Menu
Menu
Menu
Menu
Background: From 1995-2000 the under five mortality rate in Uganda increased from 147.3 to 151.5 deaths per 1000 live births and reasons for the increase were not clear. This study was undertaken to understand factors influencing the increase in under five mortality rate during 1995-2000 in Uganda with a view of suggesting remedial actions. Methods. We performed a comparative retrospective analysis of data derived from the 1995 and the 2000 Uganda demographic and health surveys. We correlated the change of under five mortality rate in Uganda desegregated by region (central, eastern, north and western) with change in major known determinants of under five mortality such social economic circumstances, maternal factors, access to health services, and level of nutrition. Results: The increase in under five mortality rate only happened in western Uganda with the other 3 regions of Uganda (eastern, northern and central) showing a decrease. The changes in U5MR could not be explained by changes in poverty, maternal conditions, level of nutrition, or in access to health and other social services and in the prevalence of HIV among women attending for ante-natal care. All these factors did not reach statistical significance (P > 0.05) using Pearson’s correlation coefficient. Conclusion: In order to explain these findings, there is need to find something that happened in western Uganda (but not other parts of the country) during the period 1995-2000 and has the potential to change the under five mortality by a big margin. We hypothesize that the increase in under five mortality could be explained by the severe malaria epidemic that occurred in western Uganda (but not other regions) in 1997/98. © 2011 Nuwaha et al; licensee BioMed Central Ltd.
Uganda is a low-income country by all indicators with current projected population of 32 million people (from the 2002 national census at annual population growth rate of 3.4%). The social-economic development has been characterised by political upheavals since 1970s, resulting into two major wars during 1978/79 and during 1981-1986. During this period there was a reversal in social-economic development with shrinkage of gross domestic product per capita (GDP) in terms of purchasing power parity (PPP) from US$ 615 in 1969 to 443 in 1980 [11]. This trend appears to have been reversed in 1986. Since then the economy has been growing at an annual rate of over 5% far ahead of population increase estimated at about 3% during the same period [9]. The economy of the country is predominantly dependant on agriculture for more than 80% of the employment. Land ownership is almost universal in rural areas where more than 87% of the population live. As a result the gini coefficient (which is a measure of income distribution) in the country is favourable and during the 1995-2000 period varied from 0.35 to 0.38 [12]. The major factors influencing childhood mortalities in the country include maternal conditions (such as education, parity, age) birth order, nutritional status of the child, place of residence (rural or urban), HIV prevalence rates among pregnant women, malaria endemicity, wealth of households and place of delivery for newborns [7,8]. The fundamental direct causes of childhood mortalities include: peri-natal conditions (such as pre-maturity, low birth weights and level of supervision during child birth), malaria, diarrhoea, pneumonia, HIV/AIDS, malnutrition and measles. These 7 conditions are responsible for more than 90% of the total childhood mortalities [10]. The HIV infection rate in the country was highest in the early 1990 but has since started declining and it now stands at less than 7% of the adult population [13]. Uganda broadly has two types of malaria transmission whereby about 90% of the country lies in stable malaria transmission (predominantly in the eastern, northern and central parts of the country) and about 10% (that is predominantly in the western region of the country) is characterised by unstable malaria transmission and prone to epidemics (see figure figure1)1) [14]. Although the land area of unstable malaria transmission is only about 10%, the population density of malaria free areas and low transmission areas in western Uganda is very high. As such about one fifth of the Ugandan population on the whole live in either malaria free or low transmission areas in western Uganda [14,15]. Malaria in low transmission areas of western Uganda is characterised by epidemics. The worst malaria epidemic characterised by very high childhood mortalities occurred in western Uganda in 1997/1998 and was greatly influenced by the el-Niño weather phenomenon [16-18]. Map of Uganda showing malaria endemicity. Data were abstracted from the 1995 and 2000 Uganda demographic and health surveys [7,8]. The UDHS of 1995 and 2000 were designed to have adequate sample sizes (7070 and 7246 women aged 15-49 respectively) proportionate to population of the regions to allow for estimation of childhood mortality indices by the four regions of Uganda. In both surveys, data were collected on characteristics of household members, socio-economic status of respondents and of households, fertility regulation, determinants of fertility, fertility preferences, reproductive health and child care, nutritional status of children, morbidity in previous two weeks, adult mortality, and HIV/AIDS. Both surveys used the same methodology and collected data on retrospective birth histories which provided direct estimates of childhood mortality [19]. The analysis of 1995 and 2000 UDHS data included disaggregation of under-five mortality data by age (e.g. within first month of life (neonatal mortality rate-NMR), between second month of life and before 12 months of life (post neonatal mortality-PNMR), during the first year of life (infant mortality rate-IMR), during years 1-4 of life (child mortality rate-CMR) and during the first five years of life (under five mortality rate-U5MR). The mortality data was also disaggregated by the four regions of Uganda (Central, Eastern, Northern and Western). Data were also abstracted on possible covariates of childhood mortality such as female education, proportion of population living below the poverty line, proportion of mothers delivering under medical supervision, percentage of mothers with high risk pregnancies, rates of usage of modern family planning, rates of measles immunization, HIV infection among antenatal mothers, childhood malnutrition, and incidence of fever, diarrhoea and cough/rapid breathing within the previous two weeks, and with the occurrence or non occurrence of the malaria epidemic in 1997/1998. Economic indicators such as proportion of population living below the poverty line were derived from Uganda national and household surveys (UNHS) of 1992 and of 1999 [12,20]. These surveys had sample sizes between 5000-10,000 households and were designed to capture regional differences particularly among proportions of the population living below the poverty line. The rates of HIV infection among women attending antenatal care (ANC) were derived from surveillance data which has been collected from a representative sample of women in various health units in the country since 1989 [21]. Other measures were obtained from the UDHS of 1995 and 2000 [7,8]. Ninety five percent confidence intervals (95% CI) of mortality rates were used to gauge whether there were significant changes in mortality indices between 1995 UDHS data and that of 2000. We further analysed association between changes of various determinants of U5MR by region with changes in U5MR between 1995 and 2000 based on the method of concomitant variation [22]. This method is used to ascertain the relationship between two variables to establish causality based on the fact that if two phenomena vary up and down simultaneously, one is causing the other or there is a third factor causing both of them [23]. The deviations of an oscillatory variable with respect to its rate of change, measures both the size and the direction of its change over time. Therefore if two variables oscillate simultaneously, because one causes the other or a third factor causes both, their rates of change will be highly correlated. In order to apply the method of concomitant variation, the variables were transformed into rates of change (i.e. the ratio X2000-X1995/X1995)*100 expressing the relative change from the year 1995 to the year 2000 into a percentage. Correlations between the transformed variables and the change in U5MR by region were then computed. Significance of change in determinants with change in U5MR were tested using Pearson’s correlation coefficient (r) using a two tailed test at 5% level of significance. The Uganda National Council for Science and Technology (UNCST) and the Makerere University Institute of Public Health (MUIPH) higher degrees and ethics committee independently approved the study. Prior to data collection permission was sought from the relevant Uganda government authorities.
N/A
