ABSTRACT
255Confounding by a continuous variable is often a problem in an analysis of rates. Common examples include age and calendar time. Continuous variables may also be an exposure of interest. Quadratic splines, restricted cubic splines, and fractional polynomials can be used to efficiently and flexibly express a continuous variable as an exposure or confounding variable. An example models the association between a continuous variable (age) and death in a traffic crash, using a conditional Poisson model.
