Simple confidence intervals for standardized rates based on the approximate bootstrap method.

This paper presents simple expressions for the confidence intervals of indirect and direct standardized rates, based on the approximate bootstrap confidence (ABC) method of DiCiccio and Efron. For indirect rates, the ABC method compares favourably with the exact methods. An exact method is not available for direct standardized rates. I, therefore, compared the ABC to other simple procedures for confidence intervals, namely: the standard normal, log-normal, and a recent method proposed by Dobson et al. Simulation studies of the coverage properties of these methods for direct standardized rates show that the ABC method performs best, with balanced tail probabilities close to their expected values.