Risk ratio regression-simple concept yet complex computation

In our article we state that if a log-binomial regression with default computation algorithm (Fisher Scoring matrix) converges then it is straightforward to read the risk ratio from the beta-coefficient. In such situations one does not need to use logistic regression or any other regression methods described in our manuscript to estimate the risk ratio. For this case, we agree with Popham that the risk ratio is a simple concept and a simple computation. However, when the log-binomial regression does not converge then one needs to use alternative methods as suggested in the manuscript. The process we refer to as complex computation involves choosing the appropriate regression/computation methods for estimating the maximum likelihood and in turn the risk ratio, standard error and confidence intervals to perform a test of null hypothesis...