A Bayesian multinomial model to analyse spatial patterns of childhood co-morbidity in Malawi

Area of Science:

• Epidemiology
• Public Health
• Spatial Analysis

Background:

• Infectious diseases pose a significant threat to children in less developed countries.
• Co-morbidity of common childhood illnesses like fever, diarrhea, and pneumonia is a critical public health concern.
• Understanding the spatial distribution and risk factors of these co-morbidities is essential for effective intervention.

Purpose of the Study:

• To assess risk factors associated with childhood co-morbidity of fever, diarrhea, and pneumonia in Malawi.
• To quantify area-specific spatial effects on these co-morbidities.
• To evaluate the effectiveness of integrated disease control approaches.

Main Methods:

• Utilized self-reported illness data from the 2000 Malawi Demographic and Health Survey.
• Applied a random effects multinomial model with a conditional autoregressive prior for spatial modeling.
• Employed Bayesian inference with Markov Chain Monte Carlo simulation techniques.
• Compared various models using the deviance information criterion.

Main Results:

• Identified significant spatial variation in childhood co-morbidity patterns across Malawi.
• Determined that risk factors for each co-morbidity outcome category differed.
• Key risk factors for co-morbidity included child's age, residence, undernutrition, bednet use, and Vitamin A status.
• Highlighted elevated residual risk levels in central and southern-eastern regions for specific disease combinations.

Conclusions:

• Childhood infectious disease co-morbidity exhibits distinct spatial clustering and varying determinants.
• The linkage between geographical location and childhood health outcomes necessitates further local investigation.
• Overlapping risk factors suggest that integrated disease control strategies could be a cost-effective approach.