HIV disease progression modelling using phase type survival trees

Abstract

Disease progression models are useful tools in analysing disease progression where HIV disease is termed as progressing if the CD4 count goes down. This is because the CD4+ T cell count is used as the main predictor of HIV disease progression and patients are classified into four states based on this count. Currently no effective cure for HIV disease exists and monitoring closely the disease progression is crucial. Hence, the aim of this dissertation is to model progression of HIV disease by clustering and quantifying the effects of several covariates and their interaction in the prediction of HIV disease progression using phase type survival trees. This is done by using a real database of 1,838 HIV infected patients and as a result the covariates which have prognostic significance on the disease progression are identified. The outcome of this research work should aid to effectively manage HIV disease, while the methods developed can also be useful for modelling disease progression in patients who have other chronic conditions or diseases.