 I'm doing this presentation on behalf of Coral Clinical and Operational Research Alliance that is a research platform gathering people from the ALIMA NGO and scientists from INSERM, the French National Institute for Health and Medical Research in Bordeaux and Abidjan. And we developed together innovative research studies aimed at improving maternal and child health in humanitarian settings. And I'm presenting today the results of the ongoing one-thousand-days cohort studies in Myria, Niger. And it's going to be on maternal factors associated with maternal, with perinatal mortality. Perinatal mortality is defined as late fetal mortality plus early infant mortality occurring during the first week of life. In low resource settings with high prevalence of home birth, it is often difficult to differentiate late fetal and early neonatal deaths. So we believe it's a useful indicator to assess the quality of obstetric and early neonatal care. Perinatal mortality in Niger is among the highest in the world. Data is scarce, especially in rural areas. We found a recent publication showing a 10% rate of stillbirth in Zander, Niger, so in the same region where we implemented our cohort. In this context, prevalence and associated factors of perinatal mortality are important to assess. And the idea is to be able to identify maternal risk factors that may help with prioritization in birth planning, meaning deciding whether the mother should delivery at a health center or at a hospital for the most complicated cases. So here on the map, you can see where we worked. So it's in the Myria district, Zander region. The 1,000 days health services were implemented in three of the 17 health zones of this district. And services were provided in 16 health structures covering a population of 125,000 and with 6,700 births expected every year in Ziharia. So the objectives of the talk today were first to evaluate perinatal mortality and associated maternal risk factors in rural Niger. And second, to assess the proportion of pregnant women presenting with dangerous signs identified during antenatal care in relation with both perinatal mortality and place of delivery. So here on the slide, you can see the full package provided within the 1,000 days health care services. And it's being, it's provided to both pregnant and lactating women and then to newborns until they reach the age of two years. So it includes, for instance, immunization, prevention and treatment of malaria. And it's important to emphasize that all mothers were receiving nutritional support during their pregnancy. And same for the kids who received a deliberation of Nutributa from six to 24 months of age. So we are in a context, as you know, where most pregnancies occur at home. So one of the ideas of this package was to convince women to deliver the baby more often within the health structure. So you can see on the graphs that coverage survey were conducted prior to the implementation of the 1,000 days package of services. And 25, only 25 percent of women delivered within the health structure. And the same type of survey was conducted one year after the implementation of the 1,000 days package services. And as you can see, in the non 1,000 days zone, the proportion of women delivering in health structure did not change. It was still about 25 percent. But this proportion increased in the 1,000 days catchment area to about 40 percent. So is it good enough? And is it a good indicator to look at? So I'm going to come back to this later. So concerning the presentation today, the cohort was implemented in five health facilities in rural Niger. And at these facilities, we included all pregnant women presenting for first ENC visit between April 2015 and June 2016. And then mother infant pairs were followed until 24 months postpartum. Concerning the statistical analysis, we run a multivariate generalized linear mixed model to examine factors associated with perinatal mortality, including study site as a random effect. And then we created a complication score based on six danger side identified during any of the ENC visits. So it includes women of younger or older age, first pregnancy, multiple gestation, severe anemia, positive malaria testing, or preeclampsia defined as high blood pressure with proteinuria. And we assessed this complication score in relation to perinatal mortality and place of delivery. So you can see on the flow diagrams that we included 1,745 women in the study. Of these women, we had 50 unknown pregnancy outcomes, 12 abortions, and four maternal deaths occurring during pregnancy. So ultimately 1,679 women gave birth. And of these, we observed 68 steel births, 4%, and 16 neonatal deaths, 1%, resulting in a total of perinatal deaths of 5%, which is elevated. Concerning baseline maternal characteristics, they were aged in median 25 years. The median parity was high, three ranging from 0 to 14. One-third of women had nutrition deficiency with the MUAC below 230 millimeters. And concerning the number of ENC visits, you can see that 15% of women only did one. About half of women did between two and three ENC visits, and 37% of women did four ENC visits. It's important to note that gestational age at first antenatal visit could not be reliably estimated due in particular to late presentation. So on this slide, you show the final model with multivariable analysis on factors associated with perinatal mortality. Nothing surprising here, but I think it's always important to emphasize. So you can see that women having done less than four ENC visits, among women having done less than four ENC visits, the perinatal mortality was 1.7 higher. It was two times higher for women with a primary parity. Women who had malaria-positive Rdt during pregnancy, the mortality was 2.5 higher. And it was three times higher among women diagnosed with severe anemia during pregnancy. So then we looked at dendrocynes that I defined earlier, and you can see that as many as 30% of women experienced at least one dendrocyne, 20% at one dendrocyne, 10% at two dendrocynes, and only 1% at three dendrocynes, and no women experienced more than three dendrocynes during pregnancy. So then we looked at dendrocynes in relation with perinatal mortality, and you can see that among women with no dendrocynes, this mortality was 3.5%, while it was 8.5% among women with more than one dendrocyne. And interestingly, the perinatal mortality was not higher when women experienced more than, when they experienced two dendrocynes. We had not enough statistical power to conclude on whether the mortality was higher with three dendrocynes. And finally, we looked at dendrocynes in relation with place of delivery, and it was disappointing for us because, as you can see, among women with no dendrocyne, 37% delivered in a health facility, and exactly the same proportion of women experiencing at least one dendrocyne delivered in a health facility. So as a conclusion, perinatal mortality is a useful comprehensive indicator in environments where a large proportion of birth occur outside of health structures, and we believe it should be more widely reported. So here we had a prospective cohort with minimal dropouts, and it shows high perinatal mortality rate, 50 per 1,000 births. We have to be cautious when we try to benchmark this figure with others, and because it really depends on the context and on the methodology. Perinatal mortality was associated with maternal risk factors that are easy to identify during ANC. So we observed that a higher proportion of women were delivering in health structures with 1,000 day services, but unfortunately pregnant women with risk factors were not more representative. So health workers should now make a better use of data collected during ANC so that the most complicated cases can deliver in health facilities. Before I close, I would like to acknowledge my co-authors listed on the slide, as well as the Coral Board of Directors who decides on the strategy and on our research strategy. Of course, all the women and children who participated in the study and the study funders. Thank you for your attention.