Modelling child anaemia and co-existing infections using log-linear models

Area of Science:

• Public Health
• Epidemiology
• Child Health

Background:

• Uganda faces a high burden of anaemia, malaria, and fever, with prevalence rates up to 33%, 34%, and 37% respectively in certain regions.
• Existing health initiatives have attempted to tackle these conditions, but the complex interplay between malaria, fever, and anaemia in young children remains under-examined.

Purpose of the Study:

• To investigate the patterns and associations between malaria, fever, and anaemia in Ugandan children.
• To analyze these relationships while controlling for the influence of maternal anaemia using log-linear models.

Main Methods:

• Secondary data from the 2018-2019 Uganda Malaria Indicator Survey (MIS) was utilized.
• A sample of 7,124 children aged 0-60 months was analyzed using a two-stage cluster and household sampling design.
• Five log-linear models (saturated, mutual independence, joint independence, conditional independence, and homogeneous) were fitted to assess data fit and identify significant associations.

Main Results:

• The homogeneous log-linear model demonstrated an adequate fit for the data (G² = 10.24, p = 0.76), indicating no significant lack of fit.
• Children who tested positive for malaria were found to be twice as likely to have anaemia compared to those who tested negative.
• The study identified significant interactions between malaria, fever, and anaemia, with maternal anaemia identified as a critical contextual factor.

Conclusions:

• The findings highlight the significant interconnectedness of malaria, fever, and anaemia in Ugandan children, emphasizing the syndemic nature of these conditions.
• Integrated interventions addressing maternal health, malaria prevention, and nutritional support are crucial for reducing the burden of these intertwined diseases.
• Log-linear modeling proved effective in uncovering complex patterns and interactions, supporting the development of targeted public health strategies.