Developing a spatial-statistical model and map of historical malaria prevalence in Botswana using a staged variable selection procedure
2007

Mapping Historical Malaria Risk in Botswana

Sample size: 17149 publication 10 minutes Evidence: high

Author Information

Author(s): Craig Marlies H, Sharp Brian L, Mabaso Musawenkosi L H, Kleinschmidt Immo

Primary Institution: Malaria Research Programme, Medical Research Council

Hypothesis

The aim of this study was to develop a malaria risk map using a systematic and practicable variable selection process for spatial analysis and mapping of historical malaria risk in Botswana.

Conclusion

The study produced a plausible model of historical malaria risk for Botswana, identifying three key predictors: summer rainfall, mean annual temperature, and altitude.

Supporting Evidence

  • 42 out of 50 potential explanatory variables were significantly associated with malaria prevalence.
  • A Bayesian geo-statistical model was fitted to the data, resulting in a final model of three predictor variables.
  • The model was used to predict malaria prevalence at unobserved locations, producing a smooth risk map for the whole country.

Takeaway

The researchers made a map showing where malaria was likely to be a problem in Botswana a long time ago by looking at weather and land data.

Methodology

The study used a systematic variable selection process, including univariate logistic regression and Bayesian geo-statistical modeling.

Potential Biases

Potential biases may arise from the exclusion of variables based on univariate analysis and the reliance on historical data.

Limitations

The study's predictions may be affected by the spatial correlation structure and the historical nature of the data.

Participant Demographics

The study focused on malaria prevalence in children aged 1–14 years old.

Statistical Information

P-Value

<0.0005

Confidence Interval

95% CI = 1.49 to 2.70 for summer rainfall; 95% CI = 4.14 to 8.08 for annual mean temperature; 95% CI = 1.49 to 2.22 for elevation.

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1476-072X-6-44

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication