The Continuation-ratio Model as a Tool for Exploratory Analysis of Clinical Research Data: An Illustrative Analysis of Delirium in a Hospitalized Elderly Cohort

• Activity Number: 152
• Type: Topic Contributed
• Date/Time: Monday, August 9, 2004 : 2:00 PM to 3:50 PM
• Sponsor: Section on Government Statistics
• Author(s): Peter H. Van Ness*+
• Companies: Yale University School of Medicine
• Keywords: continuation-ratio ; ordinal ; risk-factor ; exploratory ; clinical ; power

Abstract

The objective of much clinical epidemiologic research is the identification of disease risk factors. A second common goal of clinical research is to prevent morbidity that has occurred from becoming more severe, persistent, or disabling. To achieve this second goal it is important to determine if the same risk factors associated with disease occurrence are also associated with disease advancement and if the magnitude of their effect on the latter is the same as on the former. Fitting a continuation ratio regression model--a generalization of logistic regression for analyzing sequentially ordinal outcomes--can help achieve this second goal. Furthermore, it does so chiefly by calculating goodness-of-fit statistics and plotting related graphs, and thus can be implemented in an exploratory analysis. Interpretation of the results of exploratory data analyses involving the continuation-ratio model, especially when supplemented by relevant power calculations, can provide clinical researchers with valuable guidance in establishing protocols for confirmatory data analyses, even those using other models. This approach will be illustrated by an exploratory analysis of data.