|Friday, February 16|
|PS2 Poster Session 2 and Refreshments||
Fri, Feb 16, 5:15 PM - 6:30 PM
Software for Covariate Specification in Linear, Logistic, and Survival Regression (303599)
Margaret R. Stedman, Stanford University
Keywords: Covariate Specification, Functional Form Specification, Linear Regression, Logistic Regression, Survival Analyses, Data Visualization, Programming
One of fundamental parts of developing a regression model is to specify an appropriate functional form of a covariate. The choice to include a variable as continuous, categorical, or as a spline can be determined by model fit. We developed an efficient and user-friendly SAS® macro (%SPECI) to help analysts determine how best to specify the appropriate functional form of a continuous covariate in linear, logistic, and survival models. For each model, our macro provides a graphical and statistical single page comparison report of the covariate as a continuous, categorical, log-transformed, squared, and restricted cubic spline variable so that users can easily compare and contrast results. The report also includes the residual plot and distribution of the covariate. The analysts can include all other covariates in the model for multivariable adjustment. The output displays the likelihood-ratio test, the Akaike Information Criterion (AIC), Sawa’s Bayesian Information Criterion (BIC) and/or other model-specific statistics. The %SPECI macro is demonstrated using an example data set.