Ordinal regression model and the linear regression model were superior to the logistic regression models

Norris, Colleen M.; Ghali, William A.; Saunders, L. Duncan; Brant, Rollin; Galbraith, Diane; Faris, Peter; Knudtson, Merril L.; for the APPROACH Investigators,

doi:10.1016/j.jclinepi.2005.09.007

« Previous Next »Journal of Clinical Epidemiology
Volume 59, Issue 5 , Pages 448-456, May 2006

Ordinal regression model and the linear regression model were superior to the logistic regression models

Colleen M. Norris
- Faculty of Nursing, University of Alberta, Edmonton, Alberta, Canada
- Division of Cardiology, Faculty of Medicine, University of Alberta, 4-130F Clinical Sciences Building, Edmonton, Alberta T6G 2G3, Canada
- Corresponding author. Tel.: 780-492-0644; fax: 780-492-1219.
William A. Ghali
- Department of Medicine, University of Calgary, Calgary, Alberta, Canada
- Department of Community Health Sciences, University of Calgary, Calgary, Alberta, Canada
- Centre for Health and Policy Studies, University of Calgary, Calgary, Alberta, Canada
L. Duncan Saunders
- Department of Public Health Sciences, University of Alberta, Edmonton, Alberta, Canada
Rollin Brant
- Department of Community Health Sciences, University of Calgary, Calgary, Alberta, Canada
Diane Galbraith
- Department of Medicine, University of Calgary, Calgary, Alberta, Canada
- Centre for Health and Policy Studies, University of Calgary, Calgary, Alberta, Canada
Peter Faris
- Department of Community Health Sciences, University of Calgary, Calgary, Alberta, Canada
Merril L. Knudtson
- Department of Medicine, University of Calgary, Calgary, Alberta, Canada
for the APPROACH Investigators

Accepted 12 September 2005. published online 15 March 2006.

Abstract

Objective

Ordinal scales often generate scores with skewed data distributions. The optimal method of analyzing such data is not entirely clear. The objective was to compare four statistical multivariable strategies for analyzing skewed health-related quality of life (HRQOL) outcome data. HRQOL data were collected at 1 year following catheterization using the Seattle Angina Questionnaire (SAQ), a disease-specific quality of life and symptom rating scale.

Study Design and Setting

In this methodological study, four regression models were constructed. The first model used linear regression. The second and third models used logistic regression with two different cutpoints and the fourth model used ordinal regression. To compare the results of these four models, odds ratios, 95% confidence intervals, and 95% confidence interval widths (i.e., ratios of upper to lower confidence interval endpoints) were assessed.

Results

Relative to the two logistic regression analysis, the linear regression model and the ordinal regression model produced more stable parameter estimates with smaller confidence interval widths.

Conclusion

A combination of analysis results from both of these models (adjusted SAQ scores and odds ratios) provides the most comprehensive interpretation of the data.

Keywords: Outcomes research, Coronary artery disease, Regression models