Gravar-mail: Prevalence odds ratio v prevalence ratio--some further comments.