Modeling and presentation of vaccination coverage estimates using data from household surveys

Area of Science:

• Spatial statistics
• Public health surveillance
• Geographic information systems

Background:

• High-resolution vaccination coverage maps are crucial for public health.
• Household surveys with stratified cluster sampling are commonly used.
• Integrating survey design into geostatistical models presents challenges.

Purpose of the Study:

• To discuss crucial choices in producing vaccination coverage maps using Bayesian geostatistical models.
• To emphasize the importance of acknowledging survey design, including stratification and cluster-level variation.
• To propose novel methods for presenting estimates and uncertainties for improved map interpretation.

Main Methods:

• Fitting Bayesian geostatistical models to stratified cluster survey data.
• Accounting for urban/rural stratification and cluster-level non-spatial excess variation.
• Developing visualization methods for mapping, ranking, and uncertainty control.

Main Results:

• Accounting for survey design, specifically stratification and cluster variation, is vital for accurate geostatistical modeling.
• The geographical scale of estimates impacts precision.
• Novel visualization techniques effectively present probabilistic interpretations and overall map uncertainty.

Conclusions:

• Accurate vaccination coverage mapping requires careful consideration of survey design within geostatistical models.
• Effective presentation of estimates and uncertainties is key for informed public health decisions.
• The proposed methods improve the reliability and interpretability of spatial health data visualizations.