 Continuing with the Italian speaker theme, we have Francesco who's the epidemiologist with MSF in Belgium and was previously here in the Manson and so on, but Francesco, so over to you. Okay, good afternoon. I'm going to present a preliminary analysis of the affected or patient affected with Ebola that were admitted to MSF Management Centre during the current West Africa Ebola outbreak. So the objective of this presentation is first to provide an overall description of Ebola confirmed patients admitted to MSF centres and secondly to present a preliminary exploration of risk factors associated with this. So since the beginning of the epidemic, MSF opened nine centres at a different point in time, two in Guinea, in Gekedu and in the capital Conakry, five in Sierra Leone in Kailahun, Makburaka and two centres in the capital Freetown and two in Liberia in Foya, close to the border with Guinea and in the capital Morrovia. So other small and temporary centres were open for short periods, but that are not presented, are not being pulled in this analysis. So here is a summary table of patients admitted and confirmed positive by an Ebola PCR test up to the 12th of April of this year in the nine centres. As you may see, there was a great difference between centres ranging from more than 1,000 confirmed cases in Gekedu and Morrovia to just a few confirmed cases in Makburaka and Kisi, the two centres that were open last. In total, MSF admitted 8,520 patients of whom 5,818 were confirmed. So these are the patients including the analysis I'm going to present. So we pull line list data of the nine centres, I mentioned above. We carry out a descriptive analysis for sex, age, time to admission and case fatality. Regarding the factor associated to death, we estimated the incidence rate ratios in univariate and multivariate personal regression and estimate were adjusted for variability among centres. So a low number of children, 5.6% of total cases were reported being admitted to the centres. The majority of cases were in the 15 to 54 years age group. Ebola cases were evenly distributed between male and female. And the age and sex distribution, as you can see, were similar in all the nine centres. So this is why we put them together, not in detail by centre. Regarding the time to admission, 35% reported of coming to the centre within three days since the onset of symptoms. 45% between four and seven days, but almost 20% came at eight days or more after the onset of symptoms. For the variable, we observe some important difference between centres. Some centres had a longer median time to admission because most of their patients stayed some days in a transit or holding centres before being admitted to the EMC. So this slide presents the case fatality by centre. Case fatality range from 38% in both to around 60% if we exclude the key C for which data are still incomplete. The overall case fatality was 52.2%. If we take a reference, Ghekheido's reference where the epidemic started, the death risk ratio was significantly lower in the centres of Dongka, Kailahun and Bo. So we look at whether case fatality changed over time. In this graph, you can see the blue bars which correspond to the monthly totals of admitted confirmed cases and the right line, the overall case fatality which with a decreasing trend, which is clearly visible in the months from August to December, the period during which the majority of cases were admitted. This decreasing trend has been observed in all centre with the exception of Dongka and Bo. So we then look at the changes in case fatality according to age. In this case, we use a lowest estimate which keeps age as a continuous variable. You can see a J-shaped curve with the risk of dying around 70% in young children. The risk then decrease progressively with the lowest value in the age of 15 to 20, 15 to 20 years. And from around 20 years, there is a steady increase of the risk with older age above 60 years, having a risk of 70 to 80%. So this is for the descriptive analysis. So regarding the analysis for the factor associated with this, here in this table is a summary of the risk and incidence ratio for the univariate and multivariate analysis. This table presents an estimate from all MSF Ebola centres. You can see that there is a decrease of the risk over time or with the time, with months passing. That the death incidence rate in young children and in person at older age is twice as high as in person aged 5 to 14 years. That the risk of dying was not associated with the time to admission. And that male gender was at slightly higher risk in dying than female. So we explored the probability of survival according to the cycle threshold at admission. We made a Kaplan-Meier survival curve for each of the seven centres for which this information is currently available. Data are still not available, unfortunately, for the centres of Donka and Kisi. I remind you that we used the cycle thresholds, what we call the CT, as a proxy of the viral load. The lowest, the lower the threshold, the higher the viral load. So we split the variable into three groups, below 18 cycles, 18 to 20 cycles, and more than 22 cycles. The green curve on the graph showed the probability of survival during hospitalization for a person with a CT value above 22 cycles. The pink representing the probability in person between 18 to 22 cycles, and the blue, the probability of surviving in person with less than 18 cycles. And you can see clearly that although some differences are observed between centres, overall patients with a CT value above 22 had the maximum probability of survival, and patients with a CT value below 18 cycles had the minimum probability of survival. So looking at this, so we added the CT in the regression. And as you can see in this table, the death rate in patients with a cycle threshold below 18 at admission was more than seven times higher than in patients with more than 22 cycles, while the other factor did not record major changes. So we acknowledge that this analysis has several limitations, in particular that we have analyzed a limited number of factors. Other factors need to be added in the analysis to have a more complete picture, in particular the clinical symptoms. Obviously, our analysis referred to patients admitted to MSF centres, and results cannot be extrapolated to other centres, or to the significant number of community cases and deaths, or to cases who died on arrival to the centre gate. We analyzed the time into admission, but there is an open question regarding the reliability of this information. I think everyone is aware of that. Another limitation is that the pool data from different centres, each with their own specific geographical context, may hide some ecological factor that we are not able to control. And the support of different laboratories and lab essay make the pooling of the lab result quite challenging. However, despite those limitations, we might conclude that the cycle threshold at admission was clearly the main predictor for deaths, that the risk of dying is also clearly associated with age, with person in younger and older age at higher risk, that the death incidence rate decrease over time independently for at least the factor we were able to adjust for. And there is quite some work still ahead, in particular on the analysis of symptoms at admission and on the treatment received during hospitalization to see whether the workload in the centre had any impact in patient survival, and to keep working improving the information on the viral load and its association with other factors. So I would like to acknowledge all MSF volunteers and local staff who worked and have been working so hard to fight this unprecedented epidemic and to all epidemiologists and medical reference in the field and MSF headquarters that have made this analysis possible. Thank you very much. Thanks Francesco for that great summary. Questions for Francesco? I'm looking further back because I've found, yes, right at the back. I've got a question because I saw that my name is Oscar Serrano from Working with Goal. I saw that the highest mortality was on the under-fives. I am very curious about the mortality among infants less than six months. Well, you can see the trend I think is going higher. There is a clear, this is a regression, so you can see that really there is the younger you are, the higher the risk of dying, and also the older you are, the higher the risk of dying. So I think there is maybe the number quite small in the children, but this is clearly the trend. I don't know if someone did some analysis about the very younger, I don't know. I mean under ones the death rate was very high for the under ones, but I don't know about the under six months, I don't know that particularly. Here at the front with the yellow shirt. Thank you. I'm Hugo Smith from Ames of Holland. Thank you for the presentation. I think this is a very important study because it could be the start of more prognostic studies in depth, but for these studies you need much more information, much more variables. We have any idea how to collect more of this information and not only shouting by the fence? That yesterday we have quite several presentation on that, but I would like to say something that we have other variable in our database that we are not able to pull together because they were collected in an inconsistent way. And I think one of the big work that we really need to do is that to harmonize from the very beginning this data collection. Apart from the tools, we have to agree on what we need to do. But I also think it's a big job now to make the best use out of the data that is available. There will hopefully never be an epidemic like this again and so we will never have a better opportunity to explore these things. But it's a big task to put it even just in MSF data to put that together. Never mind across other organizations, but it's a huge task. Ruby. Hi, Ruby Siddiqui from the Ames Union in MSF UK. Francesca, did you look at treatment? I'm particularly thinking about there's been some criticism around oral rehydration versus IV. This information has been collected for each center and as I say, not in a standardized way, so we cannot put them together at this stage. I hope that with the coming weeks or months, there is some centers that we are able to analyze. But surely this cannot be collected. This information cannot be pulled in the analysis as it is. But yeah, as I say, we hope to do maybe for the next. Okay, thank you very much Francesca.