CANSA Research in Action Conference – July 2018 – Poster Session for Students – Irene Kyomugisha

CANSA Research in Action Conference – July 2018 – Poster Session for Students – Irene Kyomugisha

Title of the type A grant: Predictive models for HIV related Lymphomas

Name of PI:  Prof Farai Nyabadza from Stellenbosch University, specialising in predictive models for HIV-related lymphomas. This work has been done for the award of a Masters degree in Biomathematics


Modelling Early Initiation of HAART to Forestall AIDS-related Lymphoma in South Africa: a Cost-Effectiveness Analysis

Irene Kyomugisha1, Farai Nyabadza1, Beate Sander2, Cang Hui1 and Akin Abayomi1 (1) Stellenbosch University, (2) University of Toronto


With the advent of highly active antiretroviral therapy (HAART) in South Africa, a marked decline in HIV-related illnesses and premature death has been observed; however, AIDSrelated lymphoma incidence increased despite the large HAART roll-out programme. Earlier initiation of HAART can reduce the risk of developing non-Hodgkin lymphoma but needs to be balanced with increased upfront costs in resource-limited settings.


We developed a linked transmission and health state transition model to determine the cost-effectiveness of early HAART initiation. The CD4 count driven deterministic model predicted lymphoma incidence in HIV-infected adults (aged 15 to 80 years) over a period of ten years. The Markov model predicted health outcomes and costs. We used data from the Tygerberg Lymphoma Study group (TLSG) and existing literature. We compared early initiation of HAART at a CD4 cell count threshold of greater than 500 cells/μL to initiation at less than 500 cells/μL. Our primary outcomes were Quality-adjusted life years (QALYs), expected costs, net monetary benefit and the incremental cost-effectiveness ratio (ICER). Health outcomes and costs are discounted at 5% per annum as recommended in Southern Africa. We performed deterministic sensitivity analyses to assess parameter uncertainity.


Early HAART initiation prevents lymphoma cases and related deaths, translating to 3.76 QALYs gained over the 10-year time horizon (6.47 vs. 2.71 expected QALYs with HAART initiation at >500 cells/μL and <500 cells/μL, respectively). The incremental cost of early initiation was $14,613 compared to the alternative. The NMB of early initiation was $90,581 and $30,063 for the alternative. HAART initiation at greater than 500cells/μL was therefore cost-effective with an ICER of $3,890/QALY gained. Sensitivity analysis showed outcomes were sensitive to the effectiveness of HAART in preventing lymphoma, with early initiation being more sensitive than the base case.


Early HAART initiation would not only be effective, but also cost-effective in forestalling AIDS-related lymphoma in resource limited settings.


Do you have a question?