Department of Statistics, University of Namibia, Windhoek, Namibia.
BMJ Open. 2012 Oct 21;2(5). doi: 10.1136/bmjopen-2012-001421. Print 2012.
To estimate and quantify childhood mortality, its spatial correlates and the impact of potential correlates using recent census data from three sub-Saharan African countries (Rwanda, Senegal and Uganda), where evidence is lacking.
Cross-sectional.
Nation-wide census samples from three African countries participating in the 2010 African Census round. All three countries have conducted recent censuses and have information on mortality of children under 5 years.
111 288 children under the age of 5 years in three countries.
Under-five mortality was assessed alongside potential correlates including geographical location (where children live), and environmental, bio-demographic and socioeconomic variables.
Multivariate analysis indicates that in all three countries the overall risk of child death in the first 5 years of life has decreased in recent years (Rwanda: HR=0.04, 95% CI 0.02 to 0.09; Senegal: HR=0.02 (95% CI 0.02 to 0.05); Uganda: HR=0.011 (95% CI 0.006 to 0.018). In Rwanda, lower deaths were associated with living in urban areas (0.79, 0.73, 0.83), children with living mother (HR=0.16, 95% CI 0.15 to 0.17) or living father (HR=0.38, 95% CI 0.36 to 0.39). Higher death was associated with male children (HR=1.06, 95% CI 1.02 to 1.08) and Christian children (HR=1.14, 95% CI 1.05 to 1.27). Children less than 1 year were associated with higher risk of death compared to older children in the three countries. Also, there were significant spatial variations showing inequalities in children mortality by geographic location. In Uganda, for example, areas of high risk are in the south-west and north-west and Kampala district showed a significantly reduced risk.
We provide clear evidence of considerable geographical variation of under-five mortality which is unexplained by factors considered in the data. The resulting under-five mortality maps can be used as a practical tool for monitoring progress within countries for the Millennium Development Goal 4 to reduce under-five mortality in half by 2015.
利用最近来自撒哈拉以南非洲三个国家（卢旺达、塞内加尔和乌干达）的人口普查数据，评估和量化儿童死亡率及其空间相关性，并定量分析潜在相关因素的影响。这些国家的相关证据都很缺乏。
横断面研究。
参与 2010 年非洲人口普查的三个非洲国家的全国性人口普查样本。这三个国家最近都进行了人口普查，并掌握了 5 岁以下儿童死亡率的信息。
三个国家的 111 288 名 5 岁以下儿童。
与潜在相关因素（儿童生活地点）以及环境、生物人口统计学和社会经济变量一起评估 5 岁以下儿童死亡率。
多变量分析表明，在所有三个国家，5 岁以下儿童的总体死亡风险在近年来都有所下降（卢旺达：HR=0.04，95%CI 0.02 至 0.09；塞内加尔：HR=0.02（95%CI 0.02 至 0.05）；乌干达：HR=0.011（95%CI 0.006 至 0.018））。在卢旺达，与居住在城市地区（0.79、0.73、0.83）、与活着的母亲（HR=0.16，95%CI 0.15 至 0.17）或活着的父亲（HR=0.38，95%CI 0.36 至 0.39）的儿童相关的死亡率较低。与男童（HR=1.06，95%CI 1.02 至 1.08）和基督教儿童（HR=1.14，95%CI 1.05 至 1.27）相关的死亡率较高。与 3 个国家的较大年龄儿童相比，1 岁以下的儿童死亡风险更高。此外，还存在显著的空间差异，显示了儿童死亡率的地理位置不平等。例如，在乌干达，高危地区位于西南部和西北部，而坎帕拉区的死亡率则显著降低。
我们提供了明确的证据，表明五岁以下儿童死亡率存在相当大的地域差异，这是数据中考虑的因素无法解释的。由此产生的五岁以下儿童死亡率图可作为一个实用工具，用于监测各国在实现千年发展目标 4 方面的进展情况，即在 2015 年之前将五岁以下儿童死亡率减半。
