Menu
Menu
Menu
Menu
Neonatal (NMR), infant (IMR) and under-five (U5M) mortality rates remain high in Nigeria. Evidence-based knowledge of trends and drivers of child mortality will aid proper interventions needed to combat the menace. Therefore, this study assessed the trends and drivers of NMR, IMR, and U5M over a decade in Nigeria. A nationally representative data from three consecutive Nigeria Demographic and Household Surveys (NDHS) was used. A total of 66,158 live births within the five years preceding the 2003 (6029), 2008 (28647) and 2013 (31482) NDHS were included in the analyses. NMR was computed using proportions while IMR and U5 were computed using life table techniques embedded in Stata version 12. Probit regression model and its associated marginal effects were used to identify the predisposing factors to NMR, IMR, and U5M. The NMR, IMR, and U5M per 1000 live births in 2003, 2008 and 2013 were 52, 41, 39; 100, 75, 69; and 201, 157, 128 respectively. The NMR, IMR, and U5M were consistently lower among children whose mothers were younger, living in rural areas and from richer households. Generally, the probability of neonate death in 2003, 2008 and 2013 were 0.049, 0.039 and 0.038 respectively, the probability of infant death was 0.093, 0.071 and 0.064 while the probability of under-five death was 0.140, 0.112 and 0.092 for the respective survey years. While adjusting for other variables, the likelihood of infant and under-five deaths was significantly reduced across the survey years. Maternal age, mothers’ education, place of residence, child’s sex, birth interval, weight at birth, skill of birth attendant, delivery by caesarean operation or not significantly influenced NMR, IMR, and U5M. The NMR, IMR, and U5M in Nigeria reduced over the studied period. Multi-sectoral interventions targeted towards the identified drivers should be instituted to improve child survival.
The Institutional Review Board (IRB) of the National Institute of Medical Research, Nigeria approved the study protocol, survey instrument, and materials prior to the commencement of the surveys. Details of the ethical approvals have been reported earlier [17]. Informed consent was obtained from all parents and guardians who participated in the surveys. Nigeria consists of 6 geopolitical regions; North-East, North-West, North-Central, South-East, South-South, and South-West which are sub-divided into 36 administrative states and the Federal Capital Territory (FCT). The population in each of the geopolitical regions and states are relatively homogeneous and share similar socio-cultural characteristics. Also, health-related characteristics like access to health care, environment, housing system etc. are similar within the regions and states. We pooled data from three consecutive nationally representative Nigeria Demographic and Household Surveys (NDHS) in 2003, 2008 and 2013. The survey uses three-stage sampling technique to select the respondents. Firstly, Local Government Areas (LGAs) are selected, then the Enumeration Areas (EA), which are the Primary Sampling Units (PSU) and referred to as clusters and lastly the selection of households within the selected EAs. Primary information about households, sexual and reproductive health and history were collected from women aged 15–49 years within the selected households. Usually, the survey collects birth history of all women interviewed. More specifically, the survey collects information on all births to a woman. We, therefore, used the “child recode data” which contains all follow-up information on all children born to the interviewed women within five years preceding the survey. Among the 7620, 33385 and 38948 women who participated in 2003, 2008, and 2013 surveys respectively, there were 6029, 28647 and 31482 children born within five years preceding each of the surveys. All analysis in this study were therefore based on the survivorship of the 66158 children within first five years of their life. There are three outcome variables in this study, they are neonatal deaths, infant deaths, and under-five children (U5) deaths. According to the NDHS, neonatal deaths, infant deaths, and under-five children (U5) deaths are deaths within the first 28 days, one year and five years respectively [18]. Based on past literature, the independent variables included in this study are: The groupings of the environmental characteristics were in tandem with those adopted in the 2013 NDHS [18] and the 2010 WHO and UNICEF document on progress on sanitation and drinking water [18]. The “source of drinking water” was grouped into either improved or not. Improved sources include piped into dwelling/yard/plot, public tap/standpipe, tube-well or borehole, protected well and spring, rain water, and bottle water. The improved toilet types are “flush/pour flush to piped sewer system”, “flush/pour flush to septic tank”, “flush/pour flush to pit latrine”, “ventilated improved pit (VIP) latrine”, “pit latrine with slab or composting toilet” while any other types of toilet facilities were categorised as non-improved. Descriptive statistics were used to show the distribution of the under-five children by the studied characteristics in Table 1. We then computed the NMR using proportions while IMR and U5M were computed using life table techniques embedded in Stata version 12 as presented in Table 2. Bivariate analyses were carried out to determine the significant association between each of the outcome variables and the independent variables using Pearson Chi-square (x2) test of association and also presented in Table 2. Probit regression model was used to identify the predisposing factors to neonatal death, infant mortality and under-five mortality. In probit regression model, attempt is made to model the (conditional) probability of a “successful” outcome, that is, It is expressed as a derivative of Eq (1) as shown in Eq (2) where Φ(·) is the cumulative distribution function of the standard normal distribution. That is, conditional on the explanatory variables, the probability that the outcome variable, Yi = 1, is a certain function of a linear combination of the explanatory variables. A positive regression coefficient indicates that an increase in the predictor leads to an increase in the predicted probability while a negative coefficient is an indication that t an increase in the predictor would reduce the predicted probability. We provided the marginal effects of the explanatory variables. The marginal effects estimated using the “delta method” involves the use of calculus to show how much the (conditional) probability of the outcome variable changes when there is a change in the value of an explanatory variable, holding all other explanatory constant at their values. It is worth noting that unlike the linear regression case where the estimated regression coefficients are the marginal effects, there is a need for the additional level of computation to estimate the marginal effects haven computed the probit regression. In the case of a discrete explanatory variable, the change in the probability is Sampling weights were applied, statistical significance was determined at 5% and Stata 12 used for all analysis while multicollinear variables were removed in the final model. There are four distinct columns in the Tables Tables3,3, ,4,4, ,55 and and6.6. The first column is the marginal effects computed from the coefficients of the probit model. It shows changes in a particular category with respect to the reference category. The second column is the standard error of the estimate in the 1st column while the 3rd column is the associated p-value. However, the 4th column presented the estimated increase/reduction per 1000 live births with respect to the reference category. Mar Eff Marginal Effect, SE Standard Error Sig. Significance MPT Mortality per 1000 Livebirths to Differences in Mortality per 1000 Livebirths Mar Eff Marginal Effect, SE Standard Error Sig. Significance MPT Mortality per 1000Livebirthto Differences in Mortality per 1000 Livebirths Mar Eff Marginal Effect, SE Standard Error Sig. Significance MPT Mortality per 1000 Livebirths to Differences in Mortality per 1000 Livebirths Mar Eff Marginal Effect, SE Standard Error Sig. Significance MPT Mortality per 1000 Live births to Differences in Mortality per 1000 Livebirths
N/A
