Infant Mortality in Rural and Post-Con ict Areas in South Kivu, Eastern DR Congo: A Cross-Sectional Study

Gaylord Ngaboyeka (  amaningaboyekagaylord@gmail.com ) Ecole Régionale de Santé Publique (ERSP) , Faculté de Médecine, Université Catholique de Bukavu, Bukavu, Democratic Republic of Congo Espoir Malembaka Ecole Régionale de Santé Publique (ERSP) , Faculté de Médecine, Université Catholique de Bukavu, Bukavu, Democratic Republic of Congo Paci que Lyabayungu Ecole Régionale de Santé Publique (ERSP) , Faculté de Médecine, Université Catholique de Bukavu, Bukavu, Democratic Republic of Congo Samuel Lwamushi Ecole Régionale de Santé Publique (ERSP) , Faculté de Médecine, Université Catholique de Bukavu, Bukavu, Democratic Republic of Congo Aimé Cikomola Renforcement Institutionnel Pour des Politiques de Santé Basées sur l’Évidence, Democratic Republic of Congo, Lubumbashi, Democratic Republic of Congo Pierrot Mulumeoderhwa Ecole Régionale de Santé Publique (ERSP) , Faculté de Médecine, Université Catholique de Bukavu, Bukavu, Democratic Republic of Congo Albert Ntambwe Ecole de Sante ́ Publique, Université de Lubumbashi, Lubumbashi, Democratic Republic of Congo Ghislain Balaluka Ecole Régionale de Santé Publique (ERSP) , Faculté de Médecine, Université Catholique de Bukavu, Bukavu, Democratic Republic of Congo


Background
In the Democratic Republic of Congo (DRC), infant mortality remains one of the highest in the world, despite a remarkable drop of 60.8% over the last two decades, from 148 per thousand in 1995 to 58 per thousand in 2014 [1,2]. According to the UNICEF, several factors contributed to the decline in infant mortality rates in the DRC. These include immunization, integrated management of childhood illnesses, breastfeeding, use of long-acting insecticide-treated bed nets (LLINs) and supplementation. in vitamin A [3].
According to data from the second demographic and health survey (DHS) carried out in the DRC, there are still large disparities in infant mortality between provinces and between urban and rural areas within the same province. The provinces of South Kivu and Bas-Congo are the most affected (with 92 and 81 deaths per thousand live births respectively) unlike those of Kinshasa and North Kivu (with 50 and 41 deaths per thousand live births [1].
Since 1994, South Kivu has been the theater of a situation of war and armed con icts with dramatic consequences for the health of mothers and children. Studies had shown that the repeated armed con icts in eastern DR Congo had a signi cant impact on overall mortality [4,5]. However, a recent study on DHS data from 2013-2014 suggests that infant mortality is not speci cally higher in eastern R. D. Congo where armed con icts persist even during this post-war period [6,5].
Despite the availability of provincial estimates of the level of infant mortality in the DRC, there is scarcity of local data considering the geographic and socio-political speci cities of each health zone. Such data is important to enable political and health actors to develop health programs and policies adapted to the geographic, socio-political, historical and infrastructural speci cities of each health zone.
Recognizing that infant mortality is one of the reliable indicators of the state of health of a population [7], this study set out to measure this mortality in two rural and post-rural health zones (ZS). crisis in South Kivu, including the areas of health Miti-Murhesa and of Walungu, two media with a history of e quite different war.

Study framework
The ZS of Miti-Murhesa and Walungu are located respectively 33 and 46.9 km from the city of Bukavu (capital of the province of South Kivu), with an estimated population of 246,943 and 253,871 inhabitants respectively, according to the data. of 2016 from the Provincial Health Division (DPS). These two health zones are accessible by road. Unlike the Miti-Murhesa area which has been relatively spared from persistent con icts in eastern DR Congo, the Walungu health zone has been more severely affected by chronic armed con icts and persistent armed groups [8].
In Walungu, 17 out of 23 health centers (74%), three hospitals and a general referral hospital had been selected and in Miti-Murhesa, two health centers (out of a total of 16), 4 hospitals and a general referral hospital was included in the study. These health structures were selected in a reasoned way because of their geographical accessibility and also because of logistical di culties.

Type and duration of study
We conducted a cross-sectional study from July 2016 to September 2017.

Study population and data collection
Our study population consisted of deliveries from the two aforementioned areas during the study period in which a structured questionnaire evaluating the survival of the child from the pregnancy preceding the one leading to the current delivery. We systematically included in this study all deliveries who gave birth in one of the structures chosen during the study period and who have lived there for at least 6 months in one of the HZs of our study. Excluded were deliveries who did not consent to the study and those for whom the postpartum state of health did not facilitate the interview.
A structured questionnaire was administered to the deliveries during the period from childbirth to one week of life by data collectors who consisted of 6 people including 3 doctors and 3 nurses. The collectors were trained on the questionnaire and the latter was tested in a center which is not part of the centers concerned by the present study.

Study variables
The main dependent variable was infant mortality. The latter was de ned as the number of child deaths occurring during the 1 st year of life reported in 1000 live births [9]. It was measured as a dichotomous variable. The explanatory variables considered in the analyzes included the socio-demographic characteristics of the childbirth (age, marital status, level of education, occupation ) , information related to the household (size of household and type of house), sleeping under LLINs during pregnancy and information relating to the child from the previous pregnancy [ inter-reproductive space (SAE) , sex, duration of pregnancy, place of birth, state of child on the day of the investigation, length of life or age of death and probable cause of death].

Data analysis
The data was encoded in Access 2016 and imported into version 23 of the SPSS software for cleaning and analysis. Categorical variables were summarized as frequency and proportion and quantitative variables as mean or median depending on whether the distribution was symmetrical or not. To identify the bills associated with infant mortality, we constructed simple and multivariate logistic regression models. The variables were imported into the multivariate regression model on the basis of a p-value less than or equal to 0.2 and / or on the basis of a biological plausibility. As a measure of association, we present the unadjusted and adjusted odds ratios with their 95% con dence intervals. The threshold for statistical signi cance was set at 5% type 1 error (α). The Table 2 shows the pathology most frequently associated with death. Malaria is the most frequently reported pathology (12.7%, n = 18), followed by neonatal infections (15.5 %, n = 22), diarrheal disease s (10.3%, n = 15) and infections upper respiratory tract (6.3 %, n = 9).

Discussion
Our study indicated an infant mortality rate of 49.76‰ overall. The factors associated with this mortality are the age of the mother under 20 years old, the household size greater than or equal to 7, prematurity, home birth, the inter-reproductive space of less than 12 months and nally no. use of LLINs.
The infant mortality rate obtained in our study (49.76‰) remains lower than the infant mortality rate at the national level in 2015 (58‰) [1] and at the provincial level of South Kivu (92‰) [1]. The areas concerned by our study are two relatively stable areas from a security and humanitarian standpoint at present, with in particular more than 80% of deliveries are attended by staff. In addition, the health structures are well organized and accessible to the population. The province's overall estimate of the mortality rate is high, this would be justi ed by the fact that there are areas where the health situation remains deplorable as a result of the crisis context making basic structures inaccessible. In South Kivu, some basic health structures have been attacked by armed groups [6]. Despite this decrease, mortality remains high, justifying efforts in terms of reducing infant mortality and geographic inequalities.
The young age of the mother (less than 20 years) is associated with a risk of infant mortality of about 2 times (p = 0.022) unlike a maternal age of 35 years or more. These results corroborate with those from Brazil and France which showed that the age of less than 20 years was associated with infant death [10,11].
Prematurity is a factor strongly associated with mortality. Prematurity is responsible for more than 290,000 deaths per year in sub-Saharan Africa. According to the WHO, premature infants are 13 times more likely to die than term infants [12]. Other studies have con rmed this association [11,13,14,15,16].
The inter-reproductive space (EIG) has a great in uence on infant mortality (p < 0.001); the more it is less than 12 months the more the risk of death increases. Our results are similar to those of Naoko Kozuki et al who found that the EIG shorter s introduced him in t increased likelihood of neonatal mortality and under-ve [17]. The longer the inter-reproductive interval, the greater the chance of survival for the child. An inter-reproductive interval of less than two years not only leads to weaning of the preceding child but also to poor nutritional status and weakening.
Home birth is associated with the death of children. This result is in agreement with those of a study conducted in Uganda [18]. In addition, our results contradict those of Kambale et al who show that there was no difference in the risk of death between birth at home and that of maternity [19] as well as those of Johan et al who Neither have they demonstrated a signi cant association between childbirth in a sanitary setting by a health professional and reduction in neonatal death [20]. Home births often take place in precarious hygienic conditions and rst aid is not administered. This would explain the predisposition of children to diseases and therefore a high mortality.
Non-use of LLINs was associated with infant mortality (p < 0.001). This corroborates the data of Victoria et al [18] pointing out that all perinatal deaths occurred in women who did not sleep under a mosquito net. Malaria is the second most suspect pathology of infant mortality in our study (12.7%). In 2015, worldwide, the number of malaria cases was estimated at 2.4 million with 438,000 deaths [21]. Africa alone has recorded 1.88 million cases with 395,000 deaths [21]. The non-use of LLINs would explain this proportion of malaria among the pathologies suspected of infant mortality, but also the fact that we are in an endemic area with malaria which is the most frequent pathology in the whole country. Malaria, together with diarrheal diseases and respiratory tract infections constitute the most determining pathologies of infant and juvenile death in the DRC [22].
This work has strengths and limitations. The main strength of this work is its relatively large sample size. This study is one of the very few to study, with such a large sample size, the level and factors associated with infant mortality in post-crisis rural areas of eastern DRC. The limitations of this work include the fact that it is based on data collected from health facilities and the generalization of these results to the entire population is therefore limited. In fact, due to logistical di culties, we reasonably limited our sampling to health areas close to general referral hospitals in these two health areas. This may have introduced a selection bias in that people living further from general hospitals may have poor accessibility to health services and therefore a higher risk of infant mortality compared to those living closer to general referral hospitals. Another limitation of this work is that it is based on data collected from health facilities. It therefore does not allow data to be collected from mothers who had given birth at home. Finally, some important factors have not been studied, including vaccination, pregnancy monitoring and the course of childbirth.

Conclusion
The infant mortality rate in rural areas of health and Miti-Murhesa Walungu in South Kivu for the period 2016-2017 remains one of the highest in the world, which as less than the national average. This infant mortality remains associated with indicators of the low socioeconomic level, in particular the age of the mother under 20, household size greater than or equal to 7, prematurity, birth at home, inter-reproductive space of less than 12 months and the non-use of LLINs. The funder played no role in the study design, data collection, analyzes, interpretation or drafting of the manuscript.

Availability of data and materials
All relevant data are within the manuscript.
Ethical approval and consent to participate Ethics approval was granted by the Ethics Committee of the Catholic University of Bukavu. The principle of con dentiality was observed and the informed consent of the participants was obtained before participation in this study. For participants under the age of 18, informed consent has been obtained by the parent and / or legal guardian.
All procedures performed were in accordance with the ethical standards of the institutional ethical committee and with the 1964 Helsinki declaration and its later amendments.

Consent to publication
Not applicable, as no data is reported on an individual basis.

Competing interests
The authors declare that they have no competing interests.
Author details