Therneau and Lumley (2015) reported data on recurrences of bladder cancer, which were used to demonstrate methodology for recurrent event modelling (Wei et al., 1989). This is the data set from 85 subjects who were assigned to either thiotepa or placebo, and reports the first four recurrences for any patient. The stop variable is the tumor recurrence or censoring time in month. The event variable is 1 for recurrence and 0 for everything else (including death for any reason).
Thall and Vail (1990) presented longitudinal data from a clinical trial of 59 epileptics, who were randomized to a new drug or a placebo (T=0 or T=1). Baseline data were available at the start of the trial; the trial included the logarithm of the average number of epileptic seizures recorded in the 8-week period preceding the trial (B), the logarithm of age (A), and visit (V: a linear trend, coded (-3,-1,1,3)/10). A multivariate response variable (y) consists of the seizure counts during 2-week periods before each of four visits to the clinic.
This package allows different models for multivariate response variables where each response is assumed to follow DHGLMs (Lee and Nelder, 2006; Lee, Nelder and Pawitan, 2006, 2016). With DHGLMs, the mean and dispersion parameters for the variance of random effects and residual variance (overdispersion) can be modeled as random-effect models. Furthermore, the mdhglm package allows various correlation structures among the random effects in DHGLMs of different response variables. Functions are provided to format and summarize the mdhglm results. The estimates of fixed effects, random effects and their variances and correlations are calculated.
(Package: mdhglm) :
Repatead Measures on AIDS Data
Both longitudinal and survival data were collected in a recent clinical trial to compare the efficacy and safety of two antiretroviral drugs in treating patients who had failed or were intolerant of zidovudine (AZT) therapy (Rizopoulos, 2015). In this trial, 467 HIV-infected patients were enrolled and randomly assigned to receive either didanosine (ddI) or zalcitabine (ddC). The number of CD4 cells per cubic millimeter of blood were recorded at study entry, and again at the 2, 6, 12, and 18 month visits.
● Data Source:
● Alias: aids
Price et al. (1985) presented data from a study on the developmental toxicity, ethylene glycol (EG),in mice. Responses are the malformation (binary response) and fetal weight (continuous response) outcomes for the experiment, which shows clear dose-related trends with respect to both outcomes.
An experiment on integrated circuits was reported by Phadke et al. (1983). The width of lines made by a photoresist-nanoline tool were measured in five different locations on silicon wafers, measurements being taken before and after an etching process being treated separately. Here, the pre-etching data are analyzed. The eight experimental factors (A-H) were arranged in an L18 orthogonal array and produced 33 measurements at each of five locations, giving a total of 165 observations. There were no whole-plot (i.e. between-wafer) factors.
This example is taken from Myers et al. (2002). It involves a designed experiment in a semiconductor plant. Six factors are employed, and it is of interest to study the curvature or camber of the substrate devices produced in the plant. There is a lamination process, and the camber measurement is made four times on each device produced. The goal is to model the camber taken in 10E-04 in. as a function of the design variables. Each design variable is taken at two levels and the design is a 2^(6-2) fractional factorial. The camber measurement is known to be nonnormal.
(Package: mdhglm) :
Vascular Cognitive Impairment Data
Lee, Nelder, and Pawitan (2016) considered the Vascular Cognitive Impairment (VCI) data. As the elderly population increases, the proportion of stroke, one of main geriatric diseases also rises. The increase of the number of patients with stroke results in the rise of that of patients with the VCI which cognitive function is declined due to stroke. Through an early intervention for the VCI, the cognitive function can be improved. They considered standardized VCI scores for four domains (response variables): executive (y1), memory (y2), visuospatial (y3) and language (y4). The purpose of the study is to examine the effects of ten demographic and ten acute neuroimaging variables on the cognitive function in the ischemic stroke patients.