Gravar-mail: Validity of indirect comparison for estimating efficacy of competing interventions: empirical evidence from published meta-analyses