Standardization by a disease risk score (DRS) may be preferable to weighting on the exposure propensity score if the exposure is difficult to model, relatively novel (i.e., newly emerging or rapidly-evolving), or extremely rare. For exposures with more than two levels, methods are lacking for a DRS-based approach. We present an approach to estimate trends in standardized risk ratios (RRs) based on a regression model that uses a DRS.