Gravar-mail: Sample size considerations for historical control studies with survival outcomes