Predicted mortality patterns in cohort study populations exposed to different types of hazard: can SMRs show a dose-response?

The analysis of data from cohort studies by duration of exposure and time since first exposure (latent interval) has been investigated using a simulation approach. These two analyses are often performed when the estimator of risk, commonly the O/E ratio or SMR, shows evidence of a raised risk. Two measures of risk were tested in the detailed analyses: the O/E ratio and the excess death rate. The latter was found to reflect the 'dose-response' relationship more accurately for a variety of hazard types. Attention is drawn to possible biases and to the importance of adjustment for age in these analyses.