Geostatistical Methods for Disease Mapping and Visualisation Using Data from Spatio-temporally Referenced Prevalence...

Area of Science:

• Geostatistics
• Spatial Statistics
• Epidemiology

Background:

• Prevalence surveys generate spatio-temporally referenced data.
• Accurate disease mapping requires robust statistical models.
• Understanding disease distribution is crucial for public health.

Purpose of the Study:

• To provide a tutorial on geostatistical methods for prevalence mapping.
• To identify parsimonious spatio-temporal models for survey data.
• To guide the analysis of disease prevalence data.

Main Methods:

• Development of a variogram-based Monte Carlo procedure for model assumption validation.
• Description and comparison of likelihood-based and Bayesian inference methods.
• Extension of models for non-stationary spatio-temporal covariance functions.

Main Results:

• A framework for accounting for parameter uncertainty in prevalence estimation.
• Exceedance probabilities identified as effective for conveying uncertainty.
• Statistical software for visualizing spatio-temporal predictive summaries via interactive animations.

Conclusions:

• The proposed geostatistical methods facilitate parsimonious prevalence mapping.
• The methods effectively handle parameter uncertainty and non-stationarity.
• The approach is illustrated with historical malaria prevalence data from Senegal.