 Good afternoon, so you're going to have to excuse me on two points, say, that I'm presenting on behalf of Clotilde, who you have in your program, who cannot be here with us today. And secondly, I am not going to wrap my presentation to you. What I am going to do, however, is present MSF eCare, which is an electronic algorithm to improve the management of childhood illnesses in primary healthcare settings of MSF. So what I'd like to first do is just describe the algorithm and then take you through to the feasibility study. So I think this comes, we decided to work in this, coming from a context where we found that in most of the context where we work, there are clinical guidelines available on the field, but often these guidelines do not guide the clinician in a step-by-step process. They're often limited to severe disease, and they are quite ambiguous for the less-trained health workers. So we said, these are the problems we're trying to address with MSF eCare. So we wanted to create a clinical pathway that would be very unambiguous, a step-by-step procedure in order to help the clinician to assess, classify, and treat the patient, children two months to five years with acute illness presenting in the primary healthcare centres. And we wanted to have this based on available evidence and expert design guidelines. So we did a systematic review of the published international guidelines using both MSF material but WHO guidelines and national guidelines to develop the algorithm to include the diseases that we would find in most of our contexts and to have the best practices for the context in where we work. So these, this was then formalised into the clinical algorithm, and this was validated through a peer review process. Our main goal was to improve the quality of care and the rational use of antibiotics through this tool. So I'm just going to take you through very quickly the guideline, the algorithm as it exists. So it's a paper algorithm, this one, and it's based on a syndromic assessment. So it's based on the patient comes in, if the patient is presenting with a cough, it takes you to the cough, it takes you through the assessment of diarrhea, through ear problems and so forth, skin, et cetera. And then at the end it takes you to the fever and if we are in a malaria risk area then the malaria rapid test is proposed. If you don't find a cause of fever then a urinary tract infection is proposed for children under two. If we still don't have an end of the cause of fever, the classification comes to likely viral infection. This reminds the user that if there have been no worrying signs, et cetera, that the child will probably have a self-limiting disease and therefore they can safely say in this case do not need to prescribe antibiotics, however, to inform the mother if the fever persists that she should bring the child back to the health center. So this was the formalization of the algorithm on paper. And we then transformed this algorithm into an Android application. This application needed to be something that would help drive the clinicians through the assessment process. So it needed to be very prescriptive in doing this. Of course we wanted to increase adherence to guidelines that exist and to limit interpretation. It had to be, the application had to move away from the very linear structure of the paper algorithm to non-linear structure because it had to guide the clinician through the consultation process and we had to ensure that the consultation assessment of the child would not be disrupted by the algorithm, by the tool. So we did not want to disrupt the consultation process. The other important point was that the application had to be very flexible. It had to be able to be adapted to the context we're facing, to the locally epidemiological context and to the material available in these contexts on the field. So what we did was we added another menu into the application where the clinician, whether the user could see what material was available to him in their particular field. So in some fields people have oximeters so they could add that in and to the drugs available in their health centers. And this could then be programmed into the application and the application would be modified and adapted to that context. So now I'd like to just explain to you a little bit about the feasibility study that we conducted. Before, obviously, we needed to assess the acceptability and implementation of MSFE care. And we did a limited efficacy testing on antibiotic prescription. So for the outcomes, we wanted to look at acceptability. And with this we want to look at user satisfaction and appropriateness. We want to look at the implementation. So field factors that would either facilitate or hamper the implementation. And the limited efficacy testing, we wanted to look at the antibiotic prescription rate before and after the introduction of MSFE care. So we did a quasi-experimental pre-post study. We did this in the Democratic Republic of Congo in the eastern province in Getty, where we were in three MSF-supported health centers. We enrolled six consulting nurses. And in the Democratic Republic of Congo, nurses are able to prescribe. And the study population was children coming presenting with acute illness two to 59 months. The intervention was a one-day training and then a one-day on-job supervision with the application. And MSFE care was used over a seven-day period in April 2015. So when we looked at the acceptability and implementation, we did the qualitative data through direct observation in in-depth interviews and the quantitative data through the consultation process from MSFE care. I'd like to just specify here that it is not just a decisional support system, the tool, the application, but it also is a data collection tool. So it records what the clinician... So it records as the consultation process goes on what the diagnosis is, what the treatment is, but it also allows the clinician at the end to write their own diagnosis and treatment with what they've come up with. For the limited efficacy testing, we did a comparison pre-test, so direct observation of the consultation process, and then we did the post-test on the antibiotic prescription rates, looking at the data coming from the application. All participants enrolled in the study was through informed oral consent, and there was ethics oversight by the MSFE medical director and according to the MSFE ethical framework for innovation. So if we come to the results, the acceptability. So users through the in-depth interviews reported that initially, even though the application was hard to use in the beginning, after a few days of use, it became very easy for them to use. It allowed them to be a lot more... They reported to be a lot more systematic in their consultation, and it allowed them to access recommendations that were otherwise hard for them to find, and that it helped them to improve in the antibiotic prescription. They did say, however, that it increased the consultation time, and when we asked why, they said because they had to do a more thorough clinical examination of the child. So in terms of acceptability, we found through that they reported that it covered the majority of the clinical situations that they encountered, and that it proposed a treatment that was adapted to their context. From data coming out of the application, out of the tool, we saw that in 95% of the symptoms reported were addressed by MSFE care. Symptoms that were not covered by the algorithm were eyes, mouth, and genital areas. In the diagnosis and treatment proposed, over 90% of the consultations were addressed, with patients presenting with at least one symptom, and it was shown in terms of the adherence that the nurses in 85% of the consultations followed the recommendation proposed by the application, whether to prescribe or not an antibiotic. Looking at the efficacy testing, and it was quite limited, but in the pre-test observation of the prescriptions, we saw that in 46% of the patients' antibiotics were prescribed, and then when MSFE care was introduced, 25% of the patients were prescribed antibiotics, so we can infer that there was a more or less 50% reduction in the prescription of antibiotics after the introduction of MSFE care. So we can say that feedback was that it was well accepted, that clinicians, users found this to be a good job aid. It improved the thoroughness of the clinical assessment and helped to reduce antibiotic prescription. It was appropriate to the field realities we were facing as it covered 90% of the clinical situations, and the nurses adhered to 85% of the recommendations. It decreased the antibiotic prescription, and it was technically feasible in Getty. We didn't encounter major technical problems, but what we did encounter was problems in data transmission, and following this, we realized that we would have to develop a peer-to-peer offline data transmission because we were in Getty particularly, and in other contexts of MSF where we work, we're often in areas where there is no connectivity. So if supervisors are going, we needed to develop this offline data transfer so that supervisors could go, collect the data, and then take it back to areas of connectivity to the Amplis doctor's server. The future perspectives are that this will now be further piloted in Central African Republic in Bebarati. We would like to conduct a longer feasibility study for over a period of a year, and we would like to adapt the application for more remote settings. We also feel that this application can help in further research areas, such as helping in etiological surveys, and be a platform for new diagnostic tools such as beer markers and respiratory rate sensors. So I'd just like to thank the people involved, first for the development of the paper algorithm, the clinical algorithm for the development of the software, and for all those that helped put the feasibility study in place, and thank you for your attention. Thank you.