BuyseTest (version 1.3.2)
Generalized Pairwise Comparisons
Description
Implementation of the Generalized Pairwise Comparisons. This test
enables to compare two groups of observations in randomized trials(e.g treated
vs. control patients) on several prioritized outcomes. Pairwise comparisons
require consideration of all possible pairs of individuals, one taken from the
treatment group and the other taken from the control group. The outcomes of the
two individuals forming a pair are compared. Thresholds of minimal clinically
significant differences can be defined. It is possible to analyze simultaneously
several outcomes by prioritizing the variables that capture them. The highest
priority is assigned to the variable considered the most clinically relevant.
A natural way of handling uninformative or neutral pairs is to consider the
outcomes in descending order of priority: whenever a pair is uninformative or
neutral for an outcome of higher priority, the outcomes of lower priority are
examined In the case of time-to-event endpoint, four methods to handle censored
observations are available in this package (Gehan, Peto, Efron, and Peron).