The overall goal of this proposal is to accelerate the translation of scientific results from observational and experimental studies into public health policy. Epidemiology is at a crossroads. The vast majority of studies of exposure-outcome relationships in observational and randomized settings alike ask questions about exposure effects: questions which relate to individuals, and to lifestyle choices. For example, a study might ask about how smoking cessation affects risk of heart attack in people with HIV - a question relating largely to exposure-disease etiology. Such estimates are useful, but they are only the first step in making effective public policy, which requires the articulation of a specifi health intervention, and issues of the uptake of the intervention and its real world effectiveness. At present, few methods exist for assessing interventional effects in complex longitudinal data with time-varying exposures: the mainstay of modern epidemiology. This proposal aims to develop and apply innovative epidemiologic methods to create policy-ready, actionable epidemiologic results. Specifically, we will develop and extend two of the g-methods: the parametric g-formula, and inverse probability weighted marginal structural models. We will apply these methodological innovations to critical and timely problems within HIV in both domestic and international contexts. First, we will examine smoking-associated morbidity and mortality among HIV-positive individuals in the United States from the national CNICS database, and estimate the impact of specific smoking cessation interventions on all-cause mortality in this population. Second, we will examine the joint effects of hormonal contraception and pregnancy on response to highly active antiretroviral therapy among HIV-positive women in South Africa, and estimate the impact of specific interventions to reduce unintended pregnancy in this population on population health outcomes including response to treatment. The proposed methods work will help to bridge the gaps between epidemiology and implementation science, and will thus have extremely broad application: not only to the critical issues in HIV, but to epidemiology and public health as a whole. The proposed substantive work will fill needed gaps in the HIV literature, and our approach - rooted in the methods we will develop - will be well-positioned to drive public health policy in these areas. Thus, we propose a high-risk, high-reward effort to create methods to move from exposures to interventions and from patients to policy: through clearly communicated and usable methodological innovation, and high-impact substantive applications.
Typical population-based studies address questions of exposure-disease relationships, which are most relevant to individual lifestyle decisions and clinical care;such studies implicitly compare hypothetical populations in which everyone is exposed, and no one is exposed. Findings from such studies are difficult to apply to questions of health policy for numerous reasons. Here, we propose to innovate and extend epidemiologic methods to create usable tools for translating scientific results of typical studies into public health policy. We will apply these tools to two specific substantive areas related to HIV: mortalit and morbidity attributable to smoking among people in care for HIV;and how pregnancy and use of hormonal contraception affect women's response to antiretroviral therapy.
| Keil, Alexander P; Richardson, David B; Westreich, Daniel et al. (2018) Estimating the Impact of Changes to Occupational Standards for Silica Exposure on Lung Cancer Mortality. Epidemiology 29:658-665 |
| Breskin, Alexander; Cole, Stephen R; Hudgens, Michael G (2018) The Authors Respond. Epidemiology 29:e51 |
| Lesko, Catherine R; Buchanan, Ashley L; Westreich, Daniel et al. (2018) The Authors Respond. Epidemiology 29:e14-e15 |
| Stuart, Elizabeth A; Ackerman, Benjamin; Westreich, Daniel (2018) Generalizability of randomized trial results to target populations: Design and analysis possibilities. Res Soc Work Pract 28:532-537 |
| Richardson, David B; Kinlaw, Alan C; Keil, Alexander P et al. (2018) Inverse Probability Weights for the Analysis of Polytomous Outcomes. Am J Epidemiol 187:1125-1127 |
| Li, Xiaojuan; Cole, Stephen R; Westreich, Daniel et al. (2018) Primary non-adherence and the new-user design. Pharmacoepidemiol Drug Saf 27:361-364 |
| Keil, Alexander P; Edwards, Jessie K (2018) A review of time scale fundamentals in the g-formula and insidious selection bias. Curr Epidemiol Rep 5:205-213 |
| Keil, Alexander P; Edwards, Jessie K (2018) You are smarter than you think: (super) machine learning in context. Eur J Epidemiol : |
| Breskin, Alexander; Cole, Stephen R; Westreich, Daniel (2018) Exploring the Subtleties of Inverse Probability Weighting and Marginal Structural Models. Epidemiology 29:352-355 |
| Keil, Alexander P; Richardson, David B (2018) Quantifying Cancer Risk from Radiation. Risk Anal 38:1474-1489 |
Showing the most recent 10 out of 33 publications
