative risk in the exposed group the attributable fraction is 10% so why I say to use this formula here instead of the one in the population because in the population of course you have to wait for the proportion of people exposed but if in the population we are all exposed then it's much easier so suppose we are all exposed to air pollution which cause a relative risk of 1.1 this means that the attributable fraction is 10% so this is easy from the epidemiological perspective what is not easy is what Steele and Namsung already said in this paper one of the issue is this relative risk which is an issue of portability of the parameter so usually we estimate the relative risk in the study population in epidemiology studies is very easy to estimate the attributable risk because we are deriving the relative risk from that population but in healthy assessment we are not using this relative risk in the same study population we are exporting that relative risk we tend to say this is an issue of portability because we transfer the information from one study we transfer to another study and this issue of portability is like the issue of the external validity in epidemiologic studies so we find for instance that in our specific study let's think to the British doctor the famous study on British doctor assessing the relationship between smoking and lung cancer you know that study indicated that the relative risk was 10 no 10 so 10 fold difference between those smoking and those without smoking so the issue was has been how portable is that relative risk how can we generalize from the British doctor to the entire world population and this is the same for the health impact assessment how can we generalize and this is the issue of the concentration response function so everything starts from epidemiology but then from epidemiology the US government decided so please just because we had some discussion during coffee if you go back you said now that how can what about the generalization of the results but we are in the opposite we have to we have to import another risk coming from another population so the question will be how is generable to mine population I'm correct because I'm using data collected in other populations so relative risk coming from if I don't have my own big cohort available so that's the way if I can choose what in your opinion will be the best starting from proximity if starting from which is more available maybe from Groenland and I live in Iran I have a solution for this tomorrow but just I have a solution for this partial solution but just to address in more specific terms this problem of portability because you will see and you have read about this global estimates of the impact of air pollution worldwide and these are based on studies conducted in Europe and North America so when we as WSU we presented these results to some countries like India they may say or you are using relative risk coming from Europe and North America how can you be sure that this relative risk are applicable to our population what is the problem of adaptation of the relative risk to a specific population because you have racial differences for instance you know you probably know that the Japanese have a different relationship between smoking and lung cancer than Caucasian people so people from India may say the relative risk you are importing here you are making a trade you are importing your American relative risk to India and this is not appropriate and this is something we have to discuss but suppose you go to Dua or you go to Dubai and you are predicting the air pollution health impact in quite or you know one of the South Arabian countries and you say according to my calculation these are relative risk coming from Europe and North America but we are using the satellite estimation of air pollution in your country indicating that the PM2.5 concentration is so high and the guy comes and says our PM2.5 is 40% from anthropogenic sources and 60% from desert dust so you are taking the relative risk that comes from studies on PM from anthropogenic dust on the assumption that anthropogenic dust and desert dust have the same toxicity can you prove this because if you cannot prove your estimation is completely wrong so what would you vote for the Dubai person for the European expert the Dubai the Dubai person may be right because we are exporting or importing a relative risk which is based on studies conducted with some specific condition so the population is different but also the exposure may be different so this is one important assumption in this portability issue and this is of course for all the health impact assessment because one of the main problem we have is this portability and of course health impact assessment is very much linked to this portability because without this portability we don't do the health impact assessment okay this was very strange for the Americans because this is an editorial that came in the American Journal of Epidemiology in 1998 so before one year before the World Health Organization statement on health impact assessment and this is called the red book because the EPA in the United States decided to have an instrument for doing the health impact assessment with the risk assessment at that period John Summit says while epidemiologists and epidemiology data may have prominent roles in this field the EPI literature contains surprisingly few discussion on risk assessment so risk assessment and health impact assessment was basically conceived away from epidemiology and as a matter of fact epidemiology was out of this discussion because it was very much based on toxicology you may know that risk assessment now we use risk assessment and health impact assessment maybe Carla have been already speaking regarding this in a sort of same way but basically risk assessment was born in the toxicology lab so it was a way to extrapolate from animal studies to humans ok so and we rely that people are relying a lot on the toxicology so question to you we said that the human the portability is one of the issue of health impact assessment and my question to you do you think that portability is also an issue of the toxicology assessment of course so this portability is not only a problem of human studies is also a problem of animal studies because what we had in the past was very much based on toxicological approaches so what Summit is saying at that period epidemiologist look you have to deal with this risk assessment because up to now this has been the field of toxicology but we have a lot from the epidemiologic experience to learn and to implement in this part it's quite important because this red book had this paradigm of different steps in the risk assessment one was the other identification the dose response assessment the exposure assessment and the risk characterization it's a different order nowadays we see the exposure assessment before the dose response but it's basically the same so so what Summit is saying here we want the participation of epidemiologists so we want the participation of human data but it's very clear that epidemiologists are not well prepared to this field so that's why we are very much advocating the role of epidemiology here okay let's keep here so let's go to this discussion of health impact assessment of course this we already said this it's important to say that although health impact assessment is done within the health area especially known health people are very much interested in this because they are the proponent of the new project this is something we already said so planning a new motorway in this case when we are assessing this is quite important when we are assessing the new project we are not assessing only one specific exposure so suppose you are evaluating a new motorway it's not only the problem of air pollution but we have also other specific factors like noise or light pollution of accidents you know several different part and especially people might be interested in this physical pollutants but people might be interested in this could be much more important how this motorway will affect business and sometime if we respond to our mission of health impact assessment only with a response regarding the health impact but without considering other socio-economical part this would be seen as only a partial response I have a story here just to make it clear so in 2005 we had the new law in Italy that we called the smoking ban so this was applied in several countries in Europe but Scotland was the first and believe it or not Italy was the second believe it or not so this law was approved and I have to say this was approved during the Berlusconi government believe it or not was prepared by the other government but the Berlusconi government was the first one who signed this law the law was a 2003 law and said that in 2005 the law had to be under operation so there was a specific day which was January 10th I think or 2005 so it was one day shift, smoking ban everywhere, all the coffees and restaurants and so on so a few months before people were just questioning how this law could be what could be the compliance of this law in Italy what do you believe this law would be a complete disaster because no one would obey this law how can you imagine the Italians would follow these prescriptions this is complete nonsense and we were wrong why we were wrong because we were not considering this part the business part so owners of the restaurants and the coffee shops they were at the beginning they were protesting they were saying no this is we're ruined our business we want smoking inside because it's we're like a guest of the restaurants will go away but then they realized few days after that if you have smoking prohibited inside several people smoking outside and if you have several people smoking outside you make a crowd close to the restaurants of the coffee and if you have a crowd close to the coffee you attract people especially Italians are very attracted when they see a crowd so let's go there because there's a lot of people there should be good so this was believe me this was a strong effect for compliance of the law because the first one was not defined you know the restaurant owner they had to pay a fine if they had smoking in a close place but was much more important the crowd effect so the business and in doing quick and dirty health assessment we didn't believe this so we didn't consider that business was one of the most motivating aspect of the success of the smoking ban I don't know the other country this is something that I don't remember in France when was we had the same bill of course in France after no I'm sure I was surprised for Italy but in France you've got many cafes people smoking of course and the cafes doubled the terraces the terraces were opened now they're closed with another problem because they're just making the heating outside but they doubled the space to serve the people because people smoking just stay outside in a covered place which is a bit outside inside and it's outside and people not smoking inside really so they've got a bigger business now yeah okay thank you okay so I want to you know we already made 20 minutes so let's go to one example I want to present to you then you will have the slide I want to present to you one example of application because otherwise you're always speaking of theory let's do some work so this is a work we did within a project a European project long ago now 10 years ago on waste assessment so why waste assessment this is not air pollution this is the assessment of the process of treating waste in specific cities in specific countries and so when we did this assessment we have tried to do in a sort of integrated way so we said to ourselves let's look all the full chain from waste production to health effects and this is important because you can see all the different aspects we would see this in detail all the different aspects of you know producing waste management waste and you know producing health effect let's start here so the waste production is very much dependent on the society decisions so you have waste production from from all sectors not only from households and you have various possibilities to manage waste you know waste management start from you know collection and transport recycling you have plants with mechanical and biological treatment you have industrial process like gasification pyrolysis you have all styles incinerators you have landfills and of course you have also illegal dumping and burning so you have all sort of possible management and when you decide your policy about how we manage the waste you have to specify what you want as a policy maker of course all this process may have some emissions and you have here a list of potential emissions including of course some gases that may have a global warming effect but also some dioxin think about illegal incineration you have a lot of dioxin coming out so you have a lot of different possible exposure and they go through air water and soil and they can go also through the food chain so it's very complex and you have exposure of the people from inhalation ingestion and dermal contact and you have a list of possible health effects and you can calculate the impact as excess cases, dollars and costs and of course in making the impact you have to consider the vulnerability of people not only by age and gender by previous health status, the lifestyle behavior and socioeconomic status so this is quite complex. I took one year to develop this graph so it's not an easy graph so it's quite complex so we have been trying to address all the issues in this exercise part of this has been published, this was published in 2011 I think in environmental health and we did this assessment in three European countries and this is the effect this is quite interesting is the effect of exposure to landfills you may know that the European Union decided that landfills with I call crude waste so untreated waste has been banned in the European Union so you cannot landfill with untreated waste according to the current law this was not the case in 2008 when we did this assessment so at that time we said in each country we have an annual number of cases of congenital malformation and a number of low birth weight coming from exposure to this landfill so this was the impact you may think that this impact in comparison to the if you make this in terms of proportion of the total number of birth this is a trivial number because you think for instance Italy we have we had in the past 500,000 new birth in a year so 700 is a teeny proportion of this 500,000 this was 2008 nowadays believe it or not you know how many new births we have per year 400,000 so in 10 years we decreased the number of births it's a huge decline so in terms of proportion it's a teeny it's a small proportion that comes with some uncertainty of course and this was an easy application of the health input assessment why it's easy because you have the exposure you know the birth is occurring the same year so there is no time lag but of course when you want to estimate the number the excess cases of cancer this is the years of light lost due to exposure to PM10 and NO2 you know the number of cases is small here but when we went to the issue of cancer we had to go back to this slide which I think is very interesting slide I'll try to explain to you suppose you have an intervention here at time 0 but the exposure to this specific exposure condition started in the past suppose you are evaluating the effect of incinerators so the incinerators are old so they started to accumulate cases from the past and then at a certain point in time you decide to make an action like closing the incinerator you still you will have some new cases which are related to the exposure of the past so in the future you still have some cases related to past exposure but of course if you don't do this action of closing the plans you will have this huge increase in the cases so the number of cases avoided this part this big part but you still have some cases that will be occurring in any case so we applied this for the incinerators so this is the estimation of the additional cancer cases near incinerator in Italy due to exposure before 2001 we called that past exposure and during a period 2001 to 2020 you see that the large number of cases are due to the past exposure and only a small fraction is due to the current exposure so what we are saying is that there is some background number of cases that is occurring in any case and they will be occurring in the future and of course this decline is due to the closure of the incinerator so this exercise was very interesting to consider in our health input assessment not only the cross sectional design but also the longitudinal design you don't want only to assess the effect in a specific year you want to assess the effect in a very long time span do you have a question regarding this this is a difficult part so maybe you have comments okay otherwise yeah it was a question from Samane and she was asking in all the assessment we don't use fertility data, tables and the effect of pollutants and incoming newborns and how their mothers are exposed to the pollution so she was just wondering if it wasn't concerned in health impact assessment yeah of course fertility is one of the indicators unfortunately the database on the effect of some chemicals and some pollutants on fertility is not huge so when we have the issue of you know those response relationship we are in the field of uncertainties of what kind of estimation take but of course fertility is one of the issue because it's very sensitive in the population very sensitive issues so of course congenital malformation is an issue for reproductive health but also fertility so it's a very good point so we have been continuing this exercise on waste management and this was very complex exercise but it's complex as the issue is complex so we did this just to show the complexity of the issue of waste management so at that time in 2008 we had this baseline and we compare this baseline with a way strategy that the regional government decided to apply to the waste management in the the Lazio region is where Rome is located so this was a specific proposal from the government unfortunately the government did not last to this point but you know we did this exercise so this was a proposal for the government and in addition we said okay this policy makers they are green they want to do green but let's be extremely green and let's invent a policy which is more radical than the one that the government has already decided so the baseline was was any waste prevention policy in Lazio the baseline was no the government recommended the waste prevention policy especially for the industrial waste and packing in the supermarket you know one of the main issue of waste is the packing that you get your cheese and you get your cheese with a lot of cartoons and so on so you want to avoid that because that's a waste so the waste prevention was recommended but as in Italy if you recommend something and you don't enforce that is a lost battle so we said if we want to be green we recommend and enforce so we have to have a sort of aim of decreasing waste production 15% over the baseline so we had recycling at that time 17% the government said we should go up to almost 60% and we said let's do 70% the waste collection was mostly by beans and trucks if you have been to Rome you have beans everywhere you have these big trucks collecting from the beans and the government said okay let's go to a mix system where we have both beans and trucks but also door to door collection so people going to the specific households and collect the garbage and we said let's be green and mostly do door to door then what about recycling we had a straight collection of glass and paper the government said door to door collection and what we said let's have door to door collection but also centralized collection of specific recycling centers this was also a very big change diesel trucks were used for collecting waste and diesel trucks were very polluting the environment and only as more proportion was collected by train so the government said let's have electric vehicles and we said let's have electric vehicles but let's improve also the train used for collecting waste mechanical and biological treatment for 30% to 100% and this was a big change landfill without pretreatment was in 2008 70% of garbage was just to landfill without pretreatment so the government said according to European law 0% we also said 0% we increased the number of waste management facilities the government said let's have what is this so the government said we had in the past two incinerators the government said four incinerators and we said two incinerators okay an occupational program improved occupational program so this was the different scenarios of course the only real one was the baseline this was a real one because was demanded by the policy makers and this was completely invented by the investigators so this was quite important achievement of the three scenarios this is the baseline you see the trucks circulating all over the city and in our in the government plan you had much less traffic from trucks and in our green policy this was decreased again in terms of occupational injuries this was one part of the assessment you know we had quite high occupational injuries at baseline but surprise surprise we realized that with the government strategy because we had door to door collection so the number of occupational injuries actually increased rather than decreased for some reason I forgot why with our green policy we were you know decreasing that but the most important was this one this is we decide to have the evaluate the impact as disability adjusted life years and you have this baseline scenario and you from this baseline you have a large part due to transport then a good part from mechanical and biological treatment actually a small part from incinerators and a large part from landfills with the way strategy from the government we actually decreased a lot the contribution from transport whereas from landfills still we had some large impact and in our green policy we were able also to decrease the impact from landfills so you see this was a nice way to synthesize to make a synthesis of all the possible impact and at this stage we took the dali why we took the dali why the disability adjusted life years was the the estimate of choice because you can use different outcomes and convert them into a single number I'm just finished yeah into a single number that's why this was an easy way to finish so I have maybe the last last last slide you can keep this this was I was challenged by the one hour 15 okay okay we yeah yeah okay yeah I was scared so health impact assessment is long tradition not always well appreciated especially from an epidemiologist and one of the slide that I skipped was a discussion that David Savitz was a famous epidemiologist made was an editorial in epidemiology last year just claiming that epidemiologic studies in many cases are not useful are even dangerous if we can respond to the policy question with health impact assessment so if you can respond to with health impact assessment don't do epidemiological studies because they may be too long to provide answers and the response may be less uncertain than applying the health impact assessment health impact assessment can be conduct to address several complex issues and waste management is one of the complex issues it's very complicated and very urgent for policy makers health impact assessment is quicker than local epidemiological studies and the last stakeholders involvement is difficult why is difficult because stakeholders have different views and these different views are many times very conflicting if you have the proponent of a plant of a project and you have a community people you are sure that you will see people fighting so the most important part is how to do the stakeholder involvement in our health impact assessment and we tend to rely on journalists on what they say with this I finish thank you