Socio-economic and demographic determinants of under-five mortality in rural northern Ghana

Under-five mortality decline has improved from 1.2% per year between 1990 and 1995, to 3.9% per year between 2005 and 2012 [1]. In spite of this substantial drop in global child mortality rate, about 6.6 million children still die every year before their fifth birthday worldwide which implies 18,000 under-five children die each day [1]. There are huge disparities in child mortality among low and middle income countries and the industrial world with Sub-Saharan Africa and South East Asia carrying the highest burden of under-five mortality [2‐4].

Children in sub-Saharan Africa have the highest risk of death in the first month of life and are still leading in under-five mortality rates with one in every nine children dying before their fifth birthday as of 2011 [2].

It is worth noting that the year 2008 recorded a rate of one in seven children (144 per 1000 live births) dying before their fifth birthday with the highest levels occurring in West and Central Africa. Within 34 countries where under-five mortality exceeded 100 out of 1000 live births in 2008, all except one are in sub-Saharan Africa [5]. The rate of improvement in child survival in Sub-Saharan Africa is insufficient to meeting United Nations Millennium Development Goal 4 of reducing under-five mortality rate by two-thirds between 1990 and 2015 as it has the highest risk of death in the first month of life and is among the regions showing the least progress globally [1, 2].

In order to accelerate the decline in under-five mortality rate, specific proven interventions would have to target important causes of child death [6‐8]. Since no single factor can account for the high child mortality [9], in developing these interventions there is the need to understand the multiplicity of factors that determine child mortality especially in resource poor settings [10, 11].

Previous studies have shown that various factors influence child health and survival including place of residence, mothers age, mother’s education, place of delivery, birth order, sex of child, religion of parents, household headship and household socio economic status [9, 12‐18].

Though poverty is well acclaimed as an essential factor influencing child mortality [4, 14, 19, 20], findings on the effect of household socio economic differentials on child mortality have been mixed. A study in parts of rural Ghana and another in Tanzania did not find any significant effect of household socio economic status on child mortality [18, 19] while a study using Nigeria Demographic and Health Survey for 2008, found that relatively prosperous households were less likely to experience child death than the poorest households in rural Nigeria [13].

Existing literature has documented mixed results, such as place of delivery, birth order and sex of child. For instance, abundance of evidence suggest that women who deliver at health facilities have a lower chance of child death as compared to those who deliver at home due to the use of skilled delivery at health facilities and the none existence of such at home [6, 12‐14]. However, studies in South Africa and parts of Nigeria suggest that place of delivery does not have significant effect on either perinatal or under-five mortality [16, 17].

In Ghana, just as in many other countries in Sub-Saharan Africa with low socio economic development, under-five mortality is relatively high with a recent reported national figure of 90 per 1000 live births [21]. Moreover within Ghana, disparities exist in under-five mortality rates between regions. In the more resource rich regions of southern Ghana, under-five mortality rate ranges from 75 per 1000 live births whiles in the most impoverish and deprived regions of the north such as the Upper East Region it is as high as 128 per 1000 live births [21]. Therefore there is the need for concerted efforts especially in resource poor settings if we hope to achieve the desired improvements in under-five mortality.

Due to disagreements in existing literature on the influence of particular social, economic and demographic attributes of mothers on under-five mortality coupled with sparse information on the topic in the Ghanaian setting, this paper further examine social, economic and demographic factors that influence under-five deaths and the magnitude of their influence in a rural impoverished setting in northern Ghana. It is hoped that findings from this paper will contribute to literature on under-five mortality in rural poor settings and perhaps contribute to identifying priority interventions to address the persistent mortality in settings such as this.

In this study we have examined the influence of particular social, economic and demographic characteristics of mothers on under-five mortality in rural northern Ghana. Results showed that several factors are implicated in under-five mortality.

The bivariate analysis shows that variables such as mother’s age, level of education, marital status, place of residence, occupation, religion, autonomy, contraceptive use, national health insurance enrollment status, polygamous union and socio economic status were significantly associated with experience of under-five death in this largely rural setting. Controlling for the confounding effects of the various factors, mother’s age, level of education, marital status and polygamy, were found to have significant robust association with childhood mortality.

Both multivariate and bivariate analysis indicate that children born to older mothers (35 years and above) were more at risk of under-five death compared to younger mothers (20 years of age), this is consistent with available literature that points to the fact that maternal age is a strong predictor of child survival [12, 14, 18]. Reasons for the childhood mortality disadvantage for older mothers may be twofold: first, older cohort mothers are less likely to be educated and therefore, less likely to be associated with the enhanced mortality advantage associated with education, and second, older mothers are more likely to be associated with higher parity.

Level of mothers’ education emerged as a strong predictor of under-five mortality in both bivariate and multivariate analysis after controlling for other variables; this has long been acknowledged by other scholars [4, 9, 14, 16] though a few studies suggest otherwise [12]. The effect of maternal education on child survival is over whelming and cannot be over emphasized. The reasons may be that educated mothers are more likely to receive antenatal care and are more capable of gaining access to health care. Also education can contribute to child survival by delaying marriage and motherhood hence decreasing the total number of children a woman has [12]. Warren’s findings from Ethiopia indicate that some women do not seek care due to lack of understanding of the benefits which has effects on child and maternal survival [26].

As presented in these findings, married women were less likely to experience child death than their unmarried counterparts and this was statistically significant even in the multivariate analysis. Not many studies have been done on the effect of mother’s marital status on child mortality but one can sense that mothers in stable marriages would get support from their partners during antenatal through to postnatal care which can reduce risk of child mortality as shown by our results. Also, marriage may confer advantages such as pooling of resources to either patronize good health services or provide adequate care with respect to providing good nutrition to infants and children.

Religious affiliation did not show significant odds in the multivariate analysis though the odds points to the fact that respondents who did not practice any religion and those who practice traditional religion had a higher chance of experiencing under-five death than Christians and Muslims. This is consistent with a study in parts of Nigeria which showed that children born to mothers who practice traditional religion had increased risk of about thrice than the general population to experience perinatal deaths [17].

Contrary to Ettarh and Kimani’s findings in Kenya and another study in Nigeria which showed that rural poor mothers have a higher likelihood of experiencing child death than their urban counterparts [12, 17], the findings of this study did not show any significant effect of place of residence on ever experiencing under-five death. It is likely the place of residence for some respondents could have changed between the time they experienced under-five death and the time of this survey. The existence of wide spread community based health planning and services program (CHPS) in the region could also partly be the reason for which place of residence did not confer significant odds on the experience of under-five deaths.

In spite of previous studies establishing that maternal autonomy promotes health-seeking behavior for sick children with positive outcomes on child survival [9, 13, 27], this study did not seem to convincingly show the result of this on child survival. In this study we expected that mothers who profess to be financially autonomous would have significantly less chance of losing an under-five child but this is not so, as financially autonomous women were only 2% less likely to experience under-five death and this was not even statistically significant at the multivariate level. This may be due to our measure of autonomy. While other studies look at autonomy largely from education and employment [27] others perceive autonomy on the basis of decision making and to a large extend household headship. There is hence the need for a more concise and acceptable measure of female autonomy.

Another area which most studies have not looked at is the effect of polygamy on child mortality. This study shows that children whose mothers are in polygamous unions have 22% higher chance of experiencing under-five death than those in monogamous marriages. This may be due to possible rivalry from co-wives and neglect from husbands during pregnancy and early child care which is typical of African polygamous marriages with consequential effects on child survival. It could also be due to the fact that in this predominantly agrarian communities where women do not work and often tend to rely on the little resources that are provided by their husbands, women who are in polygamous marriages will be doubly stretched with regards to resources because the little the husband has will have to be shared among the several wives.

The effect of contraceptive use on under-five mortality though not significant at the multivariate level, it points to a 14% less likelihood of contraceptive users experiencing under-five death and a significant association in bivariate analysis. This is in line with other studies which have previously indicated a positive association of contraceptive use to low child death [9]. Therefore family planning methods are well recommended as an effective tool for birth spacing, prevention of unwanted pregnancies and affording mothers adequate time to care for the newborn and regain strength before conceiving again.

Contrary to most previous studies that had highlighted the effect of socio economic factors on child mortality [4, 9, 12, 13, 15], socio economic status did not confer significant odds in the multivariate analysis. Owais et al. in examining why families of newborns do not accept hospital care in Pakistan identified largely financial reasons as a major barrier to access and utilization of health facilities [28]. Other studies have also indicated the role of socio economic differentials in health care seeking behavior with those of higher socio economic status more likely to use health care facilities than those at lower levels [28‐30]. Which would invariably have a positive impact on child survival, but this is not what we observed here. Some studies suggest that the presence of wide spread community-based health planning and services program (CHPS) in the region has the potential to reduce the effect of socio economic differential on child mortality [19, 23]. Studies in Vietnam and Nigeria showed that mothers from poor settings with transportation difficulty have a higher likelihood of child death [14, 17]. Therefore in spite of these findings, we recommend interventions that improve the economic wellbeing of women in this setting given the prevalence of poverty in the area.

The causes of high under-five mortality in resource poor settings are complex and merit concerted efforts to clarify their implications to improve child survival. On the basis of these findings, we suggest that specific efforts with focus on under-five mortality decline need to target the individual needs and welfare of women. In order of priority, the promotion of female education, gender equity, family planning and addressing the vast socio-economic differentials in northern Ghana are important social steps to improving under-five survival.