Causal inference methods to quantify impacts of health policies: Applications to measure HIV and tuberculosis treatment and prevention effectiveness in global public health settings Restricted; Files Only
Zissette, Seth (Spring 2026)
Abstract
HIV and tuberculosis (TB) remain pressing globally despite advances in treatment and prevention. Biomedical interventions have demonstrated effectiveness in clinical trials and some real-world settings, but impacts on population-level outcomes require further investigation. Large observational data systems can help researchers answer such questions, but researchers must address complex biases that arise when analyzing these data. The overarching goal of this dissertation was to apply causal inference methods to quantify the effectiveness of biomedical interventions for HIV and TB treatment and prevention in global public health settings using observational data.
AIM 1: Isoniazid preventive treatment (IPT) is a key tool in preventing TB disease, but knowledge on its effectiveness for people with HIV (PWH) in high-burden multidrug-resistant (MDR)-TB settings is limited. We used target trial emulation to examine the effectiveness of IPT among PWH entering HIV care from 2018-2021 in Ukraine. IPT initiation was associated with reduced TB incidence, and target trial emulation helped build this evidence while avoiding biases.
AIM 2: Clinical trials have demonstrated decreased care attrition from care and other benefits for PWH who initiate antiretroviral therapy (ART) on the same day as HIV diagnosis, but observational studies have found associations with increased attrition. We examined the effectiveness of same-day ART initiation among PWH entering HIV care from 2018-2021 in Haiti using target trial emulation. Same-day ART initiation was associated with increased treatment interruption and decreased engagement in care, but the effect on treatment interruption decreased over time.
AIM 3: IPT completion is important to maximize protection, but patients may discontinue at various points. We used target trial emulation to examine the effectiveness of restarting IPT among PWH who discontinued IPT from 2018-2022 in Haiti. Restarting was not associated with TB or mortality on average but may benefit PWH who discontinued early; a lack of statistical power necessitates further study.
The results of this dissertation can help researchers and policymakers to quantify the effects of global HIV and TB treatment and prevention programs with rigorous methods designed to generate accurate, accessible, and actionable evidence. We hope that the work of this dissertation will help expand the analytic toolkit of scientists to address epidemiological questions as accurately as possible and contribute evidence to improve the lives of all people affected by the epidemics of HIV and TB.
Table of Contents
Chapter 1 : Introduction
Chapter 2 : Target trial emulation of a realistic intervention in a complex setting: An application to isoniazid preventive treatment for people with HIV in a high-burden multidrug-resistant tuberculosis country
Chapter 3 : Target trial emulation of the programmatic effectiveness of same-day antiretroviral therapy initiation in Haiti
Chapter 4 : Target trial emulation of the programmatic effectiveness of restarting isoniazid preventive treatment following non-completion for people with HIV in Haiti
Chapter 5 : Summary of Results and Conclusions
Chapter 6 : References
About this Dissertation
| School | |
|---|---|
| Department | |
| Degree | |
| Submission | |
| Language |
|
| Research Field | |
| Keyword | |
| Committee Chair / Thesis Advisor | |
| Committee Members |
Primary PDF
| Thumbnail | Title | Date Uploaded | Actions |
|---|---|---|---|
|
|
File download under embargo until 27 May 2028 | 2026-04-17 13:52:09 -0400 | File download under embargo until 27 May 2028 |
Supplemental Files
| Thumbnail | Title | Date Uploaded | Actions |
|---|