A revised SAS macro for maximum likelihood estimation of prevalence ratios using the COPY method.
J Occup Environ Med 2009 Sep; 66(9):639
In a recent article (Approaches for estimating prevalence ratios. (Occup Environ Med 2008;65:501-6) in the Education series of this Journal, Deddens and Petersen discuss several methods for modelling and estimating prevalence ratios. When the underlying distribution of the dependent variable is binomial, and the log of the binomial probability is related to a set of independent variables in a linear manner, approximate maximum likelihood estimates (MLEs) of the model parameters and prevalence ratios can be obtained with one of the methods, namely the COPY method. In 2003, Deddens et al supplied a macro for performing the COPY method calculations.1 They also provided a SAS macro for obtaining the exact MLEs for the simple case in which the model contained a single one degree of freedom independent variable.
Mathematical-models; Models; Risk-factors; Standards; Statistical-analysis; Workplace-studies; Work-practices
Martin R Petersen, National Institute for Occupational Safety and Health, DSHEFS, Mail Stop R15, 4676 Columbia Parkway, Cincinnati, OH 45226
Journal of Occupational and Environmental Medicine