Improving Health Data in Africa Using Geostatistics
Author Information
Author(s): Peter W. Gething, Abdisalan M. Noor, Priscilla W. Gikandi, Esther A. A. Ogara, Simon I. Hay, Mark S. Nixon, Robert W. Snow, Peter M. Atkinson
Primary Institution: University of Southampton
Hypothesis
Can space-time geostatistics improve the accuracy of health management information systems in Africa?
Conclusion
The modelling framework provides reliable information from imperfect health management information system data to support evidence-based decision-making.
Supporting Evidence
- The overall reporting rate of health facilities was only 35%.
- The model predicted national annual treatment burdens for presumed outpatient malaria with a margin of error of 1.3%.
- Space-time kriging can estimate large numbers of missing data to acceptable levels of accuracy.
Takeaway
This study shows how a smart math method can help fill in missing health data, making it easier to understand how many people are sick and need help.
Methodology
The study used space-time kriging to predict missing malaria case data from an incomplete health management information system in Kenya.
Potential Biases
Potential biases may arise from the incomplete nature of the health management information system data.
Limitations
The model's predictive power decreases at finer spatial and temporal resolutions, and it requires a spatially referenced map of health facilities.
Participant Demographics
Data were collected from 2,165 government health facilities across Kenya.
Statistical Information
Confidence Interval
95% confidence interval for predictions of total malaria cases was estimated.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website