|
English Français |
Potiphar Damiano, University of Malawi
Tsirizani Kaombe, University of Malawi
The burden of maternal anaemia is high in sub-Saharan Africa, which derails safe motherhood campaign efforts in the region. Thorough analysis of maternal anaemia data will help in determining the best strategies for mitigating this health problem. The study evaluates performance of robust regression methods and diagnostic statistics on quality of estimates and detection of unusual observations when applied to both simulated data and maternal anaemia data in Malawi. The study observes points of agreement and departures in detecting unusual observations upon applying the various statistics under these two techniques on the anaemia dataset, to improve modelling of maternal anaemia data. The performance of two techniques was further evaluated through a simulation study. The methods were applied to analysis of marternal anaemia data in Malawi to detect outliers and influential points in the fitted linear model. All the computations were carried out in R software version 4.3.0.
No extended abstract or paper available
Presented in Session 60. Joint Modelling of Health Outcomes in Sub-Saharan Africa