Investigating Migration as an important risk determinant of HIV infection among public school educators in South Africa

Abstract

Objective: To investigate the risk determinants of HIV, with focus on migration status of the educator, among South African public school educators.

Methodology: It was imperative to investigate the individual relationship of each of the potential risk factors with the response first. Survey analysis techniques for multilevel clustered data including Survey design based logistic regression, GEE and GLMM were fitted to the data. Comparisons were made to explore how the inferences differ under the different methods of analysis. Finally, multiple imputation technique was employed to deal with data missingness.

Results: The univariate results showed that mobility/migration is an important risk determinant of HIV. When adjusted for other potential risk factors, the different methods applied showed contradicting results. The estimates obtained when missingness was ignored and when it was taken into account are close.

Conclusion: Most of the methods that were applied in this Thesi