 So, before I start with my irritating role of academic saying that we need more evidence, I mean as Director of the Center of Global NCDs I also want to take the opportunity to welcome everyone and I think, I mean the spirit of this meeting is, I mean there are a lot of relationship between the different agencies, we want to strengthen those relationships but also foster new ones and find solutions but also find which are the gaps in terms of research. We have this double agenda and part of the complexity and the tension from researchers and agencies working but I think we need this space for discussion and find solutions and questions we need to work on. So what I'm going to be presenting also is for setting the scene of some of the programs that are going to be presenting later in this session and it's about the system and the overview that we are conducting with people from the school and UNH and people like Martin Marquis who is chairing this session and for that I just want to start covering briefly what is integration and the rationale for integration of HIV and NCDs and present some preliminary results. So you might be familiar with the building blocks from WHO, from service delivery, health workforce, information, medical products, financing, leadership and governance and integration could be many things and it's not my objective to go into that but it could be basically across a related condition so this is programmed, it could be within a building block, it could be across one or more building blocks, it could be between public health programs like promotion, health promotion and management or it could be across activities in health systems and other sectors like transport or I mean things if we think about obesity definitely we need to think about other sectors. We use this definition and you can, as I said, you can have a whole symposium on integration probably but we use this definition for our system overview. So managerial operational changes to health system to bring together inputs, delivery management and organization of particular service function as a means of improving access, acceptability, quality, coverage and cost effectiveness and we particularly focus on service delivery for diseases that are usually delivered separately but often affect the same type of end users. So this is the cartoon that Philippa was not allowed to show and it's from someone from MSF Switzerland and I think you can all read it and so you know better than I do, I mean the need of integration and I think that's common sense but it's really much better explained by this cartoon. So we know about the double burden of disease and we talk about this in particular in Sub-Saharan Africa, the biological interaction between HIV and NCDs, I will not get into that but in particular there is some evidence about increased risk on cardiovascular disease and also cancer and HIV. We know that there's strong vertical HIV programs and weak fragments for NCDs and I agree that I mean our entry point is NCD but maybe covering some of the things that Taman was mentioning we should be thinking about health in general but NCDs cover many of diseases and the commonalities and potentials to leverage from HIV systems and this shows years lives with disability, this is data from the Institute of Health Metrics University of Washington and you can see this is from Sub-Saharan Africa so you can see already in terms of years lives with disability, I mean the huge burden but again going back to Dave's comment, I mean most of this data probably is made up to be here. We don't have much data and I don't know, Dave can again do a whole symposium about the whole modeling that we have on GVD but there is some evidence that it is a problem and this is data from the same source from the HME, there is a development assistant for health so these are millions of dollars and in blue you can see HIV, this is Sub-Saharan Africa again so most of the development money is going for HIV and maternal health and malaria and NCD is quite flat. So it makes sense, I mean this is a rational for integration but also I think we need to be careful, this is a nice example from people from the school that Catholic Church led on exploring the integration between HIV and reproductive health in Switzerland and she looked at four types of models from less integrated to more integrated, actually what they found is that they didn't find an impact on health outcomes and actually there was a decrease on some of the outcome for HIV so there was less use of I think recommendation for use of condoms and what they found actually that they were not so integrated and they show with some simple method the importance of evaluating integration that it not always work. So with that in mind we said okay we are going to do a systematic view of integration and our objective was to describe different models of integration and to identify barriers and facilitators at the level of the health system that can influence the success of integration. So we look at HIV and any NCD chronic condition, we include quantitative or mixed method studies that they were either descriptive or evaluative but we didn't include editorials or reviews just should be primary research. We found more studies than in the previous one that make us to do more work and that's why it's taking us longer, 155 studies. And this slide shows on the prevalence of HIV and where the studies came from so you can see most of them came from North America, third from Africa followed by Europe and then Asia and only 1% in Latin American, Australia. These were the main conditions that were reported in the papers. Mental health, substance use disorders, cervix cancer, cardiovascular disease and others. I'm going to just concentrate the portfolio of systematic reviews that we are producing in the different topics, one that is led by Dina Balavanoba, that is the overall review and then by topic. The one I'm involved is more cardiovascular disease and that's the one I'm presenting here. As I said the preliminary result, there were 17 studies in 14 reports but only six were full papers, what made everything much more difficult to analyze. The most common integration was HIV, diabetes and hypertension. Most of the studies were descriptive. They were just talking about the program and what they've done in a descriptive way. The reporting, as I said, because we only found six full papers was incomplete, most of the follow up was short and all they measured were more process outcomes or proxy outcomes, blood pressure for example. These were, we tried to summarize the different models and the level of the patient entry point and this reflects that. Community and mobile and these shows from less intense to more intense in terms of treatment. These are mainly screening and these are treatment and referral. Community and mobile, HIV clinic, primary healthcare and secondary healthcare. You can see the distribution of these studies. Unlike the overall review, most of the studies for cardiovascular disease were in Sub-Saharan Africa. Most of them were add-on to existing HIV services, how they were designed. As I mentioned before, they were descriptive, they were not evaluative. They use proxy and process outcomes and the reporting was really poor in terms of, I mean, it was very difficult even to assess what was the risk of bias. On the right, we want in this paper to have at least some conclusions and lessons from these papers and these are things that came across those papers. They, I mean, the obvious benefits of integration and some suggestion that is feasible, acceptable and with potential impact. To be honest, I mean, the strength of the evidence is not much but, of course, and there is a huge potential of reporting bias. I mean, of course, they are doing these programs and they just report the things that work. And these were some of the challenges, the importance of training and protocols, the competing demands for primary care providers and, I mean, the importance of clear linkage. But, of course, these still, I mean, if you think about HIV and cardiovascular disease, leaves other current conditions outside the integration. So I'll start with the research recommendations. I will finish with some practical ones. So what are the research, I think, I mean, finding the right question for integration is key. And I think some of them are what are the different models? I mean, what is their impact? What are the components that work? And importantly, those that do not work. In which context, I mean, how the context affect this? And how can promising integration strategies be adapted and scaled up? And this is below the definition of what is implementation research. And I think that's the kind of research that we need for integration. So understand not only what is and isn't working, but how and why integration is going right or wrong and testing approaches to improve it. And we need to do high quality research in this context. And I think we're going to touch on this later this afternoon. This is not about, okay, we need to do, I mean, I've done randomized control trials, but I'm not saying we need to do randomized control trials about everything. We need to do the best design that we can do in that setting and minimize bias as much as we can, as they did in the paper that I presented from Catherine Church. We need to use mixed methods. I mean, the qualitative research is important beyond the quantitative to understand some of the findings. We need to do, I mean, large studies. And again, this comes to some of the data, some of the discussion we have before about the need of data, process evaluation, and clearly better reporting. Otherwise, it's better difficult to get lessons and scale them up. But this is something that we came in the main paper that, as I mentioned, Dina is leading. And it has some kind of cross-cutting themes that we think could be useful when you think about integration. The first one is about effective leadership and management. And this is so, there should be an important decision, political decision about integration at the political level, but also at the management level, in the hospital, in the district. Then the structures and resources supporting of integration. This is like the hardware of integration. And I think this is key, I mean, location. I mean, co-location might be important, but also, for example, for mental health, there should be a location that is appropriate to deal with these kind of patients where you can deal separately with patients that are in a nice, an appropriate environment. Train health workers with appropriate incentives and clear roles. The important issue is communication, formal and informal communications and collaboration. And this, again, is finding most in the qualitative results of the papers. We found there's a lot about the importance of communication. And part of the findings is the lack of communication. And then patient-centeredness. I mean, this is the most obvious one. And although I think for anyone, if anyone has been a patient and understand the importance of the, we are not just HIV patients or hypertensive patients, but I think we work in that way. I mean, as researchers, as NGOs, and I think we need to keep that in mind all the time, it's easy to say, but it's not so difficult, not so easy to implement. So as I mentioned, this is the work of a group of people.