Heterogeneous association between maternal haemoglobin concentration and child haemoglobin levels: a quantile regression analysis using multi-country nationally representative data.

Background: Low haemoglobin concentration, or anaemia, is a major global public health issue, particularly severe in sub-Saharan Africa, including East Africa, where prevalence rates are alarmingly high. This situation requires a comprehensive analysis to identify the key factors contributing to this high prevalence, especially in East Africa. This study uses robust quantile regression to examine the relationship between maternal haemoglobin concentration and child haemoglobin levels in three East African countries. Methods: The study analysed data from the Demographic and Health Surveys of three East African countries: Ethiopia, Tanzania, and Uganda. These data are nationally representative and are collected every five years from low- and middle-income countries. The analysis included a total sample of 6883 children and their mothers. The primary outcome variable was the haemoglobin concentration of the child. We used quantile regression to model the effects of maternal haemoglobin concentration on child haemoglobin levels, adjusting for potential confounders. Results: A one-unit increase in maternal haemoglobin concentration was associated with a unit increase in child haemoglobin concentration as follows: In Tanzania, the increases were 0.19, 0.20, 0.19, and 0.17 at the 5th, 10th, 25th, and 50th quantiles, respectively. In Ethiopia, the increases were 0.23, 0.22, 0.22, and 0.17 at the same quantiles. In Uganda, the increases were 0.19, 0.19, 0.18, and 0.19 at these quantiles. In contrast, a one-unit increase in maternal Hb concentration was associated with a smaller increase in child haemoglobin concentration, ranging from 0.11 to 0.13 at the 75th and 90th quantiles across all three countries. Overall, the most significant effects of maternal Hb concentration on child Hb levels were observed at the lower end of the conditional Hb distribution in each of the three countries. Conclusion: Maternal haemoglobin concentration is positively associated with children's haemoglobin levels across all countries studied. The strongest association is among children in the lower quantiles, who are at higher risk of anaemia. In contrast, the association is less pronounced among those in the upper quantiles. Therefore, improving maternal haemoglobin is likely to benefit children in the lower quantiles more than those in the upper quantiles. Programmes aimed at addressing child anaemia should consider these differing effects of maternal haemoglobin concentration. Keywords: Haemoglobin concentration, anaemia, maternal, child, quantile regression