Bridging the gap between design and analysis when making inference using observational HIV data
Abstract/Contents
- Abstract
- The human immunodeficiency virus (HIV) is still one of the main leading causes of death in the world. Depending on the geographic area of interest, different challenges around HIV remain to be addressed. The overall goal of this dissertation was to use different statistical and epidemiologic designs to address some current HIV-related issues, namely, i) morbidity and mortality of African HIV-exposed uninfected infants; ii) adherence and immunological outcomes among adult People Living with HIV in the United States of America ; iii) The use of transportability methods to improve causal inference among understudied subgroups in the context of HIV research. Chapter 1 of this dissertation is an original manuscript published in AIDS, "Breastfeeding mitigates the effects of maternal HIV on infant infectious morbidity in the Option B+ era". The goal of this chapter was to assess whether HIV-exposed uninfected infants were at a higher risk of infectious disease morbidity and mortality when compared with HIV-unexposed infants in African settings with good prevention of mother-to-child transmission and good exclusive breastfeeding rates using a prospective cohort study. I led the conceptualization, the analysis, the writing, and the preparation of the manuscript. Chapter 2 studied the relationship between adherence to Antiretroviral therapy (ART) and immunological and virologic outcome in a large population of adult People Living With HIV (PLWH) initiating ART at Kaiser Permanente Northern California (KPNC). In this chapter, an algorithm that estimates longitudinal adherence using pharmacy records over multiple six-month intervals was first developed. We then, assessed what variables were associated with longitudinal adherence and modeled the association between adherence and future CD4 counts and adherence and future viral load. Chapter 3 is a methodological paper that was informed by the limitations of chapters 1 and 2 as well as the lack of representativeness in current HIV clinical trials. Chapter 3 assessed how violations in the transportability methods' assumption affect the average treatment effect (ATE) estimation in the target population. This chapter also illustrates how we can use transportability methods and empirical Bayes to design a more efficient hybrid trial. Finally, we conclude that chapter by exploring how reconciling adaptive trial designs and the transportability framework can help us identify sources of violations in the transportability assumptions and improve our understanding of the Heterogeneous Treatment Effects(HTEs) within different subgroups in the target population.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2020; ©2020 |
| Publication date | 2020; 2020 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Toukam Tchakoute, Christophe |
|---|---|
| Degree supervisor | Baiocchi, Michael |
| Thesis advisor | Baiocchi, Michael |
| Thesis advisor | Desai, Manisha |
| Thesis advisor | Maldonado, Yvonne |
| Thesis advisor | Sainani, Kristin |
| Degree committee member | Desai, Manisha |
| Degree committee member | Maldonado, Yvonne |
| Degree committee member | Sainani, Kristin |
| Associated with | Stanford University, Department of Epidemiology and Population Health |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Christophe Toukam Tchakoute. |
|---|---|
| Note | Submitted to the Department of Epidemiology and Population Health. |
| Thesis | Thesis Ph.D. Stanford University 2020. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2020 by Christophe Toukam Tchakoute
- License
- This work is licensed under a Creative Commons Attribution Non Commercial No Derivatives 3.0 Unported license (CC BY-NC-ND).
Also listed in
Loading usage metrics...