Do biometeorological indices improve modeling outcomes of heat-related mortality?