LEVELS, TRENDS AND DIFFERENTIALS IN UNDER-FIVE MORTALITY IN NIGERIA (1990-2008)

Abstract

High Under-five Mortality (U5M) rate in Nigeria constitutes a source of public health concern. However, statistics for monitoring its levels, trends and differentials to assess the effectiveness of intervention strategies are unreliable due to inconsistencies in available estimates. Derivation of robust and reliable estimates for this important health indicator has become imperative. This study employed indirect demographic techniques to provide refined estimates of U5M levels, its trends and differentials in Nigeria. The Nigerian Demographic and Health Surveys (1990, 1999, 2003 and 2008) data on women of childbearing age (57,723) and under-five children (46,130) were analysed and weighted to adjust for the effects of multistage cluster sampling. Levels of U5M were derived using the indirect demographic techniques which involved calculating proportions of children ever born and those who died. The proportions were converted into probability of dying using Brass/Trussel demographic models. Probability of dying was refined by linear regression using a one-parametre relational logit system and the general African standard life table. A linear interpolation was employed to derive U5M levels from the relined probabilities. Locally weighted scatterplot smoothing technique was used to assess trends in U5M levels. Absolute changes in maternal education, childhood immunisation, Oral rehydration Therapy (ORT), antenatal care, and other factors associated with child survival were also computed. Cox hazard model was fitted to determine the effects of these factors using estimates of Hazard Ratio (HR) and their 95% Confidence Intervals (CIs). Levels of U5M were 167,200, 186 and 143 per 1000 live births for 1990, 1999, 2003 and 2008 respectively. Over the 19 years, U5M was consistently highest in the North-West (1990=217; 2008=182) and least in the South-West (1900=124; 2008=111). It was higher among children in rural areas (1990=181; 2008=163) and those born to mothers without formal education (1990=195; 2008=176). Mortality trends rose by 26.9% between 1990 and 1999 but declined by 23.1% from 2000 to 2008. Decline between 1990 and 2008 was greater in urban (17.3%) than rural areas (14.2%) and fastest in South-South region (62.8%). During the period, there was improvement in immunization (17.7%), medical treatment of childhood illness (17.5%), ORT (13.9%), maternal education (8.6%), and bednet usage (7.1%). Access to improved source of drinking water (24.2%), improved toilet facilities (9.0%) and antenatal care (4.5%) declined while maternal age at childbirth and childhood malnutrition remained unchanged. Antenatal care (HR=0.83; CI=0.62-1.10), contraception (HR=0.71; CI=0.52-0.97), bednet usage (HR=0.83; Cl=0.59-1.16) and urban residence (HR=0.82; CI=0,62-1.10) promoted child survival. Children with birth intervals ≤24 months were 3.3 times more Iikely to die before age five than those with birth interval ≥36 months (HR=3.3 I; CI=2.42-4.53). The risk under-five death was also significantly higher for multiple birth children (HR=8.0; CI=5.64-11.38). This work has provided plausible estimates for a clearer description of the under-five mortality trajectory in Nigeria. It has shown a gradually declining trend, slower in the north than the south. It also identified maternal education, maternal and child health behaviours and access to potable water as variables influencing the trends. Interventions targeting these factors should be intensified.