Low-value glucose monitoring in noncritically ill hospitalized patients.

Abstract

Inpatient routine point-of-care glucose (POC-G) monitoring is common given association of persistent hyperglycemia with increase in morbidity and mortality. However, some patients receive frequent fingerstick testing without needing insulin to reach glycemic targets. Determine the prevalence and characteristics of noncritically ill inpatients undergoing POC-G monitoring who do not meet guideline criteria for hyperglycemia requiring insulin initiation. We performed a secondary analysis of noncritically ill adult inpatients (≥18 years) with at least four glucose measurements, discharged from five hospitals over 4.4 years. Exclusions included type 1 diabetes, admission glucose >500 mg/dL, long/intermediate-acting insulin use, stays ≤24 h, ICU admissions, or inpatient use of noninsulin diabetes drugs. The primary outcome was the maximum POC-G value during hospitalization. Among 23,134 patients undergoing POC-G monitoring on correctional insulin monotherapy, 20.9% had all POC-G ≤ 140 mg/dL and averaged 9.5 POC-G tests per hospitalization. Patients with POC-G levels of 141-179 mg/dL underwent an average of 15.8 tests per hospitalization, with 75% receiving no insulin or only a one-time insulin administration. Overall, 149,479 (39%) of POC-G tests were conducted in patients with maximum glucose <180 mg/dL who did not meet guideline criteria for insulin initiation. The estimated labor and supply costs ranged from $241,483 to $826,868 during the study period-driven primarily by 0.8 to 1.4 FTEs annually dedicated to low-value POC-G testing. Low-value POC-G testing is common among non-critically ill inpatients on correctional insulin monotherapy. Reducing unnecessary monitoring can decrease patient discomfort and preserve healthcare resources.