 Thank you. Can everybody hear me okay? Yes. So I'm going to present a little bit of the work that we have been doing myself and my colleagues in Mali. So focusing primarily on the project in Kutiala in southern Mali, population of 629,000. So since 2009 MSFOCP has been working with the Malian Ministry of Health in the form of primarily a referral hospital and in five community centers. The project that we focus on primarily is pediatric care. Kids under the age of five. The major pathologies that we see in this population is malaria, acute malnutrition, diarrhea, and respiratory tract infections. The objective overall of this intervention is to try to reduce pediatric mortality. So a little bit about the inpatient setting. There's several wards where the pediatric patients are seen. There's an emergency room. There's an intensive care. There's a few inpatient medical wards and there's an intensive therapeutic feeding center for malmarished children. There's a neonatal ward and a burns ward or burns beds. Normally the inpatient pediatric unit has between 150 and 300 beds. However, this balloons explodes in malaria peak season between June and October to accommodate over a thousand admissions a month. Since 2014 MSF has had a microbiology lab on site in Kutiala. So OCP built this lab in 2014. The objective was to diagnose causative pathogens responsible for bacteremia, urinary tract infections, and meningitis. So the aim of this review was really to look at 12 months worth of data and to identify the main causes of bacteremia in children aged under five years in Kutiala. To identify causative bacterial pathogen and sensitivity. To try to address the question of whether they need to be adaptation of antimicrobial therapy, particularly empirically in the setting of sepsis. The background for this is documented increasing rates of antimicrobial resistance globally. To really try to address the question of whether our first line antibiotics for sepsis are still appropriate. So our methods we retrospectively analyzed routinely collected microbiological data from January to December of 2016 in children who were admitted with clinical suspicion of sepsis. These are kids that had a blood culture at admission and also those that had no improvement after 48 hours of antibiotics. All the data was collected in the Hoonet microbiological database and for those that are not familiar with Hoonet it's a free software developed by WHO which is basically like a microbiological bacteriological repository. It's a database and we use it in all projects where we have microbiology. We followed UCAST recommendations for antimicrobial susceptibility testing so these are the European recommendations for determining whether bacteria or sensitive or resistant. And for analysis of the anti-biogram which is what I'm about to present we we assessed it extracting information from Hoonet and followed the CLSI guidelines which are the American guidelines for how to develop an anti-biogram which basically in short recommend that for each patient you take one isolate per species per time frame that you are reviewing. And this analysis met the criteria of the MSF ethics review board for exemption from ethics review. Just a quick note to say that the idea of building an anti-biogram is to do this with some regularity to review ideally annually what your bacterial pathogens are and whether or not they are sensitive or resistant to your first and second line antibiotics. So in 2016 11,304 patients were admitted. We had 3,320 blood cultures performed. 14% were positives that's 479 were positive 10% of those were contaminated. In terms of guilty parties for taking blood cultures the majority came from intensive care, not surprisingly, followed by the pediatric inpatient ward, the nutrition ward, other wards and then the neonatal ward. So what we found was by far the most common pathogen was salmonella. This shouldn't be surprising. Followed by E. coli, staph warrius, entrochoccus ficalis, streptococcus pneumoniae and klepneumoniae. So this is similar and different to previous assessments of the project. What stands out in particular is entrochoccus ficalis and klepneumoniae. The other four pathogens are fairly standard fare for this age group. Among entrobacteraceae, so gram negatives largely, we found that 13 were extended spectrum beta-lactamase producing, meaning that in these patients your treatment of choice is a carbapenem for bloodstream infections. Of these 39 isolates that were ESBL producing, 36% were klepneumoniae, 33% were E. coli. So in a little bit more detail, I apologize the title is a little bit misleading. So what we're looking at actually is salmonella tested for various antibiotics that it should be sensitive to and then percentage of sensitivity. So what we see I suppose not so surprisingly is quite broad spectrum resistance to or quite widespread resistance to ampicillin. I don't know if I can point, oh yes I can. But reassuringly, almost 100% sensitivity to seftriaxone which is our first line treatment for sepsis and 100% sensitivity to ciprofloxacin, which is great. So oral agent for salmonella. More concerningly, in klebsiella and E. coli, we find higher rates of resistance. So this is not projected particularly well, I'm not sure why. But, so I don't have a, okay. So this should be, so this is sephoxetone, which is comparable to sephazolin, which is a class one cephalosporin. This is, I suppose, the money shot here, which is concerning its seftriaxone. And what we see is that only 19% of klebsiellas are sensitive to seftriaxone. 58% of E. coli's are sensitive to seftriaxone. This should be ciprofloxacin. 33% of klebsiellas are sensitive to ciprofloxacin. 50% of E. coli's, likewise. And also concerningly, 48%, only 48% of klebsiellas are sensitive to gentamycin and 63% of E. coli's. Reassuringly, 100% sensitivity to amicacin. This is the other thing that is a little bit peculiar is entrococcus ficalis. Not only did we find it in high rates, but we found it having surprisingly high levels of resistance to ampicillin, which is not really normal for entrococcus ficalis. Reassuringly, for staph aureus, we found no methicillin resistance, so no MRSA. So we have some limitations from this dataset. One is that it's routinely collected data. Blood cultures, in theory, should have been performed in all kids that were unwell, but we know that this doesn't always occur in a systematic way. And certainly there was inconsistent collection of clinical data and correlation to microbiology data. What is particularly concerning and that we hope to address further in ongoing analysis is our limited ability to differentiate nosocomial versus community acquired infections. Once again, we had inconsistent baseline blood cultures and really limited documentation about patient location. So particularly for the klebsiella pneumonia, these appear to be temporally linked. So we think that this was a klebsiella pneumonia outbreak, a nosocomial outbreak. In order to really comment on anti-biograms, they recommend, CLSI recommends that you have a minimum of 30 strains for each species. So the conclusions are what do we really do about this with this information? So for the most frequent pathogens that we see, the first line antibiotic therapy is appropriate, that applies certainly to salmonella and to staph aureus. For klebsiella and E. coli, I think for bloodstream infections, we really need to be considering in this project switching empirically to an active agent like a carbapenem or adding a second agent to improve sensitivity while awaiting formal sensitivity testing. I think there is a real need to better differentiate nosocomial versus community acquired infections, including rates of resistance. And in this analysis, this applies particularly to klebsiella, also E. coli, but also the enterococcus vicaris. And to better correlate clinical outcomes with resistance. That's it for me. Thank you. Is there any brief questions for clarifications before we go on? Yes, sir. The microphone's coming. Thank you for an interesting presentation. My name is Andreas Flores, Chief Surgeon ICRC Geneva. You speak slowly, please. Sorry. Andreas Flores, Chief Surgeon ICRC Geneva. Very interesting presentation. My question is simply this lab, is it set up as part of routine care or is it only for research purposes? It's for routine care. So we use it for any child that is admitted to the facility or any child that is referred from the five referral centers that needs to be admitted and is unwell. Routine care. Thank you very much. Lady in red at the back. Thank you. Thank you. My name is Marta. Can you put it pretty close to your mouth, please? I'm very, very close. No, it's working. All right. My name is Marta. In fact, this is Doctor. My question was regarding how are we doing the identification of the organisms and what methodology are we using for it? Right. So normally I would, at this moment, defer to my microbiology colleague who set up the lab. So we have a recently, since I think August of last year, have a backed alert system in terms of identification of resistance or of the organisms themselves. So the lab staff are trained. So they undergo, I mean, do you mean like gram staining and perhaps we'll have the full questions afterwards, is this? Yeah, that microphone is not working very well. So after they'll change microphones. Any other quick brief questions, clarification, madam there? This is Cecilia MSF. Thank you, Rupa. Excellent presentation. My question is about, apart from the lab, which is one of the big issues that we have difficulties to set up on the field, is did you manage to have more data on other main killers like malaria? I mean, because the problem is that many of these kids, I might say that 90% of them are getting 20 drugs. I'm exaggerating because I'm South America, but they might have anti-malarial, anti-biotic, anti-TV drugs, just in case ARB. So do you have some data on how many of these kids were real diagnosed with malaria and then really bacterial diagnosis? Really good question. So I, there is a study ongoing looking at malaria infected versus not kids that present with sepsis. I don't have that information for this presentation because it's not captured in HUNET. So HUNET is purely bacterial. Last question. Yeah, I'm Emily from MSF. Were you able to do a subgroup analysis between malnourished and non-malnourished kids? We have not done that yet. So the problem is again, linking clinical to microbiological. I mean, the only information that we have is location of the patient. So we can do it for those kids that were admitted to the cranny, to the feeding centre, but we have not yet done it.