 Thank you. Good afternoon. Thank you for being here and, of course, in the name of all my colleagues and from the support of our epidemiologists, I have the privilege to present a nosocomial outbreak of winter resistant ESB positive klepsiola pneumonia in a neonatal intensive care unit in Port au Prince in Haiti from July 2014 from September 2015. So, CUEO, it's an emergency center from off-state tricks, complicated objectives based in Port au Prince since 2010. It's an MSF OCA project. It's a container hospital after the earthquake and it's more specialized in maternal complication and had a neonatal intensive care unit, NICU. And from 2015, we admitted more than 20,000, 26,000 neonatal admission. So, why how would break nosocomial outbreak because we identified the germ, it was klepsiola pneumonia and give a sepsis, as you know, is an anterobacteria and gram-negative barcell, extended spectrum bectaractamas and producing strands. Often multi-resistant to antibiotics and quizzing public health problem in special as hospital. So, for your background, you know that a neonatal sepsis, a blood sepsis infection on specific clinical presentation and more or less is a nosocomial outbreak and many studies shows that klepsiola is responsible for more than 20 hour break in NICU. So, in historical for more nosocomial infection, we had the first case on early July, 2014 and you usually increase the neonate and usually high level of mortality from 10% to 26%. Suspected nosocomial infection with highly resistant bacteria. We describe the epidemiologic curve and microbiologic fighting and control over the next 15 months. Case classification was two times, clinical types. From the first of July, clinical types were more about prolonged refill propagation and we had even abdomen distended and, of course, tachypnea. For microbiological classification, we had confirmed ESBL klepsilab pneumonia by blood culture, by other sample like swabs and perinatal nose. We had confirmed other microbes by blood culture, confirmed other microbes by other sample and unconfirmed one with no microbes detect. The epidemiologic curve can show that we have the first cases from beginning of July and, of course, it can show that we had the following peak from 34, 34 cases from beginning of December and, for a lot, we had more cases but what is important on the Q epidemic that our measures make us control the infection and even we had small peaks but not from the beginning one. Our cohort analyses two risk factors. It was small gestational age and birth rate that was very specific from other factors like root of delivery, avga or others. What is important, too, that from blood specimen or swab specimen, all of them shows that imipenem was least resistant from as antibiotic and ampicillin was the last, the more resistant from antibiotic. Clinical management shows that finally we need to adjust our prophylactic and we are three line, the first line ampicentra and the second line amicacin and septacidine and finally, thanks for MSF, we had the third line with neuropenem. These are due by control measures implemented. So first one was coordination about the committee that work with Watson infection control, water and sanitation, lab medical, that was a committee to really control the infection and a lot of efforts was made to for hand hygiene use of alcohol or gel to implement it in the world and most to control clinical management of HIV medication for neonates and of course all of our efforts were made on cleaning and disinfection protocol of material in all the hospitals. Case fatality ratio was that you can see from the slide that normally blue shows you that the number of cases and red show the number of deaths. So as you can see from the three period the case ratio fatality decreased from July to September 2015 and those are made with all that efforts that make from the measures that we put on the hospital. All limitation of course this is a big hospital from most than 6,000 delivery per year and it's a high turnover hospital with high bed occupancy rates, difficulty in implementing and sustaining control measures because we are always winning of high bed occupancy. For surveillance we have send off serial code data from the beginning of July but for now we have a clinical case definition for sepsis not distinct and may have include non-specific babies with other conditions and of course we have limited microbiology confirmation. Discussion and conclusion I would break off highly resistant pathogen in the queue providing complex medical care in a low resource setting like Haiti. Access to high quality laboratory allowed us to detect and diagnose the problem on time and communication was a prime importance in fighting solution and of course multi-sectorial response was required. Antibiotic resistance and nosocomial infection are a current reality that MSF needs to start addressing. Many MSF facilities probably experience similar things and neonates in the queue are particularly vulnerable. Control measures target to reduce morbidity and reduce impact. Our next steps are systematic surveillance of all neonates on antibiotic, antibiotic treatment, appropriate treatment, monitoring of colonizing pathogen, MDR, and Yeltsbill, monitoring of impact of IPC measures what is very important and strengthen IPC across the hospital, head hygiene, Swiss to alcohol just based hygiene. So we need to really take all of patients and staff in queer hospital and of course the MRA in the country. Thank you for your attention.