This proposal addresses statistical challenges that arise in HIV prevention research. Recent progress in prevention from different uses of antiretroviral drugs (in PrEP, microbicides, and treatment), male circumcision, and possibly vaccine has been noteworthy but modest. Combining across these modalities, however, might considerably enhance their effects. To address a crucial question facing policy makers--how best to deploy these modalities--requires knowledge of their combined effects and of the factors that impact on these effects. We propose to facilitate such research by developing efficient methods for analysis of data from observational and randomized studies of prevention interventions and for investigation of factors that may modify their degree of benefit. To do so, we propose to extend methods in variety of statistical areas, including causal inference and semiparametric theory, network analysis, and spatial analysis. We show how knowledge gained from new methods in each area can contribute to the usefulness of those in the other areas.
Our first aim i s to develop methods for using baseline covariates to improve efficiency of analyses of correlated outcome data that arise in group randomized trials (GRTs) and longitudinal studies;both types of studies are widely used in HIV prevention research. Because the number of experimental units in GRTs may be constrained by resource limitations, we develop methods appropriate for small as well as for large samples, and consider both individual- and group-level covariates. Estimation of network features that are important epidemic drivers from network samples is our second aim.
This aim also considers methods for using such estimates in the construction of collections of networks--essential for realistic simulation of epidemics and of the impact of combination prevention interventions on them. These investigations will be useful for assessing whether intervention packages should be tailored to network features (and if so how), both through simulation and in analyses of data from research studies.
The third aim develops methods for using data from HIV surveillance, especially of antenatal clinics, to estimate spatial and temporal variation in age-specific HIV prevalence and incidence. Such methods can aid in evaluating the impact of deployment of prevention interventions, especially those which are rolled out in sequentially in different regions.

Public Health Relevance

This proposal addresses statistical challenges that arise in HIV prevention research. We focus on development of efficient methods for analysis of data from observational and randomized studies of prevention interventions and for investigating factors, like features of sexual networks, that impact on their success. We do so by extending methods in variety of statistical areas, including causal inference and semiparametric theory, network analysis, and spatial analysis. We show how knowledge gained from new methods in each area can contribute to the usefulness of the others in guiding deployment of prevention interventions.

Agency
National Institute of Health (NIH)
Institute
National Institute of Allergy and Infectious Diseases (NIAID)
Type
Method to Extend Research in Time (MERIT) Award (R37)
Project #
5R37AI051164-12
Application #
8586288
Study Section
AIDS Clinical Studies and Epidemiology Study Section (ACE)
Program Officer
Gezmu, Misrak
Project Start
2002-03-01
Project End
2015-11-30
Budget Start
2013-12-01
Budget End
2014-11-30
Support Year
12
Fiscal Year
2014
Total Cost
$363,375
Indirect Cost
$138,375
Name
Harvard University
Department
Biostatistics & Other Math Sci
Type
Schools of Public Health
DUNS #
149617367
City
Boston
State
MA
Country
United States
Zip Code
02115
Marden, Jessica R; Wang, Linbo; Tchetgen, Eric J Tchetgen et al. (2018) Implementation of Instrumental Variable Bounds for Data Missing Not at Random. Epidemiology 29:364-368
Dutta, Ritabrata; Mira, Antonietta; Onnela, Jukka-Pekka (2018) Bayesian inference of spreading processes on networks. Proc Math Phys Eng Sci 474:20180129
Naimi, Ashley I; Balzer, Laura B (2018) Stacked generalization: an introduction to super learning. Eur J Epidemiol 33:459-464
Balzer, Laura B; Zheng, Wenjing; van der Laan, Mark J et al. (2018) A new approach to hierarchical data analysis: Targeted maximum likelihood estimation for the causal effect of a cluster-level exposure. Stat Methods Med Res :962280218774936
Gurmu, Yared; Qian, Jing; De Gruttola, Victor (2018) A Sexual Partnership Duration: Characterizing Sampling Conditions That Permit unbiased Estimation of Survivorship and Effect on It of Covariates. Res Rev J Stat Math Sci 4:22-35
Hitchings, Matt D T; Lipsitch, Marc; Wang, Rui et al. (2018) Competing Effects Of Indirect Protection And Clustering On The Power Of Cluster-Randomized Controlled Vaccine Trials. Am J Epidemiol :
Goyal, Ravi; De Gruttola, Victor (2018) Inference on network statistics by restricting to the network space: applications to sexual history data. Stat Med 37:218-235
Carnegie, Nicole Bohme (2018) Effects of contact network structure on epidemic transmission trees: implications for data required to estimate network structure. Stat Med 37:236-248
Oldenburg, Catherine E; Seage, George R; Tanser, Frank et al. (2018) Antiretroviral Therapy and Mortality in Rural South Africa: A Comparison of Causal Modeling Approaches. Am J Epidemiol :
Balzer, Laura; Staples, Patrick; Onnela, Jukka-Pekka et al. (2017) Using a network-based approach and targeted maximum likelihood estimation to evaluate the effect of adding pre-exposure prophylaxis to an ongoing test-and-treat trial. Clin Trials 14:201-210

Showing the most recent 10 out of 57 publications