 Good morning everyone and thanks for joining us today in this session. I'm going to be talking about health systems strengthening using surveillance in Sierra Leone. So our story takes place in Tonkolili district, population of around 430,000 people, area around 7,000 km2, it's divided in 11 shipdoms, and it has around 1,000 villages, most of them in a rural and hard to reach area. MSF started maternal and child health problem here in January 2016 in the locations in the map. Why are we there? Ebola affected badly this district. We had the Ebola management center which closed in May 2015. In October 2015 we started supporting surveillance activities and after we had closed the EMC we had the resurgence in January 2016. The Ebola outbreak and this last flare up revealed some important gaps in the surveillance system. Before I explain to you what we actually did in surveillance in Tonkolili, I will briefly explain what are the main traditional approaches of MSF when working in surveillance. First we have the reporting of health events from our health facilities. Then we have the active surveillance outside of the MSF area, that's probably what you were talking about. We call other facilities and try to get their data. We don't normally work with those facilities and then we have the household level surveillance done by community health workers. So now talking about Tonkolili surveillance. In Sierra Leone, the Integrated Disease Surveillance and Response System, IDSR was adopted by MOH and WHO after the Ebola outbreak. This is a weekly paper based system reporting of 27 notifiable diseases and deaths. To the right we can see in this flowchart how the communication works, the information flows. We have the health facilities, 103 peripheral health units, PHUs and three hospitals, which is a lot. And then these health facilities send the reports to the district surveillance office and the district surveillance officers send the reports to a national level. The arrows go both ways there because the district surveillance officers can also call the health facilities that have failed to report on time. As we said before, we identify some gaps in this system, such as the lack of routine surveillance analysis and reporting. We had a large map remote area under surveillance with the challenges in terms of communication and access. Performing indicators such as timeliness and completeness were very poor. And then the fact that the system was Ebola focused, which made sense at the time but now it's a bit of a challenge because some of the diseases don't receive enough attention. For all these reasons, we decided to strengthen the public health surveillance in this large network of health facilities to improve the detection and response to epidemics and deaths beyond the MSF area of intervention. What did we do to help this surveillance system? We took a multifaceted approach to strengthening triage and case definition, quality of reporting and analysis of the reported data. How did we do this? We deployed an MSF epidemiologist, an expert, and we hired and trained an MSF National Staff Epidemiology Assistant. We decided to work closely and in continuous collaboration with Ministry of Health and Sanitation and WHO and we organized the IDSR trainings and supported the PHU supervision visits. In these visits, we would use a WHO developed checklist based in a mobile data collection system. And furthermore, we developed an open source dashboard software tool to identify poor performing PHUs in hospitals in terms of completeness and data quality and to help us detect outbreaks and unusual trends. Apart from helping them in reporting analysis, etc., we decided to also strengthen the response and activity planning capacity. We did an exhaustive mapping with field validation from MOH. We joined case investigation of epidemic prone disease, supportive vaccination campaigns, helped with reviewing malaria and maternal deaths data, and helped the district surveillance officers when preparing the weekly surveillance meetings and the epipooletines. This is an example of the dashboard. You can see the last version outside, and Idris can't give far details, but you can see how easy it is to visualize disease trends and the geographic distribution of the cases. This is the map that we did. First, we mapped the health structure in Red Crosses. Then we mapped the villages, the gray dots, and then we matched those villages to their corresponding health structure to define the catchment area of each PHU. So, what did we achieve? What were our results? We managed to organize four IDSR trainings to support more than 130 PHU visits. In this picture, we can see an example of one of the trainings. In this picture, the epidemiology assistant is checking the registration consultants' books. In these PHU visits, we would also use the dashboard to feedback to the PHUs about their performance in terms of data quality, etc. One of the most important achievements of our intervention was the improvement in the reporting completeness in these figures. We can see how the complainants went from around 50% of health facilities reporting in Epic Week 6, 2016 to 100 of them reporting in Epic Week 2017. In this slide, we can see how we managed to reduce the misclassification of the area severe dehydration and severe pneumonia after the efforts and continuous training on the use of the correct case definition. Before this, most of the PHUs were reporting all the areas as the area severe dehydration and all respiratory presentations as severe pneumonia. We managed to support 63 case investigations, most of them missiles, and that is because at the beginning of 2016, Sierra Leone had an outbreak of missiles, and we supported the subsequent vaccination campaign for missiles and all the vaccinations for polio. In this picture, we can see the district surveillance officer in a case investigation processing sample. We managed to identify priority areas for a community-based malaria program using the dashboard to analyze the malaria burden and the geo-distribution of cases. We helped the MOH when reviewing and analyzing the maternal death data, and by doing this, we reinforced maternal death surveillance, improved documentation of deaths, and identified areas that needed attention such as the timeliness of maternal referrals. We faced some important challenges. For example, the mapping data was of poor quality and mainly related to Ebola. The lack of local resources, MOH didn't really have the capacity to allocate any extra resources, human or logistical, to the district surveillance office. Some communication issues at the beginning with WHO and MOH, but now it's working very smoothly. The fact that MOH is not using the dashboard independently yet despite our efforts to train them and they still need us around to help them, and the fact that there is a new national data health system in place now, which is very good, but some of the features are not working yet such as mapping, so they still need our dashboard to do that and our support, so this is a bit of duplication of systems. In conclusion, we realized that surveillance is a powerful public health tool to strengthen clinical skills. For example, by reinforcing case definition, we reinforced triage and correct diagnosis. The national AP assistance was key to build relationships, acceptance and to ensure the continuity even when the expat is not there. The dashboard supported data quality training and more importantly, the real-time analysis, alerts, outbreak detection and intervention planning. The engagement with MOH and WHO was crucial, working closely and in continuous collaboration with them was fundamental for this approach. And finally, and more importantly, to our knowledge, this is a novel approach for MSF with limited human and technological resources. We managed to support a large network of rural health facilities in their ability to improve their ability to report and respond to epidemics outside of the MSF area of intervention. And finally, I would like to acknowledge the rest of colleagues who participated in this project and especially the PHE staff and the people of Tancolili. Thank you. Do we have any questions? We have one at the back there. So the community part of the surveillance network, could you explain that a bit more, how that works? Because we know that often health facilities have access problems, so any surveillance system only based on health facilities might be missing out or detecting late any outbreaks. Could you explain a bit more about the community part? Yeah. In fact, one of the items that were checked during those visits, that checklist developed by WHO that we were using to supervise the PHE, includes the routine communication with community health workers from all the small villages that health facility is covering. So when we were going to these facilities, we were also checking that they were actually in communication with them and they were also performing their outreach activities and their active case search in communities. So yeah, at times it was a challenge and they were not always able to perform all the activities, but that was one of the points that we were trying to improve. One question. Hi, Claire from MSF. Just given the presentation before you, can you see a place for using SMS in this kind of situation as well? Because having worked in Sierra Leone, I know that one of the issues is the access and getting data. Yeah, yeah, definitely. And actually we might talk to Zia after this presentation. It's a very interesting system. In fact, as part of the new system that the MOH is putting in place, there is a plan to use some type of mobile data collection at the PHE level, but I think this is a simpler way rather than having a tablet in each health center, having an SMS might be a much better idea, keeping the context.