ARTnet: Model Parameterization with the ARTnet Study Data

ARTnet is an anonymous cross-sectional web-based survey of HIV-related risk behaviors, testing, and use of prevention services among men who have sex with men (MSM) in the United States. It recruited MSM who have completed the American Men’s Internet Survey (AMIS) study, and therefore the dataset contains variables merged from that study as well. Full access to the dataset from ARTnet will allow the researchers to conduct analyses and disseminate results using the data.

Access to the data requires a Memorandum of Understanding (MOU) that outlines the personnel analyzing the data and purposes of the data analyses. This dataset may not be shared without the consent of the ARTnet Study PI (Samuel Jenness, Emory University) as outlined in an MOU.

ARTnetData Dependency

The ARTnet package depends on the ARTnetData package, which contains the limited use public dataset. Because of the restrictions of the dataset, the ARTnetData package must be installed separately, before installing the ARTnet package, using the following directions.

Installation

The suggested method for accessing the dataset is to directly install the ARTnetData package in R, using the remotes package as follows:

remotes::install_github("EpiModel/ARTnetData")

Because this repository is private, installing this package may require setting up a Github Personal Access Token.

This package comes with two included datasets, a wide-form dataset (rows = study participants) and a long-form dataset (rows = partnerships, with multiple rows per unique study participant).

Dataset Loading

The suggested method for accessing the dataset is to directly install the ARTnetData package in R, using the remotes package as follows:

library("ARTnetData")
d <- ARTnet.wide
l <- ARTnet.long

The built-in dataset names for the two datasets are ARTnet.wide and ARTnet.long, and they are "lazy loaded" into global memory when the ARTnetData package is loaded. To use or modify the datasets, you might start by assigning those datasets short-hand names. Then any R operations may be performed.

str(d)
str(l)

The ARTnet Package

The ARTnet package contains standardized scripts to analyze the ARTnet data for the purposes of parameterizing the epidemic modeling with EpiModel and EpiModelHIV. There are three primary functions, detailed below, that conduct statistical analysis of the data for a specific target population of MSM defined by geography, age, and race/ethnicity. Users may also conduct analyses of the ARTnet dataset without the ARTnet package, but this package automates several standard analyses needed for many epidemic modeling projects.

Installation

The ARTnet package may be installed with the remotes package:

remotes::install_github("EpiModel/ARTnet", build_vignettes = TRUE)

Example Uses

Some of the example uses are then as follows:

# 1. Epistats: Specify geographic features, as well as race stratification 
#              and total age range
epistats <- build_epistats(geog.lvl = "city", 
                           geog.cat = "Atlanta", 
                           race = TRUE, age.limits = c(30, 50),
                           age.breaks = c(35, 45))

# 2. Netparams: Specify age categories if needed, or let ARTnet determine 
                age categories by number of categories desired
netparams <- build_netparams(epistats = epistats, smooth.main.dur = TRUE)

# 3. Netstats: Finalize network setup 
netstats <- build_netstats(epistats, netparams, expect.mort = 0.0005, 
                           network.size = 1000, edges.avg = TRUE)

More details of which may be found in the package vignette:

vignette(package = "ARTnet")