Prognostic factors and risk groups: Some results given by using an algorithm suitable for censored survival data

The problem of identifying the role of various prognostic factors when the response variable is the patient's survival time is discussed. Morgan and Sonquist's approach of constructing groups at similar risk is useful and an algorithm which implements the method for survival time data with censored observations is described. Examples of its application to two sets of medical data are given, arising from studies of prognosis following renal transplantation and treatment for stage I malignant melanoma of the skin.

The problem of identifying the role of various prognostic factors when the response variable is the patient's survival time is discussed. Morgan and Sonquist's approach1 of constructing groups at similar risk is useful and an algorithm which implements the method for survival time data with censored observations is described. Examples of its application to two sets of medical data are given, arising from studies of prognosis following renal transplantation and treatment for stage I malignant melanoma of the skin. Copyright © 1983 John Wiley & Sons, Ltd.