This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Stat 771 Homework 4, due Monday, April 18 We will consider data from an AIDS trial comparing two drugs, ddI and ddC (Abrams et al., 1995). This was a multicenter, randomized, open label trial to compare ddI to ddC in HIV-infected patients who were intolerant to or had failed ZDV therapy. Participants were men and women, aged 1567, who had either advanced to AIDS or had a CD4 lymphocyte count 300 (this is low). The primary objective is to find out which therapy is better, ddI or ddC, at increasing lymphocyte counts; CD4 T-cell counts are widely-used biomarkers measuring the progression of HIV infection and AIDS (low cells counts are the hallmark of AIDS). There are m = 467 patients that could potentially be observed at 5 time points: 0, 2, 6, 12, and 18 months after therapy started. At 0 months, no there are no differences in the populations, and we will include the CD4 count at 0 months as a base line measurement in the model statement. Instead of working with the CD4 counts directly, we will take the square root and work with Y ij = p CD 4 ij . The square root is a common variance stabilizing transformation for counts, often used with Poisson data. This is a fairly large, messy data set with missing observations, and observations taken at unequally spaced times. The overarching goal is to find out whether there is a treatment effect ( randgrp...
View Full Document
This note was uploaded on 12/14/2011 for the course STA 771 taught by Professor Hanson during the Spring '10 term at South Carolina.
- Spring '10