If the study data contained no SAEs at all, or if there were no non-serious AEs at all, then the Stata code previously would crash with errors. This has been corrected with some conditional statements.