The study aims to develop a novel quantile regression models to identify critical risk factors of under-five severe stunting (severe chronic malnutrition). The study also attempts to examine the effects of risk factors on different quantiles of HAZ to inform appropriate malnutrition policies and intervention strategies.
Analyses were conducted on 2716 children out of which 144 (5.3%) were identified as severely stunted. Critical risk factors associated with severe stunting were type of birth, sex, age, place of delivery, size at birth, maternal age and education, maternal national health insurance status, household wealth status, and number of children under-five in households.
The use of simultaneous quantile regression to model severe stunting (i.e. 0.10 and 0.20 quantiles of HAZ) as well as ranges of 0.10 to 0.90 quantiles of HAZ as opposed to ordinary least squares regression is of paramount interest in this study. The simultaneous quantile regression modelling approach employed in this study provided a richer characterization of the data, thereby revealing the effect of a covariate on the entire distribution of HAZ, making it possible to identify the more vulnerable groups and to formulate more effective interventions to these groups.
The results have shown that quantile regression models provided much more information about the underlying associations better than the ordinary least squares regression, suggesting that the conditional distributions of HAZ did not only differ by their means but also by their lower and upper tails as reported in previous studies[20-24]. Thus, the ordinary least squares regression missed critical aspects of the associations that exist between the conditional distributions of HAZ and its determinants as shown in Figure 1. For example, the impact of child age and the number of children below 5 years in households on HAZ was higher at the lower quantiles of HAZ and lower at the higher quantiles of HAZ. This finding suggests that children at the lower tail of the HAZ distribution who are more likely to be severely stunted benefit more from interventions aimed at addressing issues surrounding child age and number of children below 5 years in households than those in the upper end of the distribution. Thus, child age and number of children below 5 years in households interventions to address under-five severe stunting could have more impacts on children who are at higher risk of severe stunting. Also, the impact of multiple births and mothers with no formal education on HAZ was lesser at the lower tail but higher at the upper tail, suggesting that children at the lower tail of the HAZ distribution who are more likely to be severely stunted benefit less from interventions aimed at addressing issues surrounding multiple births and improvement in maternal education than those in the upper end of the distribution.
The study broadly supports earlier studies that examined determinants of under-five malnutrition in developing countries. Multiple births, older children, been delivered at home, born small at birth, increase in number of children aged <5 years in households, belonging to poor households, and having mothers with no formal education increased the risk of under-five severe stunting. Factors that were protective of under-five severe stunting include female child, older ages of mothers, and having a mother who had health insurance[1, 4, 28, 35-42].
Children belonging to mothers with no formal education had increased risk of severe stunting compared to their counterparts with primary education or higher. Increase in the level of maternal education has been shown to be protective of child malnutrition[1, 4, 40, 43, 44]. The plausible explanation could be that mothers who had better education may have knowledge about better childcare and childhood feeding practices, better health seeking behaviour, and judicious in the management of the limited household resources. Well educated mothers are expected to be more conscious about their children nutrition and health[19, 36-38, 40, 43, 45].
The finding that children from poor households are more likely to be severely stunted compared to their counterparts from average/rich households is consistent with previous studies[1, 36, 38, 40]. This could be as a result of low quality and inadequate food consumption, exposure to infections and lack of access to health services commonly associated with poor households. Thus, the effect of poverty could be manifested through lower purchasing power, low literacy, and food insecurity resulting in poor nutritional outcomes among poor households [1, 16, 36, 37, 43].
This study supports previous studies that reported that children from Mothers who had health insurance had lower risk of severe stunting compare to those who had mothers with no health insurance[1, 44, 46]. The possible explanation could be that having national health insurance could propel mothers to seek better health care for themselves and their children, visit health facilities and this is expected to result in improved health outcomes for children and their mothers[1, 47, 48]. The association observed among children from older mothers and the decrease in severe stunting suggests encouraging adult motherhood as young mothers are less likely to have good knowledge on appropriate health care and feeding practices of their children. Also, young mothers need adequate nutrition for their optimal growth and that of their children but this is not always the case as mothers have to share the little portion of food with their babies[40, 49-51]. An increase in the likelihood of severe stunting found among children from households with higher number of children below 5 years could be due to increased competition for household food with a high number of people and limited resources to respond to health needs of household members. Households with more children are generally socioeconomically disadvantaged coupled with poor quality of life, suggesting the need for family planning and improvement in socioeconomic conditions of households[18, 19, 40, 41, 52, 53].
The finding that the risk of severe stunting was higher among children who were born multiple compared to those born singletons is consistent with previous studies[1, 28, 42]. This could be attributable to premature births, low birthweight, cerebral palsy or competition for nutritional intake commonly observed among children who are products of multiple births[1, 28, 42, 54] which has the tendency of limiting child growth. Consistent with previous studies, female children had decreased risk of severe stunting compared to their male counterparts[28, 38, 42]. Generally, female children are less likely to be influenced by environmental adverse effects than male children. Thus, gender inequality in childhood malnutrition is more likely to be observed in hostile environments such as diseases and exposure to environmental pollutants. Apart from epidemiological evidence that male children are biologically more vulnerable to morbidity[55, 56], gender-based cultural practices and perceptions could influence the behavioural patterns of care-givers in which female children are given preferential feeding treatments[57, 58].
Increase in the likelihood of severe stunting observed among older ages of children in this study supports previous studies[1, 4, 28, 40, 42, 59]. A plausible explanation could be a deficit in timely and adequate complementary feeding and presence of progressive childhood diseases[1, 4, 28, 59]. The finding that the children who were born small at birth had higher likelihood of severe stunting is consistent with previous studies. Size of child at birth is a proxy for child birthweight and that low birthweight is a risk factor for under-five child malnutrition [1, 60-62]. Children born at non-health facility had higher likelihood of severe stunting compared to their counterparts delivered at health facilities, a finding which is consistent with previous studies[35, 60, 63, 64]. A plausible reason could be that mothers who delivered at non-health facilities were not assisted by a skilled provider[35].
The strengths of this study include the fact that it utilised data from a nationally representative population-based survey which is globally respected for its sound survey methods and sound data quality on children, their households and communities in which they live. The large samples drawn nationwide permits generalization of findings to the population of children under-five in Ghana and that of children from other similar populations globally. The study also used a novel simultaneous quantile regression approach, permitting the study of covariate effects across different quantiles of HAZ. Thus, providing much more information about the underlying associations between HAZ and the risk factors which could not have been possible using ordinary linear regression approach. Despite these strengths, the study has a limitation. The survey employed cross-sectional study design which could not establish cause and effect relationship between severe stunting and risk factors considered. Also, it is very difficult measuring income of households in developing countries like Ghana. As a result, an asset-based wealth index commonly used as a proxy for household income was used in this study.
One of the major public health issues in Ghana is under-five child malnutrition with the prevalence of stunting been the highest. The findings in the present study provided vital and current information on critical risk factors for under-five severe stunting that can be used by policymakers and health practitioners for better understanding of under-five severe stunting and its prevention and management. The study highlights the need for improvement in maternal education and general socioeconomic conditions of households, intensification of family planning education and its services, encouraging adult motherhood and targeting nutrition and health policies and intervention strategies at teenage mothers and their children, educating and encouraging women to delivery at health facilities and supporting women and their children with free national health insurance premiums can be helpful in reducing the risk of severe stunting and its associated morbidity and mortality.