Statistical models for CD4 cell counts and HIV RNA viral load in HIV-infected pregnant women and children

Supervisors: Dr Claire Townsend, Dr Mario Cortina Borja, and Dr Claire Thorne

Statistical models for joint immunological/virological markers in HIV-positive women and their infected offspring provide important information to aid decision-making regarding clinical management and treatment, and policy (1). The proposed work is methodological in nature, based on real-life data on HIV-positive women and children collected in several on-going European cohort studies. This research will focus on developing appropriate and efficient statistical models for analysing markers of HIV infection, including HIV viral load. Such models will contribute to a greater understanding of how HIV infection develops in childhood and pregnancy, especially with respect to treatment, and its impact on the immune system: CD4 cells, the preferred target of HIV, the depletion of which leads to the characteristic immunosuppression seen in HIV disease (2). Data patterns that emerge as a consequence of the joint modelling of CD4 cells and viral load may lead to improvements in our understanding of HIV disease progression and response to treatment, and ensure that clinical management is optimised (3). This is particularly important for HIV-infected children, who may have started life-long treatment during infancy (4).

The successful candidate, based within the Centre for Paediatric Epidemiology and Biostatistics, will develop and implement statistical methods based on local dependence and copulae (5) to analyse joint variation in longitudinal CD4 cell counts and viral load measurements in HIV-positive pregnant women and children and disseminate these analyses in the form of conference abstracts and scientific papers. Note that although the project focuses on HIV-positive pregnant women and children, its methodological findings on modelling bivariate distributions will apply to other situations; e.g., where no gold standard for a definitive diagnosis exists and results of two immunoassays are used to assess infection status. The project aims to develop complex statistical methods directly addressing real-life problems posed by observational data. This is an exciting opportunity to learn methodological skills whilst also gaining hands-on experience of working on data from large cohort studies with real and useful applications.

References:
1)European Collaborative Study prepared by Patel D; Cortina-Borja M; Thorne C; Newell M-L. (2007) Time to Undetectable Viral Load after Highly Active Antiretroviral Therapy Initiation among HIV-Infected Pregnant Women. Clinical Infectious Diseases, 44, 1647-1656
2) European Collaborative Study prepared by Bunders M; Cortina-Borja M; Newell M-L; (2005) Age related standards for total lymphocyte CD4} and CD8+ T cell counts in children born in Europe. The Pediatric Infectious Disease Journal, 24, 595-600
3) Thiebaut R; Jacqmin-Gadda H; Babiker A; et al. (2005) Joint modelling of bivariate longitudinal data with informative dropout and left-censoring, with application to the evolution of CD4+cell count and HIV RNA viral load in response to treatment of HIV infection. Statistics in Medicine, 24, 65-82
4) Lewis J; Walker SA; Castro H; et al (2012) Age and CD4 Count at Initiation of Antiretroviral Therapy in HIV-Infected Children: Effects on Long-term T-Cell Reconstitution. Journal of Infectious Diseases, 205, 548-556
5) Yan J (2007) Enjoy the Joy of Copulas: With a Package copula. Journal of Statistical Software, 21, 1-18.