actg.RdA dataset obtained from the Aids Clinical Trials Group.
actgA tibble with the following columns.
id: patient identifier
sex: Male, Female
race: White, Black, Hispanic
idu: 1=history of injection drug use, 0=no history of injection drug use
art: 1=triple therapy, 0=dual therapy
art_f: factor variable for treatment group (1=triple therapy, 0=dual
therapy)
r: 1=in subcohort, 0=not
age: years of age at randomization
karnof: Karnofsky score (100=normal, 0=dead)
days: days from randomization to endpoint of censoring
cd4: CD4 count in cells/mm^3 at randomization
stop: day therapy stopped, missing if not stopped
Y: time of event (AIDS or death)
C1: time of censoring due to drop-out
C2: time of censoring not due to drop-out
There is one record per patient in the dataset. Note that r is a
constructed variable provided to illustrate confounding.
The AIDS Clinical Trials Group (ACTG) 320 compared a combination of three antiretroviral drugs (two nucleoside analogues, zidovudine and lamivudine, plus a protease inhibitor, indinavir) against a pair (zidovudine and lamivudine alone). The dataset consists of 1,156 HIV-positive patients enrolled in the trial between January 29, 1996 and January 27, 1997. At screening, all patients had a CD4 cell count <= 200 cells per cubic millimeter, at least 3 months of prior zidovudine therapy, and no previous treatment with lamivudine or protease inhibitors. Patients were stratified based on CD4 count (<= 50 vs. 51-200) and randomly assigned to the three-drug or two-drug regimen. The outcome of interest was the occurrence of AIDS or death, and the findings from the trial were reported in Hammer's 1997 paper in the New England Journal of Medicine. Trial participants were followed until AIDS diagnosis, death, loss to follow-up, or administrative censoring. Covariates measured at baseline and available in the dataset are age, sex, race, ethnicity, Karnofsky score, CD4 cell count, and history of injection drug use.