Frequently repeated glucose measurements overestimate the incidence of inpatient hypoglycemia and severe hyperglycemia.