Gravar-mail: Influence of blinding on treatment effect size estimate in randomized controlled trials of oral health interventions